Bucker Engines

The engines that provided most of the data for this report are the E.N.M.A. Tigre G-IV-B5.

They were both overhauled before we started to use them – one in Germany the other in Spain. Since flying began with one engine in 2003 and the other in 2004 they have used/and are equipped with the following

• only NEW (as of 2004) Teflon hoses,
• only 100LL aviation fuel, with TCP added,
• only Aeroshell 100 Mineral oil, changed every 20 to 25 hours,
• only Aeroshell 14 grease or German HeiBlagerfett Extra (diamant) for the rockers,
• rocker arm axle tube ends modified to take grease internally via zerk fittings
• the factory- equipped fuel pump in series with a Christen wobble pump,
• the factory-equipped 5 stage oil pump, but the oil cooler has been removed,
• the factory-equipped (Cuno type) oil filter,
• the factory-equipped square-style intake manifold,
• the factory-equipped square-style carburetor inlet plenum - not the 120 degree tube
• the factory-equipped magnetos  (there at least 3 different types!),
• magneto timing retard restrictor – for a smoother idle,
• Slick Magneto Harness kit M-6201-4B (spark plug wires),
• Champion REM37BY spark plugs (or Unison/Tempest equivalent),
• Ellison EFS-4 throttle body injector (TBI),
• Hoffman HO23F-212 147K propeller
• ……………..specifically designed  for the Tigre G-IV B5
• …………….maximum RPM of 2300 in level flight at full throttle,
• …………….Vne of 350 km/hr with throttle at idle – respecting 2300 RPM max.
• EI or GRT 4 cyl. Engine Analyzer,
• oil tank with flop tube for inverted operation,
• no starter or electrical generation system,
• 12v battery for radios and analyzers,

  

Other than the problems documented in this report, both of our engines have been operating very well – one for 320 hours and the other for 200. Many things we thought were important turned out to be insignificant and both of us have done needless maintenance. At 270 hours I took my engine apart to clean, paint and measure things. The Spanish overhauled theirs at 300 hours and I wanted to see if there was a reason for this. The concept of Reliability Centered Maintenance (RCM) is gaining popularity amongst small aircraft operators and I don’t want to overhaul unless needed.

I measured all cylinders and pistons as they would in an overhaul and found almost all within tolerance. I am happy to continue operation.

Therefore this report does not have much to say about cylinders, valve guides and seats, pistons, piston rings, connecting rods, bearings, crankshaft (other that the fact it is heavy - 45 pounds or 20.5 kg), camshaft or the oil pump. We haven’t had problems with these components and don’t want to present data that may be incomplete or wrong.

There’s still a lot of other stuff to talk about!!

ENGINE MOUNTS

There are at least 2 different types of engine mount installations on the Tigre. One type employs a large steel washer and two rather large rubbers with holes that are mounted over each of the 4 studs on the crankcase (see figures 1 and 2). The other uses two small rubbers that look like donuts and are compressed to a total thickness of 28 mm. This  system uses a different type of stud (see figures 3 and 3a).

Both systems have been used on our engines and neither is very good at smoothing out engine vibration to the airframe. It seems that an inverted in-line four-cylinder engine with the propeller positioned for easy hand propping is going to shake the plane no matter what we do with the engine mount rubbers.

Fig. 1.    The large rubber engine mounts with special washer, stud and aluminum back plate.

Our experience has been that as the rubber ages we experience metal-to-metal contact. In other words, the engine mount stud contacts the engine mount cavity because the rubber fails to isolate the two components. Another factor seems to be aerobatics. The more aerobatics are flown the sooner the plane starts to feel the engine vibration. Neither of the two systems seems to have an advantage in this respect. We may experiment by increasing the diameter of the hole in the engine mount so that there is a greater distance between the mount and the stud.

Most components on these engines are high quality – in manufacture, material and finish. However, quality control for the manufacture of the engine mount studs and nuts is somewhat lacking. Figure 3 shows two studs with the same part number, and eight different types of nuts that we have found on these engines! The cotter pin holes on some of the nuts are located in such a manner as to make proper tensioning of the rubbers impossible.

One must be careful to use the correct washers and nuts for each type of mount, and to properly compress the rubber if using the small “donut” type. In this aspect the small rubber donut system has an advantage in that it's easy to confirm the rubber has been compressed to a thickness of 28 mm. (One way is by doing a “dry run” and simply counting the number of threads the nut is turned before the distance between the washer and stud surface is 28 mm.)

We have not found any information on how to compress or install the “big rubber” mounts. The rubber used in these mounts is much softer and probably cannot take much compression.

We are both now using the small “donut” type.

Fig. 2.   Large Rubber mounts.

Fig.3. Engine Mount Studs (both have identical part numbers!) for the small rubber donut mounting system. Also the nuts (8 different types have been seen on these engines!)

 

Fig. 3a  The small rubber “donut type” engine mounts, and special washer. 

PROPELLER

Tigre engines vent the crankcase through the crankshaft at the propeller spinner. There is also a vent in the oil tank. One of our airplanes has a modified venting system that allows all gases to be vented through the spinner - in effect the oil tank is now vented to the crankcase instead of overboard. Figure 4 shows one way of modifying the oil tank to achieve this.

 

Fig. 4.    Oil tank modification allows gasses to vent to the engine crankcase.

By doing this, oil is not forced overboard from the oil tank. The return pressure of the oil from the pump is so high that an oil mist is always escaping from the normal tank vent at velocities up to 70 mph!. By directing the vent to the crankcase – an allowance must also be made for inverted flight venting – one doesn’t have to clean the belly of the plane as often!

              ……HOWEVER…..

By venting all gasses through the spinner, a lot of nasty stuff is being sprayed on the inside of the spinner, and the propeller face, and the prop mount flange, and some of that gunk works it’s way to the back of the prop where it starts to corrode the crankshaft mounting flange! This was discovered after only one year of mostly aerobatic flying when pulling the prop off for the annual inspection.

Modifying the propeller face-plate with a neoprene diaphram that fits tightly over the prop cone nut has eliminated this problem. Figures 5, 6 and 7 show how the face-plate has been machined to allow a diaphram to be sandwiched between the face-plate and an aluminum plate that holds the neoprene in place. This prevents any gas or liquid from getting behind the face-plate.

Fig. 5.    Prop face-plate with machined recess. Neoprene diaphram is in the middle.

 

Fig. 6.    This side goes against the propeller.  Rivets are flush.

Fig. 7.   Installed  - no liquids or gasses get behind the face-plate.

OIL SYSTEM

For the purpose of understanding the oil circulation, we will describe how the oil moves inside a Tigre.

The oil tank has a capacity of 8 litres and a recommended minimum of 6 litres. There is a check valve in the tank outlet that is supposed to prevent oil in the lines and pump from draining back into the tank during long periods of inactivity. Essentially we don't want the oil pump to loose its "prime". This check valve is sometimes installed in the flop tube of a tank that has been modified for inverted operation. The tank supplies the engine with oil via a hose (recommended MINIMUM internal diameter of one half inch or 13 mm) to the back-center of the Tigre's sump (see Figure 8a). The sump is the unit with the 3 shiny chrome filters at the lower back of the engine.

Fig. 8a. The oil sump at the lower rear of the Tigre – the 3 shiny chrome filter covers are not shown installed in this picture. The 3 temperature probe ports are visible on the right side of the sump.

The oil goes to the centre filter (all 3 filters have metal screens that are removable for cleaning) and then enters a cavity into which a temperature probe can be installed. This is the centre of the 3 ports on the right side of the sump (this access port is normally not used and is plugged). The cavity then supplies the oil to one of the 3 stages in the Tigre oil pump (Tigres set up for inverted flying normally have a 5 stage oil pump; easily identified by 2 external hoses that run from the top of the crankcase directly to the oil pump). The oil pump's 3 stages are fed through 3 holes on the top right side of the sump (see Fig. 8b). The oil that started its journey at the tank now enters the pump's stage that sends oil via an internal gallery to the Cuno oil filter – which, in turn, supplies the oil to all internal parts of the engine. Oil pressure information is sensed at the Cuno oil filter.

After the oil has done its job of cooling and lubricating, some of it falls to the lower back of the engine into the square hole on the top of the sump (see Fig.8b). This cavity includes the top temperature probe port – which is also usually plugged as there is rarely any oil in the vicinity of the probe port, and any temperature information obtained here is erratic. The oil in this cavity enters the feed of the left sump filter (engine left when viewed from the cockpit) and then goes to another cavity that houses the bottom temperature probe access port, and also feeds another stage of the oil pump.

Inside the crankcase there is a tube that runs from the front thrust bearing area to the oil sump. It collects used oil at the front of the engine and sends it through the third (right) filter after which it moves to the pump via the left of the 3 pump inlet holes. Exiting the pump, this oil also returns to the tank.

Fig 8b. The top of the sump, showing the square inlet hole for crankcase return oil, and the 3 oil pump inlet holes at the right. The center pump inlet hole routes oil from oil the tank to the pump that in turn feeds the internal engine parts. The right inlet hole takes oil from the crankcase and sends it through the pump then back to the oil tank. The other big hole (visible above and slightly left of the 3 pump inlet holes) directs oil from the front of the crankcase to the right filter. After this filter the oil moves to the pump through the left hole of the 3 pump inlet holes, and also returns to the tank.

As mentioned, the oil tank can be modified to allow venting into the crankcase. This is not difficult! The vent line is run below the tank and then enters the top stack that has a hose to the top of the engine (Fig. 4). The bottom of the top stack has a one-way valve that won’t allow tank oil out when inverted. The normal main oil feed port becomes the inverted vent and is also run to the top stack. A large boss is welded to one end of the tank and the flop tube is inserted (see Figure 8).

 There are 2 ways of incorporating the main oil feed line one-way valve. This is the valve that prevents oil in the main supply hose from emptying when the engine is not used for a few days. We both put it on the end of the flop tube – using it as the weight to move the tube up and down with g-loading (see Figure 9). One aircraft has done nothing but aerobatics and the other has not done any aerobatics. The aerobatic machine had the flop tube break during flight and this allowed the weight to thrash around inside the tank. We have no idea how long this went on but there are lots of little dimples in the tank where the weight hit the tank with some force!

The good news:  I monitor oil pressure while inverted and never once saw the pressure fluctuate! The main supply boss is in the middle of the tank, and while inverted with no flop tube, the oil feed remained steady. Bad news:  another part had to be made! My flop tube now has little weight and the one-way valve is installed in the big brass fitting in the tank boss (see Figure 9A).  (Note: Actually it is the exact same concept as the non-inverted tank before modifying it for the flop tube.)

Fig. 8.  Main oil supply hose connected to middle of the tank.

 

Fig. 9.   Oil tank flop tube. The one-way valve and end fitting (effectively the weight that moves with g-loading shown at left) broke off and bounced around inside the tank.

 

Fig. 9A   The new flop tube. The one-way valve is now inside the big brass fitting.

The opening for the oil tank dip-stick cannot be made oil tight! Well maybe it could  -  but  there is always oil around it, so we both welded the hole shut and no longer use the dipstick. Our high-tech method of checking the oil level in the tank is to open the refill cap and stick our index finger in. If we can feel oil we fly. If not, we add oil!

Oil consumption for our engines has averaged 1 liter every 3 hours of air time.

Twice in 300 hours of mostly aerobatic flying I have had a big blob of oil come out the spinner after doing some “perfect storm” combination of positive/negative spin/snap  maneuvers – probably only 3 cubic centimeters (a tablespoon) but it made a small mess. The rest of the time there is only a slight oil mist around the propeller area of the nose bowl. Every couple of flights the front seat windshield has to be wiped clean of oil mist.

The tank has a few small rivets that hold stuff together internally. These rivets leak! Pressure testing the tank at low pressure will sometimes not show much of a leak but if you are patient you will find most of these tanks are leaking. Another common leak area is the boss on the top of the tank. Mine was cracked half way around! And at low pressure one could hardly see any bubbles, but when the area was stressed it was apparent. These tiny leaks are responsible for some of the oil mist that always seems to be everywhere inside the engine compartment. Welding the rivets and ensuring no leaks eliminates some of the mess.

The newest type of hose that is available in both metric and imperial sizes is called “Convoluted” Teflon. It is internally ribbed and can withstand high and very low (suction) pressures, as well as temperature extremes. The exterior is steel-braided and it can be bent around a very small radius. The internal diameter of the main supply hose is ½ inch

(12.5 mm) and it can be bent around a 20 cm (8 inch) diameter! There is no shelf or max service life, and a firesleeve can be added - but the firesleeve will cause the minimum-bend radius to increase! This type of hose is highly recommended (See Figure 8).

OIL SYSTEM PRIMER

There seems to be a consensus that a lot of engine wear occurs during start-up when there is often no oil in the bearings because of long periods of inactivity. We have both plumbed a hydraulic quick-disconnect fitting into the oil pressure line that feeds the cockpit gauge. Instead of an elbow at the firewall, I installed a bulkhead “Tee” so that one line goes to the Cuno-type oil filter (where oil pressure is sensed) and another goes to the quick disconnect fitting (see Figure 10). If the engine has not been started for more than 2 days, I prime the oil system with about a fifth of a liter (one cup) of oil. I use a small plastic pressure vessel that is sold at any garden store for applying weed spray.  You connect the sprayer (pressure vessel) to the hydraulic quick-disconnect, manually pump up the pressure, then open a valve to force the oil into the engine oil circuit. When the engine is started there is instant oil pressure, and hopefully less engine wear.

Fig. 10.   Hydraulic “Quick-Disconnect” fitting.

We have not attempted any modifications to the oil filter. The Cuno-type of oil filter is possibly one of the very best kinds of filters, and because we are essentially lazy, nothing else has yet been considered. With oil changes every 20 to 25 hours using only straight 100 mineral oil, we have found no debris in the filter or the 3 oil screens after 300 hours of operation.  Mineral oil keeps the particles in suspension and frequent changes flushes the contaminants. Use of different types oil may change this data.

The housing on the back accessory case that holds the 3 oil screens has, at the right side of the engine, 3 entry ports that allow probes to be inserted for oil temperature information. Every installation we have seen uses the bottom port for the temperature probe. My engine analyzer has oil temp capability so I used the top port for the analyzer and the bottom port for the mechanical (standard) oil temp probe. We have found the information from the top port is erratic and recommend neither of the two ports normally left plugged be used to get valid data. If more information is wanted about the oil cooling process in these engines, we suggest temperature data should be obtained from the oil tank or perhaps at a banjo fitting somewhere convenient in one of the hoses.

One last observation on the topic of oil: The Spanish Air Force used mineral oil exclusively in these engines. In every one we have taken apart there is almost no corrosion! Where the oil covered any part, there seems to be a coating, or residue from the oil that appears to have effectively prevented the part from rusting away. If, however, the engine has been disassembled and the parts “cleaned”, or perhaps the coating has been somehow scratched, etc., there is almost always rust –sometimes to the point where the part is now garbage! So it could be a good thing if that boat anchor in your shop has never been taken apart…

THE MAGNETO

This component causes a lot of the frustration that people who fly behind a Tigre experience. We have discovered a few things that, to our knowledge, have not been printed anywhere.

We will discuss how the spark is made and distributed, and look at ways of improving the reliability of the unit. We will also attempt to explain how to time the Tigre magneto.

Before starting this exercise you should be aware of the 3 different types of magneto that we have found on these engines, and the high number of possible combinations of parts from different mags that can be found in any one unit! From our experience,

the most common type is               MVE 4-1 (left mag) and MVE 4-2 (right mag)

less used is the                             MVE 4-109A (left)   and MVE 4-109B (right)

least seen                                     MVE 41-1N (left)   and  MVE 41-2N (right)

There are two completely different types of impulse assemblies. There are differences in the timing retard mechanisms. The 109A and 109B versions have springs that pull the weights towards the centre stops when the engine is at rest.  We have reports that the MVE 41 maximum retard is 61 degrees (!), while our MVE 4 max is 42 degrees and the -109A/B units have 47 degrees. These values may be based on mags that were already incorrectly assembled when we started to analyze them.

Fig. 11.  Common magneto parts. The square cam that operates the points, capacitor, points mounted on the rotatable point plate, and bottom right is the rotor. These parts appear to be common and will work on all the different types we have seen so far.

Much of the information in this section has been gathered with the MVE 4-1 and -2 type. With this type alone there are many variations of parts and dimensions so you must be careful when assembling parts. If something is not going together easily, it could be because one of the parts actually has different dimensions and cannot fit! Even if the part does fit, it may not be big enough! Measure everything and satisfy yourself the parts are supposed to fit before forcing them together.

(Keep in mind here that just because the writing on the mag  SAYS  it is a -109A type  or a 4-2 type…. does not mean everything inside will actually  BE  a -109A type or a 4-2 type!....)

One big precaution before we start playing with a magneto:

DO NOT ROTATE A MAGNETO ARMATURE UNLESS THE “P” LEAD IS GROUNDED.

Magnetos generate electricity. Not enough to harm you but it will get your attention. But that’s not the reason for grounding them. ANY time you rotate the magnet you are generating electricity that has to go somewhere. If it has nowhere to go it will start to destroy the coil. So, make a jumper wire with alligator clips that will stay on the unit as you work with it. These mags have a feature that is designed to ground the mag when the “P” lead is removed from the mag. Our experience is that it sometimes works, but most of the time the mags stays “live” (ungrounded), so do not trust this feature.

When using a magneto timing device, you remove the jumper because the timing device will automatically ground the mag for you.

THE MAGNET

A rotating magnet can induce electricity in a coil of wire placed close by. The magnets in the Tigre mags are all over 50 years old and may not be as strong as they once were. There are apparently shops where one can take an old magneto to be “re-charged”. I have not yet done this, as it seems all my magnetos still offer resistance when manually turned – an indication that the magnet still has some strength.

THE COIL

There are many ways this part can fail or not work well. Perhaps the most abuse one can give a magneto is by turning it with no load connected to take the energy the unit produces. If the mag is not grounded, it will develop a high voltage, and the energy developed WILL go somewhere and do damage! The original coils are 50 years old and even if they have never been abused the materials are deteriorating with age and heat cycles.  This deterioration will cause a loss of insulation, which then allows arcing and burn through  - shorting the windings. The coil can also be damaged if the contact points or capacitor are bad. (We have not found a replacement capacitor that will fit in the existing housing but experiments are on going.)

The coil is a weak link but there is hope. I have been running my left mag for a while now with a Bendix coil (see Figure 11a). The part number is IO 357165.  The spark is VERY strong – much stronger than any Tigre coil has given. By simply hand tuning (no impulse engaged) the spark will jump almost 1 cm (3/8 of an inch)!

Some work may be required to modify the Bendix coil so that it will fit in the available space, and a high-tension lead must be soldered to the output tab, but the work is possible in most home shops. A drill press is helpful for machining the slot that allows the coil to be mounted and affixed (See Figure 12).

Fig. 11a.  The Bendix coil will work in all 3 types of magnetos. Shown above is an MVE4-109A magneto. (The impulse mechanism on the mag pictured above is normally found on the  MVE 4-1 or 2, and MVE 41-1 or 2 type mags.) Note the upside-down “U” bracket underneath the hold-down screw that goes over the core’s laminations to hold them together.

The cavity into which the coil fits (49.5mm or 1.950”) is a little bit smaller than the length of the red epoxy section of the coil (50.5mm or 1.990”). However, there are manufacturing irregularities in the actual length (Bendix coils may have been made by different sub-contractors over the years) and if you are lucky, you may find a coil that will fit without taking some material away! If the one you get is too big, about ½ mm (.020”) must be shaved off the red epoxy coating at each end of the coil in order for the coil to sit properly on the magneto core frame. A Dremel tool was used for this.

Drill a 1.0mm dia. hole (.040”) through the high voltage output tab and solder the centre wire of a high tension spark plug wire to the tab affixing it through the hole (make sure any shielding is removed). Add a few layers of heat shrink tubing to build up the outside diameter of the wire.

Fig. 12.  The slots are simply drilled with a .150”  (3.8 mm) drill.  A “C” clamp should be used to hold the core  laminations together while drilling. Drill once per end – the excess will fall away after the “C” clamp is released.

Fig. 13. High-tension wire soldered to tab, covered with heat shrink tubing and run forward with brass end cap.
This cap could be similar to the little “can” made to go on the ends of the Slick wires described later and shown in figure 18a.  Make the “can” a bit shorter (about 6mm) for use on the coil – to prevent arcing to ground.

Ensure it will feed through the hole in the frame and be long enough to touch the distributor cap connection point. Add some sort of a brass end cap to make a better electrical connection. See Figure 13 and Figure 19

Special “U” brackets were made to hold the laminations together and form a washer underneath the coil hold-down screws.

INTERNAL TIMING

The internal timing ensures the points open where we will get the strongest spark. This can be done on the bench in the shop. There is a little red index mark by the condenser (capacitor) and the points should open when the rotor is pointing to the red mark. The plate that the points are mounted on can be rotated. This magneto is very sensitive to the point gap: it must be 0.4 to 0.5 mm (.016 to .020”). If it is more than 0.5 mm there can be a lot of erratic mis-firing. Changing the gap changes the position where the points open as does rotating the point mounting plate – which is periodically done as points wear.  This sounds a lot more complicated than it actually is.  

Fig. 13a.  Point plate showing pivot arm at right, fixed screw at left and concentric screw with the dab of paint on it. Also visible is the square cam that drives the points and the felt pad that lubricates the cam at the top right.

Lets assume the points and capacitor are serviceable, and the mag is assembled correctly.

Start by getting the point assembly plate centered (the retaining screws centered – not at either end of their slots). This is shown in Figure 13a. The point assembly pivots on an arm that needs a small drop of thin oil every year, is fixed at the other end with a slotted screw, and is “gapped” by turning the other slotted screw located in a concentric hole. The points are opened and closed by a square cam on the main shaft. This cam has a felt pad pressed up against it for lubrication and also needs a small drop of thin oil every year.

Rotate the main shaft by turning the rotor (do not try to turn the shaft with the coupling gear at the other end of the mag – there are timing retard mechanisms that affect the main shaft rotation and these will be discussed later) until it appears the points are at their widest opening with the rotor pointing close to the red mark (see Figure 13b). Rotate the magneto rotor by turning it the way it will turn during engine operation. This is CLOCKWISE when viewed from the front of the airplane looking aft. There can be confusion here because the propeller turns COUNTER-CLOCKWISE when viewed from the front of the plane looking aft.

Fig. 13b.  The rotor pointing to the red mark (sometimes not red!) on the edge. This is an “aft-facing” view (if the mag were on the plane it is being seen as if you are standing in front of the plane looking aft).
The rotor rotates clockwise, and the goal of internal timing is to have the points open when the rotor is in this position – aimed at the index (red mark). NOT BEFORE!

Loosen the fixed screw and rotate the concentric screw to get a gap of 0.4 to 0.5mm (0.016 to 0.020”). The rotor should be moved back and forth to get the maximum gap and then measure by using a feeler gauge (see Figure 13c), then re-tighten the fixed screw. Once you have the desired gap, you may have to loosen the entire point plate assembly (2 slotted screws on the big, round plate’s edge – see Figure 13a) and rotate it a bit to get the desirable gap with the rotor pointing at the red mark when the points open. This often involves lots of “tweaking”.

Fig. 13c. Using a feeler gauge to set the point gap.  A wire-type gauge is usually easier to use.

Hook up a magneto timing device to the mag (one lead to ground and the other to the “P” lead) and slowly turn the rotor (in the normal direction) to see when the points open (see Figure 13d). The sound from the timer will change as the points open.  Some magneto-timing units have lights that turn on, or turn off as the points open. Make sure you know which way your tester works!

At the instant this happens the rotor should be pointing at the red mark or just slightly past it. Tighten the screw then re-check the gap and point opening. The Tigre magneto taunts us mischievously by almost always changing its set position when the screws and nuts are re-tightened!

You may find the shaft “locks” while rotating the rotor – this is good! It means the impulse mechanism is engaging (more on this later). To disengage it, just turn the mag upside down, turn the rotor backwards a bit, then continue.

You now have the points opening at the right time and the point gap set. Make sure all screws are tight.

Fig 13d. Magneto timer hooked up to the mag; finding the point opening position.

TIMING RETARD MECHANISMS

The typical Tigre slow idle can be very rough and often sounds as if the engine will soon die! There are several possible causes of this – one of which is timing.

On the G-IV B5 the right mag is timed to 38 degrees BEFORE top dead centre (TDC) and the left mag is timed to 34 degrees BEFORE TDC. These values will not allow the engine to run smoothly at low (idle) RPM. Therefore, both left and right mags have timing retard devices. The armature and rotor are driven by a device that retards timing at low RPM. This apparatus can retard the timing by up to 61°!!

Figure 14 shows the originally installed phenolic bushing in the retard mechanism of an MVE 4-1 type, which is in the rear part of the magneto. Typically, there are 2 weights that move to the centre at low RPM - which retards the timing; and move outward as RPM increases - which removes the timing retard. On the MVE-4 this retard is typically 42 degrees of crankshaft rotation after normal point opening. We had thought the retard was a “sort-of” linear reduction of advance, but with very accurate computer modeling have discovered that this design only allows for full retard or no retard. This has been proven by experience – when VERY slowly advancing the throttle there will be an RPM (usually around 700 to 800 RPM) where the engine will suddenly speed up by about 150 RPM. The weights have overcome the force that was holding them in, and they move outward thus removing the retard.

Fig. 14.  Small, 15 mm diameter, originally installed phenolic bushing.  MVE 4-1.  type mag.

While idling, with no modification to the timing retard mechanism, the right mag will be firing at 4 degrees AFTER TDC and the left mag will be firing at 8 degrees AFTER TDC. The driving cam is not symmetrically square so there is up to 2 more degrees of variation. Even with these retard devices the idle is often rough. Although it sounds kind of neat, it is not good for the rubber couplers that connect the magneto drive gears to the magneto. It can also shake the plane quite a bit.

Bucker owners from about 30 years ago discovered that the retard device has no effect once the RPM is around 1000. A typical idle is 500 RPM, and they found if the maximum retard was limited to 36 degrees the slow idle would be smoother. There is the possibility that the people who discovered this were solving a problem (the rough idle) that actually had other causes (a manifold leak, weak coil, weak valve springs, bad spark plugs, etc…). Before we knew much about the engine we decided to operate the engine with timing retard restrictors. Although it is probably not necessary, both our engines are operating very well with the restrictors.

Restricting the retard can be done several ways. We have personally used 2. They both limit the distance the weights can move toward the centre on the MVE 4 and some MVE 41 mags.

The simplest way is to pull out the old phenolic bushing (Figure 14), make a new one copying all dimensions exactly except the top. Make it with a bigger diameter – 25.4 mm or one inch (see Figure 14a). It’s not pretty, but it limits the retard to 36 degrees and at 500 RPM the Tigre purrs!

Quality control issues again rear their head: every bushing I have removed from original mags has slightly different diameters, lengths and hole dimensions so you must copy each mag bushing EXACTLY!

Fig. 14a.    25.4 mm (one inch) diameter phenolic bushing.

After 270 hours I inspected the 25mm bushings I had made, and found a bit of phenolic residue in the area, and the bushings were worn just a bit at the points of contact with the weights. I just cleaned them up and rotated the bushings a bit! Should be good for another couple of hundred hours…

The other way we have tried is much more elegant – but of course is more difficult too! The restrictor in this case is a plastic block machined to 17.25mm  (0.680”) wide so that the weights contact a long surface – not just one point (see Figure 14b).

The magneto must undergo some disassembly in order to mount this restrictor. The nut that secures the armature pick-up drive (see Figure 14c) must be removed and the plastic bushing installed. However, there won’t be residue and it will never wear out!

Some things to look for if you have this unit apart. The driving pins on the weights can be 9 or 10 mm in diameter, and the 9 mm pin will assemble with the slots for a 10 mm pin but the retard mechanism’s operation may be erratic. Use a 9 or 10 mm drill bit to see what size the slots are.

Fig. 14b.  The most elegant solution! This plastic piece is actually installed on thearmature. It is shown here for information only.

Fig 14c. Showing the armature drive pick-up.

The pins on the weights in Fig 14 go in the slots on this pick-up. This directly drives the rotating magnet, cam and rotor on the other (front) end of the mag.   CAUTION: The diameter of the pins that go into the slots on the pick-up can be 9mm or 10mm. Trying to force a 10mm diameter pin into a 9mm diameter slot will not work, but the reverse – a 9mm pin into a 10mm slot will enable the parts to be assembled but may cause other problems.

The weights pivot on pins that have washers and a cotter pin. The cotter pins can interfere with the armature drive’s slots (see Figure 15). This interference may change the total retard available but will probably not affect engine operation. We found one set of weights with the drive pins not fully seated. It looked as though someone discovered the interference with the cotter pins and raised the pins. The measured distance between the armature drive and the top of the mag is 40 mm. With the pins raised, there was now only 39.5 mm available (see Figure 15a).

Fig. 15.  Cotter pins can interfere with the armature drive pick-up.

Sometimes these weights have been taken apart , cleaned and re-assembled with too many or not enough washers. The weights must be free to move but not so free as to wobble on the pivot. They also seem to work well with no lubrication.

The units that have 9 mm drive pins on the weights have numbers from 1 to 9 and an “F” or an “E” stamped on them. They all look the same size (we have not attempted to measure them accurately), and they all seem to be about the same weight and material. The container they sit in also has the same number and an “F” or “E”. So does the armature drive pick-up, the rotor bulkhead (behind the armature drive pick-up) and 2 components inside the impulse assembly. Presumably, if one thing has “5 o F “ stamped on it, the other pieces should have the same… (however this has never been seen in practice – they are all hybrids!)

The units that have “12 o F” stamped on them have 10 mm drive pins.

Fig. 15a. Raised drive pins (possibly done to reduce the effects of the weight’s pivot cotter pins) don’t leave enough clearance, and will cause the armature to bind. The retard may not work.

One mag we took apart seems to have new manufacture fly-weights and container – no identifying marks stamped on any piece – and the washers and cotter pins are eliminated! Perhaps someone saw the interference caused by the cotter pins and decided to re-invent the wheel. It looks as though it may have worked but the parts do not look as though they have seen much use (see Figure 15b). We are unable to comment on this set-up.

Figure 16 shows the retard mechanism for the -109 A and B units. This design uses small springs underneath the flyweights to always return the unit to the max retard position (see Figure 16a). This seems to be a better design for the retard unit and, if given the choice, I would use this type. It is also much easier to install on the engine, as the rotor does not keep ‘wandering around’ while you try to get it in position. This unit gives 47 degrees of retard on my left mag and I have not attempted to restrict its range. It’s possible that for very slow RPM – less than 400 – this much retard is necessary for smooth operation. After a flight, I can select carb heat on and mixture full rich and the engine will idle smoothly at 310 RPM on the left mag only!

Figures 16b and 16c show a completely different type of timing retard mechanism that can give up to 60 degrees of timing retard. We have only seen this in one of about 20 mags we have taken apart, but may in fact be applicable to the MVE 41-1N and 2N types if they are still in “new-out-of-factory” condition (meaning they have not been tampered with).

Fig. 15b. Weights installed with no cotter pins or washers. History unknown. This unit is attached to an impulse normally seen only on the -109A/B mags.

Fig. 16. The MVE109B timing retard mechanism. Note the different shape of the weights and armature drive pick-up.

Fig. 16a. MVE109B  The springs pull the weights back to full retard position.

Also shown on the right is the impulse mechanism that is sometimes (but not always!) on the -109A/B mags. It’s a good idea to be able to tell the difference between the 2 types of impulses so that you can predict the angle between engagement and release.

Fig. 16b. MVE 41-2N timing retard mechanism – possibly 60 degrees of timing retard!

Fig. 16c.  MVE 41-2N flyweights. Note the different design and drive pin diameters.

MAGNETO TIMING

The magneto must now be installed and timed to the engine.

The first step is to find Top Dead Centre (TDC) for Cylinder #1. This is a commonly used procedure and is explained on several websites. John Schwaner’s “Sacramento Air Ranch” website and parts supplier is an excellent source of many things aeronautical. Included in this report (for the next 4 pages) is a method of finding TDC using an electronic tool called the RITE Magneto Timing System.

Timing a Tigre mag (and just about any other aircraft engine) just got a whole lot easier!

This is a VERY nice unit. The yellow, plastic “dome holder” is attached over the prop spinner with 2 rubber bands that can be tightened. There’s no need to raise the tail on a taildragger.

I just “eyeballed” the approximate centerline/axis of rotation and attached the bands, but found they could not hold the weight of the entire unit, so I added some good quality masking tape to stabilize it. The electronic part of the entire device is a 2 inch cube (approximate size) and it is now inserted in the front of the “dome holder”. It is held there with an internal magnet but does move around a bit so I jammed a piece of rolled up tape on the edge to stop the movement. When turned on, it remembers it’s angular position from when it was last turned off.  To “re-zero” it just push the “Zero” button!

The new “RITE System Magneto Timing Kit” - sold by Aircraft Spruce – taped to the spinner.

Turning the prop is much easier than turning a small timing disk when timing. Accuracy to a tenth of a degree is possible (but not needed for the Tigre!). Most importantly, you don’t have to remove the propeller to time the engine! The RITE kit can also be used for determining flight control deflection angles, and includes a separate holder to hold the electronic part while the device is on a flap, etc.

Like electronic torque wrenches and GPS, we will all soon wonder how we ever got by without this amazing tool!

DETERMINING TDC

I made this Top Dead Center (TDC) indicator (picture next page) out of a thin strip of aluminum. The piece narrows so that it can pivot on the internal thread of the cylinder’s spark plug hole. The left side of the indicator is in the cylinder, and is pushed down by the piston as it moves toward TDC. If this piece “wanders” while moving, the measurements will be bad. I drilled a 1.5 mm (1/16”) hole where it pivots then screwed a self-tapping #4 screw into the hole. This pushes out some material on the bottom of the hole, which will then help prevent the piece from wandering around while it is moving up and down. I did not leave the screw in because it could fall into the cylinder.

The first step is to take out the front spark plug in each cylinder and rotate the prop until the piston in Cylinder #1 is visible and moving toward Top Dead Center (TDC) with both valves closed.

1. Tape a piece of white cardboard to the pushrod tube and the nose bowl (see picture).
2. Attach the RITE system dome holder to the spinner and tape it so it doesn’t move, and also make sure the electronic unit doesn’t move inside its cavity.
3. Move the prop to where you think it might be within a few degrees of TDC and “Zero” the unit, and insert the TDC indicator into the spark plug hole of Cyl. #1.
4. Move the prop backwards (opposite normal direction of engine rotation) about50 degrees.
5. Move the prop forward slowly and stop when the unit reads 12.0 degrees plus or minus a degree or two. (If you have to move the prop backwards at all move it 50 degrees and start over - so that the effects of gear backlash are eliminated.) “Zero” the unit again and make a thin pencil mark on the cardboard where the TDC indicator is pointing.
6. Continue rotating the prop and stop when the TDC pointer is pointing to the mark you just made. The indicator will have moved up then down while you were moving the prop. Read what the unit indicates. Whatever number is on the unit now is the total number of degrees the crankshaft has rotated since you made the mark. Divide that number by 2. Lets say the result was 23. Divide that number by 2  =  11.5
7. Move the prop backwards about 50 degrees, then move it forward – past the zero reading - and stop when it reads 11.5
8. That is TDC!   “Zero” the unit again, and you are ready to go!

An approximate TDC position is estimated and the TDC indicator is placed in Cyl. #1.                       

The RITE unit is “zeroed”.

The prop is backed up 50 degrees, then moved slowly forward until the unit reads 12 degrees – plus or minus a few. A mark or line is drawn on the cardboard, and the unit “zeroed” again.

The prop is moved forward until the TDC indicator points to the mark or line that was just drawn. The unit now reads the total number of degrees the prop just rotated. Divide this number by 2. If the result is, let’s say 11.5, move the prop back 50 degrees or so then move it forward so that you go by zero and stop at 11.5. This is TDC for Cyl. #1. Now “zero” the unit again and you can start timing.

TIMING THE MAGNETO TO THE ENGINE

If mounted for easy hand starting, the prop will be vertical when cylinder #1 is at TDC. Unfortunately, the prop has to come off for most timing disks to be used.

It is important to determine TDC accurately because of GIGO (Garbage In Garbage Out)! Having one spark plug out of each cylinder makes the process much easier and more accurate.

For this procedure, and ALL subsequent movement of the timing disk (or prop if installed), the effects of gear backlash must be eliminated by ONLY turning the engine in the normal direction of rotation (counter-clockwise when viewed from standing in front of the plane). If it must go backwards even a bit, you must move it backwards about 50 degrees, then start to move forward again.

The goal is to have the points just opening at 34 or 38 degrees before TDC, with the 3 magneto attach nuts in the centre of their slots. Your patience will be tested here as it can sometimes take a while to get these results. The tolerance of the magneto-mounting studs in the slots is obviously from one end to the other, but subsequent measuring and timings will mean you have to eventually take it off and get it where it was supposed to be in the first place!

For the left mag (Number 1) using an MVE 4 or 41 type retard mechanism, install the timing device and position Cylinder #1 at 34 degrees before TDC. If the tail of the airplane is low you can insert the rubber coupler into the mag cavity and it should stay in position while you install the mag. Ensure any gaskets and “O” rings are on. There may be quite a bit of fiddling required, so make sure a jumper is installed between the “P” lead and ground.

When placing the mag into the cavity you must hold the rotor in the advance position (this is done by holding the magneto’s driving gear at the back and then rotating the rotor clockwise when viewed from the front of the mag). Make sure the impulse is not engaged. The rotor should be pointing at the red mark with the points just about to open with the rotor in the advance position. It often takes several attempts to get it in the correct position! Once the mag is in position, you can hand-tighten the 3 nuts then check to see that the points are in fact opening close to 34 degrees before TDC.

This is done by hooking up the magneto-timing device and slowly rotating the disk/propeller, while at the same time turning the rotor (clockwise when viewed from the front) by hand to ensure the mag stays in the advance mode. This is perhaps the most mis-understood aspect of timing the Tigre mags. The internal timing retard unit must be overridden to correctly time the mag. That means you have to “assist” the rotor’s rotation to eliminate the timing retard unit’s effect. When viewed from the front of the plane, the prop/timing disk is being moved counter-clockwise and you are turning the mag rotors clockwise. You must apply a constant pressure to the rotor. Moving the rotor back and forth will just screw things up!

When the sound changes the points have opened. If it’s close to 34 you can carry on, but if it’s nowhere near 34 degrees you have to take the mag off and perhaps rotate the rubber coupler one notch then re-install. Try to hold the rotor steady so that it doesn’t move while putting the mag on. Again, make sure the impulse is not engaged. This usually takes a few attempts and it seems like you never have enough hands!

The -109A/B units have a spring that returns the weights to the retard position, so it is easier to install this type with the prop at about 13 degrees AFTER TDC  -  then check to see if the advance position is 34 degrees BEFORE TDC.

Once you have the points opening within a few degrees of 34 you can start to “fine tune” the mag to get it exactly at 34 degrees. Rotating the mag within the slot is the way this is done. Rotating the mag clockwise (looking at the rotor facing aft) retards the point opening. Counterclockwise rotation of the mag increases the advance. Every time you tighten the 3 nuts the timing will change! Sometimes as much as 2 degrees! This is where a lot of trial and error happens, but eventually the points will open at 34 degrees with the 3 nuts tightened. The correct torque for these nuts is 12 to 15 Nm or 9 to 11 ft.lb.

Remember that every time you take a measurement you must be applying clockwise pressure on the rotor and the impulse must not be engaged.

The same procedure for installation and timing is used for the right mag except the angle will be 38 degrees.

If you try to rotate the rotor the opposite way while still turning the propeller (or timing disk) in the normal direction of engine rotation, you will be retarding the spark by the amount the mag can retard – up to 61 degrees for the MVE 41-1/2N, 42 degrees for the MVE 4-1/2 or 47degrees for the MVE 4-109A/B ; 36 degrees if a restrictor similar to ours has been used. One would only do this in order to see how much timing retard each mag has.

Fig. 17.  The Timing Disk with TDC pointer taped to the nose bowl. 

Do the timing only on Cylinder #1 because, as was also earlier-mentioned, the “square” cam that opens and closes the points is not square. You WILL end up with slightly different numbers for EVERY cylinder!!!  On each mag!!!

After you have finished it is a good idea to get some data. Take time to accurately see what each mag is doing for each cylinder. Remember to hold the rotor in the advance position for all readings. When the impulses engage you must keep moving forward and let them release, go backwards about 50 degrees (to eliminate gear backlash) then move forward again to get readings for each cylinder.

If the readings are not good, you may have to start all over with a new reference value. I had to “adjust” the reference to 37 in the right mag and 35.5 in the left mag so that the other cylinders did not exceed limits. If I had timed the right mag cyl. #1 to 38, then cyl. #3 would have fired at 39 degrees before TDC – out of limits. And if I had timed the left mag cyl. #1 to 34, then cyl. #2 would fire at 31 degrees before TDC or about 3 degrees later than it should.

Here’s a chart of what my mags are actually doing now. I have just used the new electronic timing unit that Aircraft Spruce sells (the RITE Magneto Timing System).

 

Firing Order                         Right Mag                      Left Mag                          Difference

Cylinder # 1                               37                                    35.5                                    1.5

Cylinder # 3                               38                                    34                                        4                                        

Cylinder # 4                               36.5                                34                                         2.5  

Cylinder # 2                               37.5                                32.5                                      5

        Max. retard angle                    36                                    46 (no restrictor installed)

Perhaps because of the low compression ratio and Max RPM of 2300 these variations are not significant.

IMPULSE MECHANISM

An impulse mechanism is needed to generate a spark strong enough to start the engine when there is very little rotational speed. The Tigre has 2 different types - both using a spring steel coil that engages at speeds below approximately 250 RPM. In the type normally found on the MVE 4 and MVE 41 mags there are two little dime-sized (about 1.3 mm. diameter) weights (see Figure 17a - 1) that rattle around inside the mag when you are hand propping the engine. If you can’t hear them rattling around there is a good chance they will be dirty and cannot engage the impulse - which means the associated plug(s) will not fire. These little weights are made with at least 2 different diameters. The limited experimenting we have done indicates all units will work with either diameter weights installed.

The other type is mostly seen on the -109A and -109B mags (See Figure 17a – 2) and these units do not make any noise. 

Fig. 17a - 1.The inside of the MVE 4/41 impulse mechanism showing the 2 dime-sized weights

 

 

Fig. 17a – 2. The inside of the -109A/B impulse, and the  engage/release device with spring.

At slow speeds they will engage every 180 degrees of magneto rotation. This means that each mag can only use the impulse for two of the four cylinders. Therefore the right mag is different from the left mag, and both are needed in order to have an impulse-generated spark available in all 4 cylinders.

It is quite likely that EVERY Tigre magneto you will ever see has been taken apart by at least one person who has absolutely no idea what he is doing, and been re-assembled incorrectly. You may get lucky and have a mag that works, but by the time they make it to your shop a lot of people have played with these units.

Some of the problems you may find are:  the wrong impulse mechanism has been installed in the mag; the unit has wrong orientation; the release point is 60 or 120 degrees out of sync with itself; and the many ways of screwing up the re-assembly if the impulse spring assembly has been taken apart. There are also issues concerning the manufacture of different impulses for different engines and applications that may have somehow found their way into our Tigres but were never meant to be in a Tigre! They just happened to fit.

When I started trying understand the engine I put a big pulley on the prop mounting flange and drove the Tigre with a 2 HP electric motor using v-belts and reduction gearing so that the engine would turn at various speeds from 50 RPM up to a max of 250 RPM. This allowed confirmation that fuel and oil were going where they should, and valves were opening and closing properly, but more importantly it enabled a look at the timing. Because the engine had supposedly been overhauled by the Spanish Air Force, I assumed it would run - so all I would be doing was seeing how the magnetos and impulses were working, etc.

Fig. 17 b. The impulse assembly is at left (this is an MVE4-1 mag), next is the retard mechanism, magnet, cam and rotor.

I discovered the impulse mechanisms would still occasionally engage at 250 RPM. I also discovered that the impulses were firing at about 23 degrees after TDC on one mag and 57 degrees after TDC on the other (these were strobe light values – not too accurate). I had no idea what the correct value was and nobody I asked could give me a reasonable answer. And at that time I had no idea how screwed up a mag could be.  My gut feeling was that a cylinder charge would not burn very well if any of the valves were open. I operated for a long time not knowing that only one of my mags was capable of starting the Tigre – the other impulse was way out.

So, how do you check the impulse? If the mags are not on the engine, you can place the mag in the exact orientation it will be on the plane and see that the impulse engages (stops rotating the rotor) as the drive gear is turned. This should occur with the rotor pointing close to the high voltage wire that comes out of the coil on the left mag (see Figure 17c), and again 180 degrees later. The right mag will be up side down compared to the left, and the impulse should engage when the rotor is pointing close to the capacitor wire (see Figure 17d) and again180 degrees later. These indications are with the advance/retard mechanism in the full retard position (do not assist the rotation as you do when timing the mag to the engine).

 

Fig. 17c. The left mag (MVE4-109A) rotor has stopped in this position as the impulse mechanism engaged for Cyl. # 1. The prop was 64 degrees before TDC. The impulse released at TDC. (Total: 64 degrees of prop rotation – 32 degrees of mag rotation)

For all remaining tests and calibration the units must be on the engine. The Tigre manual states that in order to have the impulses work properly the mags must be installed in such a way that a long straight edge can be placed on top of both mags and the coil housings will be close to being parallel to the ground.

Fig. 17d.  The right mag (MVE4-2 type impulse) rotor stopped here at Cyl. #3 impulse engagement. Prop position was 53 degrees before TDC. The impulse released at 23degrees after TDC. (Total: 76 degrees of prop rotation – 38 degrees of mag rotation)

Our experience indicates both types of impulse mechanisms will engage and release properly over the entire range of magneto rotation – moving the mag from one end of the slot to the other. This will change the timing: with my mag correctly timed to 34 degrees before TDC, I rotated the mag clockwise (viewed looking aft) to the end of the slot, which moved timing to 24 degrees before TDC and changed the point of impulse release to 13 degrees after TDC. Rotating the mag the other way to its limit moved timing to 40 degrees before TDC and the impulse release position to 4 degrees BEFORE TDC.

The statement in the manual could be there to avoid positioning the impulse release position BEFORE TDC.

You can change the impulse release position a few degrees by rotating the point assembly plate and re-setting the point gap (a total of 10 degrees of impulse release position change is available – approximately +/- 5 from the plate's centre position). This will move the position where the points open. Unfortunately, we have not been able to figure out a way of accurately determining where the impulse will release while the unit is off the engine. It must be installed AND timed reasonably well before you can see what the impulse is doing.

There are only a few ways that we know of to move the point at which the impulse fires: changing mags or changing and intermixing impulse assemblies are 2 possibilities. I have considered – but not yet done – drilling 6 new holes in the mounting plate of the impulse assembly in order to rotate the assembly into a better position. Also see the section "Magneto Impulse Surgery" at the end of this section.

We have noticed the problem of poor impulse release position occurs mostly on the left mag. This could simply be coincidence.

It seems the magnetos made for the right side almost always fire in the 20 to 25 degrees after TDC range - regardless of the type of impulse installed. When taking apart these mags, it appears that the armature drive pick-up is positioned on the main shaft such that it is pointing the same way the rotor points – they are both kind-of “in line” (see Figure 17 g). When looking at the insides of left mags, the rotor points in a different direction and this direction is not common among the units we have taken apart.

Once the left mag has been set to 34 and the right to 38 degrees before TDC (on cylinder #1), you can rotate the prop/timing wheel (and for this measurement there is no need to “help” the rotor – keep your hands away!) and note when the impulses fire - do it for all 4 cylinders.  We have found that the values should be between TDC (0 degrees) and 25 degrees after TDC for the engine to start easily with hand propping. I have no idea how an electric starter would influence this data or change the operating envelope.

The left mag impulse is supposed to fire cylinders 1 and 4, the right fires 2 and 3. We have seen installations where this is reversed, and engine starting/operation is normal.

My engine now has the left impulse (MVE4-109A) firing at 0 degrees for cylinder #1 and 5 degrees after TDC for #4, while the right impulse (MVE 109B – there is no “4” engraved after the “MVE” on this one and it has the MVE 4-2 type impulse!) fires both #3 and #2 at 22 degrees after TDC.  If primed properly, the engine will fire on the first blade – regardless of which cylinder is up.

Fig. 17g.  Rotor is approximately “in line” with the armature drive pick-up just visible at the top of the picture. This is an MVE4-2 (right) magneto. In the left mags they are not “in line”.

It should also be mentioned that the impulses must not fire BEFORE TDC!  This could cause the prop to spin backwards…

The following is a bit of extra information to help you understand what is going on in the impulse.

The MVE 4 and 41 impulses engage when one of the little steel weights falls (with the help of gravity) into the proper place. If the unit is turned upside down the weights cannot fall into place and therefore the impulse will not engage. That is why the right and left mags are different. The right mag is almost the same as the left mag – just turned up side down with the impulse mechanism rotated 180 degrees on the aft end of the unit (as well as a main shaft/ armature pick-up drive re-alignment).

There are limits as to how far you can rotate the entire magneto and still have the weights engage. The MVE4-2 mag shown in figures 17e and 17f had limits of 15 degrees one way and 85 degrees the other way. This totals 100 degrees, and that total seems to be common – the plus and minus numbers from the vertical may vary with different mags. The point of impulse engagement is exactly the same with respect to the magneto – NOT IT’S ORIENTATION TO GRAVITY - throughout the entire 100 degrees.

Fig. 17e.  MVE4-2 Magneto rotated in a lathe to determine impulse engage limits. 

15.6 degrees one way shown here.

Once engaged, it will stay engaged for 35 to 40 degrees of magneto driving gear rotation (gravity has no effect once the weights have locked in position). That is 70 to 80 degrees of propeller rotation.  Although the inside of the impulse mechanisms all look the same, there are manufacturing tolerances and wear characteristics that give us the variety of engagement and release angles we have seen. . (General rule: MVE4 and 41 impulses stay engaged 35 to 40 degrees. The 109A/B units stay engaged 30 to 35 degrees)

It then releases, and the armature, cam and rotor that stopped moving when the impulse initially engaged are now re-connected and spin quickly. I have seen many different impulse release positions - possibly because so much has been changed, and added to, and mix-matched with various parts that sometimes were never even meant to be on a Tigre.

The 109A/B impulses also need gravity, engage twice per revolution (which means 2 cylinders per mag – the same as the other type) and can only engage when the unit has the proper orientation.  The operating window with this impulse is much bigger – 45 degrees one way and 115 degrees the other  - using the same “mag-in-lathe” technique . The impulse coil is  engaged when a little arm, assisted by gravity, falls and its other end locks in position on the inner edge of the cavity. After a further 30 to 34 degrees of magneto rotation (60 to 68 degrees of propeller rotation), the arm releases and the armature spins forward.

If everything is set properly, this spin will happen sometime between TDC and 25 degrees after TDC (of engine/propeller rotation). The cam will open the points as it turns - generating a high voltage that will be sent to the appropriate spark plug.

Each of the two types has its advantages. The -109A/B units are not as affected by grease or oil in the mechanism – in fact it helps! The little weights in the other type, however, are negatively affected by grease and oil in the cavity – they are prevented from moving freely. With this second type, if you hear the weights moving, the impulse is probably working. It is difficult to know if the -109A/B type’s impulse is working because you can’t hear much or feel anything when propping the engine.

Fig. 17f. 85.9  degrees the other way shown here.  The impulse engaged at all positions between 85.9 and 15.6, and stayed engaged for 38 to 40 degrees of magneto gear rotation before releasing.

Magneto Summary

Essentially, we want the magnetos to allow the engine to start easily, develop rated power and run for a few hundred hours before maintenance is required. So here’s a “to do” list:

Make sure the points and condenser are OK;

Confirm the points open when the rotor is pointing at the red index mark;

Confirm the point gap is 0.4 to 0.5 mm; This should be the first step in any timing operation, or annual maintenance check, as we have found the gap increases with engine time. This usually retards the timing. Reset this gap before timing the mag and you may not have to re-time the mag-to-engine!

Ensure the left mag is timed to 34 (or 36 as required) degrees before TDC, and the right mag is timed to 38 degrees before TDC (with allowances made for irregularities in the cam that drives the points);

Install a Bendix coil, or make sure the original coil is in perfect condition;

Measure and record ALL cylinder firing positions for both mags, as well as impulse engage and release positions;

For ideal impulse engagement we want:  the left (No. 1) mag’s rotor to be pointing at or up to 10 degrees after the coil’s high voltage output lead, and the right (No. 2) mag’s rotor to be pointing at or up to 10 degrees after the condenser’s wire with the timing retard mechanism in the “retard” mode. This should give impulse firing angles between TDC and 25 degrees after TDC.

If impulse release positions are later than 25 degrees after TDC and parts exchange is not an option, consider some magneto surgery…

Magneto Impulse Surgery

We have discovered some small differences in the construction of the impulse assemblies. In the MVE 4 and MVE 41 types, the impulse engage position is determined by a “ridge” in a machined plate (see Fig. 17h) on which the little weights of the impulse rotate (there are other potential factors that can influence engage/release positions but they are presently not that well understood and are beyond the scope of this report). The Spanish manufacturer actually changed the angles for a number of different applications but some have not been useful for the Tigre. We are making some new plates that have an increased angle. This change should allow any mag that is currently firing the impulse between 25 and 50 degrees after TDC to fire 0 to 25 degrees after TDC. Essentially, this modification will cause the impulse to engage 12.5 degrees of MAGNETO rotation (that is 25 degrees of ENGINE rotation) earlier than the one that’s presently installed. This can make hand-propping MUCH easier!

                                                                             WARNING!

ENSURE you have made accurate measurements and understand the entire process, because if your impulse is presently firing at ANY angle LESS than 25 degrees after TDC, making this modification will move the impulse release position to a firing angle BEFORE TDC – this can cause injury or death if the engine fires and moves the propeller backwards!

Fig. 17h. The machined plate on the right is attached to the frame on the left. The impulse  assembly “weights” move around the outside edge of the machined plate. The impulse engages when a weight jams against the rounded edge of the ridge (at top) of the plate.

 

The following explains a way of changing the impulse mechanism’s release angle, and applies only to engines that are started by manually swinging the prop. Electric starters change the rules of impulse release positions. We will also assume that you have been able to operate the engine normally – it’s just starting that is difficult. This process applies ONLY to MVE 4 and MVE41 type mags. The MVE109 mags have an entirely different impulse design.

Before attempting this “fix”, make SURE you actually need it. An engine that has good impulses may be hard to start because of weak valve springs, bad spark plugs, a weak coil, intake manifold leaks, bad ignition wires, etc. Once you are certain the magneto impulse is the problem, proceed as follows

Accurately determine TDC, then time and record the impulse release positions carefully for all four cylinders, and determine which mag impulse fires later than 25. Try to start the engine by using only this mag – leave the other mag off. Remember that with one mag off every 2^nd blade will do nothing. If it starts and stays running, don’t waste you time trying to fix it – go flying! If it fires each time you prop it (that is every second blade because the other mag is off and nothing should happen when propping the other mag’s  associated blade) but the engine doesn’t stay running, the impulse  might not need modification – something else may be causing the engine to die. If you get a small “poof” every 6 or 8 blades and cannot start the engine, then the impulse probably needs work (assuming the coil is OK).

Attach a jumper wire from the “P” lead connection to the case of the mag in order to ground it (prevents coil damage) and remove the mag from the engine. On the rear of the mag there are 6 screws that hold the impulse/retard unit in place. With the mag sitting on the bench in the EXACT orientation it would be on the engine (identifying marks - MVE4 etc. - on the mag facing up), make a small identification mark such as a centre punch dot beside the mounting hole that is at the top. This is important for re-assembly.

 

Fig. 17i.  The 14 mm nut on the magneto’s drive gear secured by the special “star” washer that is held in place by the circlip

Next loosen the 6 small screws that hold the impulse/timing retard mechanism to the magneto about 5 or 6 turns – don’t take them all the way out. Hold the mag upside down and tap the mag edges with a rubber hammer. The impulse/retard assembly will fall out but still be in position on the screws.. There is a VERY thin gasket between the two – try not to break it! Now carefully remove all 6 screws so the gasket is not destroyed and remove the unit from the mag.

In order to take apart the impulse mechanism, remove the circlip around the 14 mm nut on the aft end of the unit (see Figure 17i). This only retains the special star washer that prevents the nut from loosening. Gently pry the washer out using a dental pick. The cavity that holds the retard fly-weights must be held securely in order to loosen the 14 mm nut. I use a 3-jaw chuck in a lathe to prevent distortion. The 14 mm nut is a standard right hand thread and is quite tight. Once it loosens it will turn freely for a few turns then become very tight again. It now will act as a gear puller as you continue loosening the nut. It pulls the impulse away from the rest of the unit. The 2 impulse weights will fall out the side if the unit is not horizontal.

Fig. 17j.  This machined plate has a 9 degree ridge. The one in Fig 17h has an 11 degree ridge.

There is a small ‘key’ that maintains the correct orientation between the two parts on the shaft. Do not loose it! You will now be looking at the machined plate with the ridge as seen in Fig 17j. This plate can be removed by loosening the 3 retaining screws. These screws are locked in position by slightly deforming them with a punch in the slots. You should try to get rid of the deformation before loosening them

The modification is to install a new plate with the ridge rotated from whatever angle is in the existing mag to 23.5 degrees. The magneto that contained the impulse shown in Figure 17j had an impulse release of 45 degrees after TDC. By increasing the ridge from 9 to 23.5 degrees the impulse release will be moved to 16 degrees after TDC.See figures 17 k and 17 l.

Fig. 17k.  The “new” plate with a 23.5 degree release angle

Fig. 17l. This shows a 9 degree, a 11 degree, and a 23.5 degree machined plate. These angles determine where the impulse releases.

Some notes for re-assembly:   The plate must be installed on the frame correctly. Refer to Fig. 17h. The plate has 3 holes for screws. The frame has 4 threaded holes. It is possible to put the two pieces together 90 degrees out of alignment. The proper position is when the rounded edge of the ridge on the plate is very close to one of the 6 mounting holes on the outer edge of the frame - the one that you put the identifying mark beside when you took the unit apart. If it is more in the middle between 2 of those 6 mounting holes it is wrong and the impulse will not work at the right time. The 3 screws are tightened and secured by using a punch to deform their edges in the machined slots. Remove any surface roughness so the weights can slide freely over top of the screws.

When mating the impulse spring unit back to the shaft you must make sure the key is in the keyway on both the shaft and the impulse spring unit. Trying to tighten the 14 mm nut with the key even slightly out of alignment will ruin the slots and/or the key itself. Go very slow and be sure the weights are in their separate, respective cavities before tightening the nut. I don’t know the correct torque for the nut but, if you remember how much force it took to remove it, make it pretty tight - however don’t distort the fly-weight cavity trying to hold it securely. Insert the special “star” washer and circlip.

Before placing the entire unit back into the magneto, hold it with the specially identified mounting screw hole (the one you marked) up and rotate the spring unit with a rubber coupler in the gear teeth. It should stop at two locations per 360 degrees of rotation – this is effectively the two impulse engagements for every rotation of the mag armature. If it doesn’t make those two stops with the marked hole up, there is something wrong inside. One of the stops should be when the impulse spring’s outer visible edge is seen close to the specially identified hole (see Figure 17m).

Fig. 17m. For proper impulse operation, the visible end of the impulse spring will be closely aligned (+/_ 10 mm) with one of the 6 outer edge screw holes (the specially identified one) for one of its two engagement positions. The other engage position is 180 degrees from this one.

The last part of the exercise is to put the impulse/retard unit back into the mag. This is tricky. I start with the mag vertical, and lower the unit into the cavity so that the pins on the flyweights go into the slots on the armature drive pick-up. You will find they don’t want to go into the slots. If you have the approximate alignment correct, hold everything still and rotate the rotor at the other end of the mag just a little bit back and forth. This may allow the pins to fall into the slots.

This may not seem too difficult – but try to do it without damaging that very thin fragile gasket! And you must have it in perfect alignment with the 6 holes so that the screws will go in (and don’t forget to put that specially marked hole at the top). That’s where your patience will be tested. It is best to do a few practice runs first without the gasket in order to see how things can get screwed up!

This may seem like a tremendous effort for such a small (?) problem. But it was definitely worth it.Both of us used to dread the “start ritual” our Tigres seemed to require. Having the impulses firing between TDC and 25 degrees after TDC means if we have done the fuel prime exercise properly (and that is truly a ‘black art’), the engine will almost always fire on the first blade!

DISTRIBUTION

We will look at the wires, distributor cap and spark plugs in this section.

The magneto harness is inside a massive, rigid and heavy conduit. I eliminated that conduit and replaced it with a Slick Magneto harness kit M6201-4B.  (This kit is just  8 long wires with the spark plug ends already installed. One cuts the wires to the proper length and adds a little brass cap to the other end and this is inserted into the distributor cap terminals.) Figure 18 shows the plastic spacer that holds the wires in place as they exit the mag. Timing the mags is much easier with no conduit installed; and if so desired, you can look at one mag without having to take off the other mag’s distributor cap

For pictures showing one method of making a high tension lead terminal end (used on the distributor end of a spark plug wire as shown below in Figure 18a) go to the Bucker website:  “www.bucker.info” select   Photo “Gallery”   and look at “Tigre ignition” pictures.

There is also information on a “thread” on the Bucker Forum “Maintenance” section called: “Slick HT leads and plug spanners”  that was started in 2007 but has seen recent activity.

 

 Fig. 18.  Plastic bushings that allow the high tension leads to enter the mag.

 

            

Fig. 18a. Brass tube soldered to Slick part number M-1798 – electrode screw.

If using this “can” on the coil’s high tension wire make it about 6mm shorter to prevent

arcing to ground

 

Fig. 18b. The Slick M6201-4B Ignition Harness.

The little needle is a Slick part  (Slick M-1798 electrode screw) – your local maintenance shop has thousands of them – and it is silver soldered to a piece of thin wall brass tubing (0.250” or about 6.3 mm with a wall thickness of about 0.010” - .25 mm) you can buy at any hobby store. 

You end up with a tiny “can” that has the electrode screw sticking out of the centre. This is simply screwed into the end of the high-tension wire. Once this “can” is installed on the wire, it is placed in the Tigre magneto distributor cap and is “crushed” by the setscrew. This ensures a good electrical connection. Originally, the setscrew would pierce the wire. With the “can” on the wire end, you only want to dent the can. A small spacer or washer (about 0.060” or 1.5 mm thickness) is added before the setscrew’s lock washer so that the screw doesn’t go in too far.

In the magneto, the shielding from each of the 4 leads is grouped together and grounded inside the mag. Radio transmission and reception is very good using Becker radios. 

It is also possible to take out the old wires and put new ones in that conduit (see Figure 19).  This is more complex and difficult - which is why my partner did it!  He loves a challenge. He used the existing conduit and machined special fitting that are exactly what you would normally see at the mag on a Slick harness. Effectively, the conduit becomes the “magneto” as far as grounding the shielding wire is concerned, and the central core continues through the conduit to the mag. (See Figure 19a)

It looks very authentic and “old”!

 

Fig. 19. The ignition conduit with Slick harness extensions. Looks very authentic.

Fig. 19a. The harness shields (from the spark plugs to the conduit) terminate at these specially made fittings attached to the conduit. You have to make these – they cannot be bought

  

DISTRIBUTOR

So far, all distributor caps we have used have worked fine. It appears that no matter how many different types of these we find, they all work on all the different magnetos. They are well made, and if no scratches appear on them it seems to they may last a very long time (see Figure 19b). When attempting to clean the copper terminals make sure you don’t scratch the very smooth phenolic surfaces. Very high voltage is attracted to a rough surface! It is OK to sand or file the copper but not the phenolic. This is where carbon arcing/tracking will begin and this “arcing” takes energy out of the spark.  Once it starts you can never get rid of it. A new unit will have to be installed. The only thing to do on this part is keep it clean!

Fig. 19b. The Tigre distributor. The coil’s output enters the cap at left.

Figures 19c. and 19d. belowe show what can happen if the high tension wire from the coil slides out of contact with the distributor and allows the voltage to start travelling to the magneto case. This occurred after about 50 hours because the wire soldered to the Bendix coil’s tab was not rigid enough and started to expose the long brass “cap” on the wire’s end. The fix is to make the end “cap” as small as possible so the electricity can ONLY go into the distributor contact; and to make the wire rigid enough to prevent ANY movement.

Fig.19c. The high tension wire from the coil moved out of position.

Fig.19d. The subsequent arcing eventually quit firing all spark plugs - dead mag as far as the pilot was concerned,  but the electricity was still being generated and destroyed the cap.

SPARK PLUGS

The original European spark plugs are, in our opinion, not usable. I was lucky to hear of the “NEW” Champion REM37BY spark plugs as I was finishing my airplane and spent a bit of time talking with the factory. The first production runs gave plugs that were slightly longer than the final version, but I was lucky enough to get the short one right from the start, and I couldn’t believe other people were saying the plugs were too long!

These plugs are fantastic. They were developed to fight lead fouling and they do that in the Tigre, but mostly – THEY FIT! And they are much more reliable. Using 100LL and TCP with constant leaning, they simply do not get dirty! At 50-hour checks they do not even need cleaning.

However, if you run full rich all the time with no TCP, you can dirty them up a bit! See Figure 20.

Unison and Tempest also make these plugs, but we have found the Tempest plugs have a slightly longer “waist” which can interfere with the pushrod tubes during installation (see Figure 20a) depending on how thick the installing 22 mm socket is.

Fig. 20.  Full rich operation with no TCP!

 

Fig. 20a.  The REM37BY Spark Plug. Tempest on left, Champion on right. The Tempest has a slightly longer “waist”. 

I bought a Tempest plug (their actual part number is UREM37BY) to see if it would fit, and modified my socket so it could be maneuvered past the pushrod tube. This involved making the hole in the socket (see Figure 20b) AND taking away material so that the entire socket is much thinner. This socket is only used for Tigre plugs so it does not need a lot of strength. It worked, so I bought 3 more. I am now using 4 Tempest plugs in the front (exhaust) hole of each cylinder and 4 Champion plugs in the rear (intake) holes.

My partner cannot get the Tempests installed because of push-rod tube interference (see Figure 20b) and he has a thicker socket. I have noticed that on my engine there is a bit more clearance between the plug and the pushrod tube on the FRONT (exhaust) hole than there is on the BACK (intake) hole. This means that when installing the plugs, I have to put the exhaust plugs in first. If I put the intake plug in first and try to put an exhaust plug in right beside it (for example:  the REAR plug of Cyl. #2 and the FRONT plug of Cyl. #3) I can’t get the socket on to the exhaust plug – even with my thin socket. The other spark plug gets in the way of the tool. The exhaust plug must be installed and torqued first. THEN the adjacent intake plug can be installed.

When it’s time to replace the Champions, I will buy another Tempest and see if it’s possible to install two of them beside each other.UPDATE: My engine is now using 8 Tempest plugs.

Fig. 20b. Special socket for torquing spark plugs. Note interference with push rod tube.

As a side note, we strongly suggest ALL wiring associated with the mags, radios and engine analyzers be shielded – even wires that the manufacturers say do not need to be shielded. Additionally, one should try to get all ground wires together BEFORE they are actually connected to a ground. Effectively, all ground wires should NOT touch a ground (usually the airframe) until they have all been grouped together. And something else to consider is only one end of any shielding wire should actually be connected to a ground! The other end should “float”! Doing all of the above will eliminate many radio “noise” problems. So say my expert friends….(however, I’m told that BOTH ends of the spark plug wire should be grounded.)

One last comment while still on the subject of magnetos: The magneto switch in the Jungmann can appear to be off when in fact one or both mags could be live. The little ball on the stick that we move up and down to select the operative magneto(s) can be removed from the switch, or fall out of the channel that holds the long straight edge in the switch, while the switch remains in an active position.

Fuel selector off, throttle closed, mixture at idle cut-off, the little ball on the stick freely swinging on its chain (completely out of the mag switch) giving the illusion that the mags are off, …AND… it is truly amazing how many days can go by with a combustible charge sitting in a cylinder (after engine shutdown), and it can still be ready and able to combust! “How do I know this?” you might ask…

                                                                        BEWARE!!

FUEL SYSTEM

The engine-driven fuel pump (see Figures 21 to 21c) will be looked at first, then we will examine the rest of the fuel circuit and its components including the wobble pump and flop tube, and talk about some of the problems we experienced.

Aerobatics is the reason we need a complicated fuel system. If there was always 1 “G” positive, the Tigre wouldn’t even need a fuel pump. Negative “G” complicates things: In order to get fuel to the carburetor we now need a flop tube in the fuel tank and a pump. The engine driven pump on the Tigre is a complex device that usually works very well. However it can fail (recall Tony Smith’s Australian flight.…), and because of it’s design, some failures (one of the copper plates breaking or a frangible link “freezing”) will stop most fuel flow. We therefore need a back up – the wobble pump – to force fuel through a failed engine driven pump. The pump’s design also requires the addition of a fuel recirculation circuit. This is because at idle power the pump is still moving fuel, but very little is needed. The pump would simply be trying to move fuel internally from one side to the other without a recirculation circuit.

Fig. 21. Fuel pump – 3 hoses leave the back of the pump: inlet, outlet (both brown), and fuel pressure (with orange firesleeve).

There is also a fuel pump drain hose at the bottom front of the pump (the silver hose shown at the bottom left corner of the picture). Sometimes you see the fuel re-circulation line coming out of the fuel pump outlet hose’s banjo. Ours starts at the Ellison TBI and continues to the tank ( see Fig. 22a). Note the one-inch diameter tube for cooling air (this is connected to a small ram-air “scoop” on the top of the engine cowl).

While building my Jungmann I was often working in Europe. I used those opportunities to travel to several small airports where Buckers were based and see how people had modified their Jungmanns. I subsequently became good friends with many of the folks I met.

I noticed that in Tigre-powered 131’s nearly everyone had installed a Christen wobble pump/fuel selector. Those that did not have a Christen invariably had fuel leaking out the Spanish wobble pump, and the Spanish fuel selector. I was in the process of designing my fuel system and had many Spanish pumps and valves.

Being a cheap pilot, I decided that I could overhaul these units so they would not leak. I designed and manufactured the complete fuel system pretty well the same as the CASA Jungmanns – before finishing the “overhaul” of the main components. This was not smart. No matter WHAT I did – and I spent a LOT of time and money trying - I could not get the wobble pump and fuel valve to stop leaking. The seats and circular eccentric cams were simply worn out. So I bought a Christen and re-designed the entire system. The Christen wobble pump is not being made any more, so if you see one for sale – buy it!  It will be a better investment than just about anything, even if you never use it…

Fig. 21a.  The Tigre engine driven fuel pump. Disassembly starts.

The Jungmann fuel system is fairly simple. An aluminum tank with a sight gauge on top, a bottom feed, and a flop tube feed – selectable at the Christen wobble pump. This unit is the low point in the system and has an internal filter accessible from the bottom. The tank also has a fill receptacle with an overflow drain tube, a fuel vent port and a fuel return line port.

Fig. 21b.  The pieces that make up the pump.

Fig. 21c. Fuel pump rotor, cavity and valves.

The fuel tank’s bottom feed has a finger strainer over the port to prevent big pieces of tank debris from entering the fuel circuit. The sight gauge is a very delicate unit and must be handled with care to prevent the glass from breaking. The thin gaskets are prone to breaking while installing them, so go very slowly and do not over tighten. It will likely leak the first time you go upside down. Even if you can’t see it you will smell it. Don’t be too hasty in tightening things – just a little bit - then go fly again. It may take several flights of upside down leaks, then repairs, until the sight gauge stops leaking.

Here is an interesting problem I had with my plane. The Tigre would sometimes hesitate and run rough at high power settings, high “G” loading and occasionally almost stop. Once, the engine stopped completely at the end of a landing roll and would not start again. I changed the right mag coil (thought it might be weak); changed the engine-driven fuel pump (thought it had failed as fuel pressure indications were sometimes erratic when the engine was mis-behaving); tried a different wobble pump (I thought it had failed); looked at cooling problems for the mags and fuel pump; changed the left mag impulse mechanism (starts were getting difficult, and during re-timing the mags I discovered the impulse was firing at way too far after top dead center); tested all ignition wires, sparkplugs and distributor caps - looking for a spark jumping to ground; checked and cleaned all the fuel lines and filters; looked for air leaks at all hose connections; checked the Ellison TBI for dirt; checked the intake manifold for leaks; checked the fuel in case we got a bad batch; checked the gap on all rocker arms and changed the grease while I had the covers off; did a compression test and looked at all valves; checked all wiring of the P-leads all the way from the mags to the switches and analyzer connections; made sure all the fuel vents were clear; and re- gapped the points and cleaned all internal mechanisms of both mags.

Fig. 21d.  Fuel tank flop tube. Hose is the new ‘convoluted’ Teflon type 

After most of these attempts to fix the problem I would have to go fly to see if the problem was fixed. There was a lot of time taking the plane apart – fixing – then putting the plane back together and go fly. Then repeat, …repeat, …repeat.

The VERY last possibility we could come up with was a flop tube failure. That was indeed the problem! Flop tubes do fail and this one, after flexing too many times, developed a pinhole leak right at the top of the tube. So the engine would run fine if the tank was full, or if the selector was on bottom feed, or if the plane was upside down.

Otherwise small air bubbles would be drawn in once aerobatic feed was selected and a minute later, when the bubbles had worked their way to the injector, the problems would start. Old flop tubes are made out of rubber and they do have a service life. When was the last time you have heard of anyone changing one?

I changed mine – to the new convoluted Teflon type of hose (see Figure 21d). I wonder how long this one will last…  at least we now know the failure symptoms. However, there are still uncertainties. The convoluted hose is more rigid than previous flop tubes and consequently, I need more fuel in the tank to prevent fuel flow interruptions during aerobatics. I used to have a 35 liter minimum for aerobatic displays or competition flights. I now need 50. I may have to go back to a rubber flop tube. …..(last minute update! HAVE  gone back to using rubber…..)

THE FUEL FILTER

Once the fuel gets to the Christen wobble pump the source is selected and the fuel then goes through a replaceable fuel filter at the lowest point in the system. After 2 years and about 100 hours of flying, the fuel pressure (as indicated on the cockpit gauge) started to decrease. Over the course of a few weeks it dropped to zero! But the engine stayed running fine. I thought maybe the gauge was failing.

I finally decided to have a look at the filter. I was quite surprised to see the filter was almost completely clogged! The gauge was accurate – there was almost no fuel getting through, but the interesting thing is the Ellison TBI will work fine in the Jungmann with just about zero pressure – a comforting thought if ever a lot of things decide to go against you at once.

The debris was the result of building and rebuilding the fuel system. No matter how hard one tries to clean the tubes before first flight, a lot of gunk can hide. I change the filter every year now. Cost is about $3, but the filters are not dirty anymore. Just that first one – so perhaps one should check the fuel filter about 25 hours after any modification or repair to the system…

THE RE-CIRCULATION CIRCUIT

Fuel then travels to the engine driven pump. This is an intricate unit that can be overhauled by someone who really knows what they are doing and has the special tools to do it. Fortunately I have not had any problems yet – other than imagined ones. The back of the pump has 4 ports:  Fuel in; Fuel out; Fuel pressure; and one for the device by which fuel pressure is set (typical design and easy to use). There is also a pump drain on the bottom forward part of the pump. This vents overboard.

Fig. 21e.   Example of metric banjo welded to AN hose fitting. Also shown is the oil tank Quick-Drain with a “safety” clamp that prevents accidental emptying.

Excess fuel – that not needed by the injector - is routed back to the fuel tank. This is a re-circulation circuit that keeps the engine driven pump happy, and the pilot happy by reducing the possibility of vapour lock. In effect, fuel that is being heated during it’s travels through the engine compartment, is always moving through the lines and back to the tank where it is cooled again. Some installations have the re-circulation hose attached to the fuel pump’s outlet port banjo fitting, and routes from there to the tank.

There is typically a restrictor installed in these fuel circuits. So what type and size of restriction is needed in the re-circulation line to keep fuel pressure within the TBI or carburetor limits? One installation we know of has NO restrictor installed. Engine operation and fuel pressure indications appear to be normal using a 3 mm inside diameter fuel return line. We found that a 1.0 mm (0.040”) diameter hole drilled in a 5 mm long  piece of  aluminum that is then forced inside the firewall bulkhead fitting works well with the Ellison as well as the CAT carburetor.

All my hoses are imperial dimensions, using metric-to-imperial fittings I made – usually by welding metric banjo fittings to AN hose fittings (see Figure 21e). After I built the plane, I found there are specialty shops that sell these types of fittings already made and anodized.  Sigh…

BAT Inc.  in Sarasota, Florida sells these fittings.

THE ELLISON EFS-4    (ellison-tbi.com)

Without this Throttle Body Injector (TBI), many of us would not be flying with Tigres. It’s a great design, works in all attitudes, does not need much pressure to operate, purchase price is the same as an OVERHAUL cost for many other units and it does not weigh much.

There are considerable efforts going on to “fix” perceived inlet turbulence problems with the Ellison. It’s possible that:   intake manifold leaks, sniffle valve leaks, primer leaks, manifold pressure gauge leaks, bad spark plugs, ignition timing , a weak spark and fouled plugs because of high lead content are more dominant in engine mis-behavior than TBI issues. Having two similar installations with the standard square Tigre intake plenum and manifold producing extremely good and identical temperature and power results over the entire normal operating range may indicate that the TBI and/or inlet turbulence is not the culprit.

Fig. 22. This design allows operation with both Ellison TBI and CAT carburetors.

The Tigre has two types of inlet configurations. Our -B5’s have a square, fairly big box that has an “airflow angle” of about 95 degrees between inlet air axis and TBI axis. Most –A and –B models have the long “tube” type inlet that bends 120 degrees and effectively has no plenum.

When this second type of inlet was first paired with an Ellison TBI, it was discovered that a spacer was needed to allow the Ellison to fit in the allotted space. The placement of the spacer proved to be important. Operators found that the engine/TBI combination would work reasonably well only if the spacer was placed after (downstream of) the TBI.  We do not have experience with this type of inlet, but understand that several Tigre operators in Europe are actively involved in modifying the entire inlet design to include a plenum. Research is also being done on modifying the TBI to make it more compatible with the Tigre. The fact that these efforts are ongoing would seem to imply that not everyone is totally happy with this type of inlet’s performance. Because our experience has been so positive with the Ellison using two different data acquisition units giving identical results, our thoughts are to ensure measurements are accurate and correct ALL other possibilities before modifying the inlet.

As far as we know the 95 degree-square-box type was only used with the -B5 engines, although advertising pamphlets from E.N.M.A. show the B5 engines with the “round tube” type installed.  The 95 degree-square-box type also needs spacers. I added a 19 mm (0.750”) spacer on the top in order to contour the mating of a square box (the Tigre plenum) to a round hole (the TBI inlet), and a 6.3 mm (0.250”) spacer on the bottom – again to contour the slight change in mating profiles (see Figure 22a).

 

Fig. 22a.  95-degree inlet box/plenum with spacers and Ellison EFS-4 TBI

My partner did the opposite (see Fig. 22b) on his installation: a thin spacer on top and the thick one on the bottom. Performance-wise there are no differences between the two. This is based on 18-hour formation trips with identical fuel consumption and engine/airframe performance.

During the initial set-up after installation on my airplane, it was discovered that the TBI performed very well until it approached the full throttle position. At that time, the installation manual for the Ellison TBI suggested that in order to improve wide-open throttle operation, it might be necessary to limit the amount that the slide in the TBI actually opened. This means that “full-throttle” in the cockpit would become “full-throttle-minus-a-bit” at the TBI.  

The manual now says the EFS-4 “should not require any open throttle adjustment”. It also says: “engines operating with poorly designed air inlets may demonstrate engine roughness at wide open throttle, inability to tolerate lean mixtures, and substantial variation in cylinder to cylinder head temperature or exhaust gas temperature.” Adding a big plenum cured the problems I had in my Pitts when I initially used the Ellison. In the Jungmann, however, I decided against adding a plenum – too much work!

Fig. 22b.  Ellison TBI with thin spacer at top and thicker spacer at bottom.

I have restricted the full throttle position by 4 mm at the TBI. Wide open throttle now gives no roughness, no excessive EGT or CHT on any cylinder – as confirmed by the analyzer; cylinder to cylinder variation stays steady as power is increased from idle to max and is not at all “substantial” when compared to Lycoming values; full throttle operation gives 2300 RPM in level flight at sea level (this is the design specification); full power static run-up at 1000’ ASL gives 1950 RPM (85% of max RPM) and manifold pressure 1.0 to 1.5 inches below pre-start reading. In short, the engine seems to be developing rated power.

We are happy with the way the Ellison works using the “square” intake box.

We know of at least 2 operators in Europe that have made big plenums to replace the 120 degree round tube type of inlet – one of aluminum and one of carbon fibre (see Figure 22 c.). Results so far have been inconclusive – perhaps a slight improvement in mixture distribution but still rough running at full throttle. An airflow straightening grid was tried on one of the plenums but was taken out as there was no improvement with it.

Fig. 22c. Carbon fibre plenum feeding the  EFS-4

There have been experiments with the fuel metering tube as well. There are 4 rows of holes in the airstream at wide-open-throttle, and bench testing in England has revealed the full-throttle mixture may be too rich. At least one operator has soldered the extra two rows of holes closed and is apparently very happy with the engine’s performance (see Figure 22 d.)

Fig.22 d. The EFS-4 Metering Tube with the 8 holes in rows 3 and 4 soldered closed.

We are aware that several operators have experienced engine roughness just after the aircraft leaves the ground during take-off. This has apparently occurred in many Tigres using the Ellison TBI. The earlier-mentioned potential causes may contribute to this roughness, but there may also be another factor.

Neither of our planes have experienced rough running after T/O. We have noticed that our 2 airplanes are the only Tigre installations we’ve seen that incorporate heat shields between the Ellison TBI and the cylinders. As seen in Figures 22a and 22b there is a metal plate that gives the TBI some protection from the very hot air that leaves the cylinders, as well as provides a barrier for the radiated heat from the cylinders. According to an excellent article by Lyle Powell (available on the Ellison website), radiated heat in the engine compartment is one of the big reasons for vapour lock in fuel system ahead of the firewall. We attacked the problem independently, and didn’t even realize until recently that nobody else appears to have done what we thought was an obvious necessity!

I made my shield out of 0.025” 4130 steel, and my partner made his from 0.025” 2024 aluminum. Figure 22 shows how the forward part of the aluminum shield is bent over a bit to protect the diaphram housing of the TBI.

Of course we could be completely wrong! Perhaps it is not needed, but it might be one more thing to try before you contemplate getting rid of the Ellison.

Another possible explanation of the rough-running after take-off was recently told by some of our friends in New Zealand. The engine had an EFS-4 and on the second test flight take-off with this unit the engine ran very rough at about 400’ AGL. The (extremely calm) pilot slowly reduced the throttle and found that at about 1/3 throttle the engine would run fine, and thus managed to complete a very low (and stressful) circuit. This happened several times and the unhappy operator was in contact with the manufacturer. A delayed response prompted our friends to do an autopsy themselves, and they found a small defect in the fuel metering diaphram that may have allowed the diaphram to flutter. This is anecdotal and not documented by approved sources, but may be one more possible avenue to investigate before giving up on what we think is a great injector.

Downstream of the Ellison things can get interesting! The intake manifold in the -B5 is a sort-of ‘square-tube’ design as opposed to the ‘round-tube’ type of the other engine variants.  Neither is very efficient. The fuel/air mixture must make two 90-degree turns before it gets close to a cylinder, which causes a lot of turbulent flow. During engine start, this turbulence, combined with a cold container (the manifold walls), can cause the fuel suspended in the mixture to leave the mixture and “puddle”.

There is a “sniffle” valve located at the aft end of the manifold to drain this “puddled” fuel. It is hidden inside the manifold and most of us just think of it as a drain. It is important that this valve does what it is supposed to do. As soon as there is low pressure inside the manifold (engine is operating) the sniffle valve should close. If it does not close there will be a constant leak of air into the mixture. A small leak seems only to affect cylinders 3 and 4.  A big leak may affect all, but ANY leak should be fixed.

It is very easy to see if the sniffle valve is working. With the engine operating at idle, remove the hose from the valve and place your finger over the valve. You should not feel anything. If you feel suction (low pressure) on your finger, the valve is not working – it’s letting air into the manifold. This is not reason for grounding the plane, but it’s much better to have everything working properly so that problems don’t compound.

Fig. 23.  The new sniffle valve.

Every Spanish sniffle valve I have (3 of them) leaks! I could not repair them so I made a new one (see Figure 23). These valves are in use on some Lycoming engines so I just copied that design.  Essentially, there is a small ball bearing that gets pulled up into a seat when there is low pressure inside the intake manifold. When the engine shuts down the ball falls down and any puddled fuel can exit.To be sure there is never a leak (even at high manifold pressures), I have also added an Andair check valve in the hose that runs from the sniffle valve to the drain mast.  The above-mentioned hose must now be able to withstand very low pressures. Manifold pressure at idle RPM (500) should be between 10 and 11 inches. If it is higher, something is leaking.

Here’s another VERY interesting observation: I have been using one magneto that does not have any retard restriction because it has a good impulse location (0 and 5 degrees after TDC), it has the Bendix coil, and the weights in the retard mechanism connect to the rotor drive with long “pins” (see Figure 16) instead of the more familiar small stubs shown in figure 14. When the engine is shut down, I always select the 2 mags individually at slow idle to see if they are grounding. Typically, when using the left mag (the one that now has no restriction), the engine would do the rough, about-to-die shake and rattle. After I installed the new sniffle valve, the idle on the left mag was very smooth! No roughness at all! With carb heat on and mixture rich I have seen the left mag idle as low as low as 310 RPM (very accurate Engine Analyzer indication)! That is incredibly slow for a smooth idle using only one mag on a Tigre!

Could the historic rough idle of the Tigre be caused by an intake manifold or sniffle valve leak? Leaking primer? Leaking manifold pressure gauge? Or maybe weak valve springs? Old, weak ignition coils? Bad spark plugs? Perhaps timing too?  OR a combination of all these…….

One last observation with respect to the sniffle valve: While operating with the old valve, low power settings, such as descent for landing with the throttle closed, would often induce backfiring. Playing with the mixture would sometimes reduce the backfiring but it was always there. Since the new sniffle valve was installed there has never been any backfiring. I have seen manifold pressures as low as 5.5”. That is a lot of suction, and if there is an intake leak there will probably be backfiring!

The design of both types of manifolds causes starting problems for the Tigre in winter. We have discovered that pre-heating the manifold with high wattage halogen lights works well. The colder it gets the longer the lights must stay on – close to, and pointed at, the manifold. I use pre-heat if the OAT is below 5 degrees C (40 F).

VALVE SPRINGS 

A Tigre valve spring story:

I wanted to do the best for my newly overhauled Tigre engine and thought that the most expensive high-temperature grease would be the perfect for use in the rockers. I thought I knew what I was doing. Events proved otherwise…

The engine was running great, and all parts were to spec - including the valve springs. I measured them myself using a nice compression fixture for the scale. Load was a little low at the end but still within limits. The lengths were very slightly less than the lower limit.

 

A new good spring should meet following values:

Compression spring:                               G-213   

Uncompressed Height :                         >59.0 mm +/-1 mm       

Force gradient:                                     > 5.2  N/mm       

Compression force at 28 mm height        >160 N  <170 N     

ATTENTION: DO NOT COMPRESS MORE THAN 28 mm TOTAL HEIGHT

Compression spring:                             G-212  

Uncompressed Height :                         >56.0 mm +/-1 mm                                                                Force gradient:                                      > 5.2  N/mm                                                                                    Compression force at 27 mm height        >150 N  <160 N  

ATTENTION: DO NOT COMPRESS MORE THAN 27 mm TOTAL HEIGHT

Use a proper fixture so that you do not kink the spring or over-compress it. It will lose its strength if over-compressed.

I did not have anything else at the time and thought it would be close enough???

End result: è engine runs really nice!!!

After a few hours of flight the engine did not want to idle anymore. It was running rough and misfiring at idle. However, full and cruise power was no problem!

So began 4 months of investigation: new ignition harness; reworked the magnetos twice; spark plugs changed to REM37BY; installed a new carburetor; installed a manifold pressure gauge; blamed the Ellison TBI’s sensitivity to inlet turbulence (wrong); blamed leaking primer nozzle (wrong); blamed the sniffle valve (wrong); suspected a leak in the welded manifold tube and changed the manifold (wrong).

A compression check showed almost perfect results on all cylinders, but the engine would not indicate less than 15” of manifold pressure.

Thoughts of a Lycoming darkened my mind!

I bought and installed a 1000$ engine information system from Grand Rapids Technology. This was perhaps the best money ever spent on the airplane!

Next flight, while descending (throttle closed and the prop driving the engine), the EGT of Cyl # 4 just died and dropped to ambient outside air temperature. After landing and adding power to taxi the cylinder comes alive again! Interesting…so another circuit was flown – same result.

Studying the EGT’s closely revealed they were high (normal) at full power on all cylinders. At cruise power all EGT’s were lower and nearly equal. Slowly reducing power showed an increase in cylinder 4 EGT, then a sharp drop of cylinder 4 EGT to ambient temp. This was followed by a rising cylinder 2 EGT and rough running. Manifold pressure never went lower than 15”.

I eliminated fuel starvation as it was OK at full power. Ignition was OK and changing parts had no effect. Could be the Ellison, but it worked well for the first 2 hours. The manifold pressure at idle was high and the dying cylinder at low pressure led me to the conclusion that I either have an intake leak or one of those mysterious problems with the Ellison intake turbulence??

Next morning the outside air temperature was a little cooler. I turned the engine over by hand to feel the compression when cold. I heard a strange noise. A “plump” noise  - just twice, and could not repeat it. What can make a “plump” noise, which was easy to hear but is not repeatable??

Is the engine falling apart and are there parts “plumping” around inside??

About an hour later I tried again with a stethoscope on the engine. There was the “plump” again. On cylinder # 4 ????  The “plump” noise comes out the exhaust.

I opened the rocker arm covers to check the valve gap. The grease looked as if it was brand new, unused and pushed aside. There was a clear path for the spring and valve train. The valve gap was OK (0.3mm). Pushing on the exhaust rocker and lifting the valve off its seat makes that “plump” noise! And that push was not hard at all - just a slight push with the thumb was enough. However the intake valve needed a bit more push. Both valves should need the same push! I was unable to see if a spring was broken.

So I decided to check the other cylinders to compare the push needed for valve opening on all rockers. I removed the other 3 covers and found all the grease looked unused and pushed aside. I removed all grease to see what was going on.

All intake valves were hard to push. All exhaust valves were easy to push. Cylinder #4 was the easiest followed by # 2.  This pattern was similar to the EGT readings at low power on the engine analyzer. First cyl # 4 dies then cyl# 2.

THE LIGHT COMES ON!

The intake leak is confirmed. However, it is not in the intake system. The leak is in the cylinder during the intake stroke. If the manifold pressure is below a certain threshold, the suction is enough to lift the exhaust valve off its seat. This will allow the burned gases in the exhaust stack to be sucked back into the cylinder and mix with the fresh intake charge in the manifold tube. The end result is a mixture far too lean for combustion. The EGT will drop at the manifold pressure required to lift the exhaust valve. The “plump” noise is created when the exhaust valve is lifted up by the suction and then falls back in its seat.

Now I had to check the valve springs again! Upside down, under the airplane was lots of fun. With a lot of sweat and hurting fingers I got them out.

The valve springs are shot. With a length of 45 mm and visually collapsed windings, I had the explanation of my 4-month odyssey. My great high temp grease did not melt. Therefore the grease did not contact the springs and so could not cool them. They overheated, lost their temper and collapsed. Total flight time was only12 hours.

I found 4 brand new sets of springs in Germany and borrowed another 4 sets from a used engine. The length of the new springs was 59 and 56 mm. The used ones 57 and 54 mm.

One month later everything was back together. The engine’s manifold pressure goes down to 10” and at 450 RPM it idles like a sewing machine. Full power is still the same and very responsive to throttle movement. The EGT’s are all close together – much closer than the spread seen in the O320 Lycoming on my Citabria.

I now use the grease recommended by the German Bucker pilots.

The bright side of this story is that the engine still developed full power with completely shot valve springs! No sign of valve flutter, the max static rpm remained unchanged and EGT readings at full power were unchanged. The Tigre seems to be a very reliable engine with some great fail-safe features.

But where can I find some more good springs for the exchange of my 4 sets of used ones.

I started to measure and re-engineer the existing valve springs. It was very apparent that the focus had to be the force at max compression stroke and the force in installed closed position.

The installed close position needs to be as high as possible. The full compressed force should be at the lower end of the tolerance to not overload the camshaft drive system. There seems to be no problem at the higher rpm or even at aerobatic rpm. However this requirement is difficult to meet given that the manufacturing process leaves only a small window to improve the situation by a given design and the properties of steel.

I had to remove one of the new springs on the engine to verify the “new product” dimensions. To my surprise I found that the valve spring had “settled” from its original length of 59 mm to 57 mm. This indicates that the original Spanish springs have not been pre heat set and they settle in the first couple of heat cycles to the length I measured on all used springs.

To increase fatigue life I planned for a shot peening treatment together with a high temp heat setting in the manufacturing process. The wire dimensions, winding and pitch are identical. I increased the pre heat set length to 59-60 mm. This allows now for a slightly higher closure force when the valve is in closed position and just the same force as original when the valve is completely open, ensuring the valve train does not see any higher loads.

After a long wait the brand new springs are in a box on my workbench. They look very similar to the G213 and G212. The loads and dimensions are in the original spec tolerance with a focus on high closer pressure and lower end force, but still within tolerance. Fatigue life, heat set and temperature resistance have been improved to what modern technology can offer.

I paid for a complete set of 16 springs CAN $224. Not bad - compared to a set of Lycoming springs!

I hope this trouble-shooting story helps to build confidence in the Tigre engine.

THE VALVES

The valves, particularly the exhaust valve, have been the source of many problems over the years. These problems, though, may not be the valve’s fault. If one uses the wrong grease for the rocker arms, or allows the cavity to lose all the grease and a valve sticks or fails, is it the valve’s fault?  If a valve spring breaks or looses its temper such that the cylinder doesn’t work properly and a valve fails, should we blame the valve?  What if the timing is off, or spark plugs bad and we allow lead to build up? Perhaps the biggest variable for the Tigre valves is the accuracy of the initial manufacture. If the cavities for valve guides and valve seats are not machined PERFECTLY – and/or the guides and seats are not reamed properly, wear will be uneven and we will eventually have problems. Everything affects everything else - and this is quite apparent in the Tigre. Maybe if the cylinder is made properly and we give it the correct mixture, burn it at the right time, and lubricate it correctly the valves would last a bit longer.

There is usually not much we can do about the engine we have, unless we plan to completely overhaul it. But there are ways of treating the valves such that they last a little longer.

We have used an engine analyzer for most of our engine’s lives. It is a fantastic tool. We can lean aggressively in almost ALL phases of flight and be confident we are not damaging the engine. We also use TCP fuel additive. As a result we are not filling the cylinder with excess fuel – which keeps unwanted combustion by-products to a minimum. It is also immediately apparent when a problem develops, and we generally know which cylinder is the culprit.

As was also mentioned, the type of grease in the rocker box is very important. In order for it to lubricate and take away some heat, it must be in a semi-liquid state. Most modern greases do not do this. They want to stay solid for as long as possible as temperature increases. There is a German grease, named “HeiBlagerfett Extra (diamant), that has been used in one of our engines that is specifically made for this type of operation. I have used Aeroshell 14 because it is almost the same, a lot cheaper and easier to get. I have just heard about a North American equivalent to the German grease called “Jewell” oil. Looks like it might be what we are after, but I have a whole case of Aeroshell 14 that will be used before I switch!

When the grease turns to a liquid, it can leak if the rocker covers are distorted or if the gaskets are in bad shape. If it all leaks out you are inviting trouble, so I have started greasing the rockers every 3 hours.

However, if there is too much grease in the rockers it will start to leak at the top, or bottom of the pushrod tubes. This can make a mess all over the inside of the engine compartment, but perhaps more importantly, the grease can stick to the cooling fins on the cylinders - and I noticed when the back side of my cylinder #4 had a lot of grease in the cooling fins (from a leaking pushrod tube) the CHT was about 10 degrees warmer than when it was clean. Removing the grease immediately returned the readings to normal. Our CHT probes are the spark-plug gasket type. If a CHT indication is higher at the spark plug because of grease on the cooling fins, it is difficult to say what is going on in the other parts of the cylinder – maybe strange, thermal stress things are happening??

There could be other problems if there is too much grease: we have reports that in the Kinner engine too much grease has caused valve “stretch”. Perhaps with no place to expand (with heat), the grease is being forced into the valve guides…

The Tigre manual says when putting new grease in the rockers to put in 200 grams. This is a small yogurt cup full of grease. Presumably, the amount should stay at 200 grams - not more or less. We are using 150 grams as 200 looks like too much

The pushrods that drive the valves are in a housing that cannot seal properly. It is a mystery to me why they would design something like this, but the pushrod cover tubes are mis-aligned vertically by 1.5 degrees! Old Bucker newsletters detailed a way to prevent some grease loss by inserting an “O” ring at the top of the tube.

Believe it or not, you can install these “O” rings while the engine is on the plane – without taking off the cylinders! (You have to remove the rocker arm in order to lower the pushrod enough to get the “O” ring installed – difficult and time-consuming, but possible.) I did this when I first started operating the Tigre, and for the initial 300 hours have had very little loss of grease through the pushrod cover tubes. Actually, if there is grease leaking from the tubes, you know that particular cylinder HAS lots of grease! Too much may be better than too little – it is easy to run out of grease and there is no immediate indication. We have no data on how long the engine will run before damage occurs, but the currently available information suggests a visual check every 25 hours.

Several Bucker owners have switched their rocker boxes to oil. These cavities are designed to hold about 0.3 liters of oil. One report I have says that 8 hours of flying – during which there were 2 aerobatic flights – caused a loss of 20% of the oil. I suspect that more aerobatics, especially negative “G”, would cause more oil loss. There may also be long term problems because of oil foaming and low pressure in the exhaust valve guide. (We need more data!)

ROCKER ARMS

We have seen that the axles on which the rockers rotate often show signs of abnormal wear. This is usually not apparent until engine overhaul. Their only method of lubrication is rather dubious at best: they are supposed to get some of the liquid grease into the rocker’s bushing through a few VERY small holes in both the axle and the rocker arm. This design might work well if the parts are submerged in oil, but it is obvious that not much grease is getting in.

We have modified the rocker axles with the addition of a grease (Zerk) fitting on one end cap, and made the two hollow rocker axles in each cylinder into one continuous cavity by adding an “O” ring-sealed “bridge” between the 2 rocker axles (see Figure 24). When this zerk fitting is pressure greased (with a grease gun), the grease can only exit the cavity by squeezing out through the rocker bushing – thereby lubricating the axle.

The bridge is made of 6061 aluminum as per the dimensions shown in Figure 25. Make the part as close to the values as possible – the 0.285” dimension is dependent on the space between the 2 retaining screws (shown in figure 24) being 21.0 mm. Note:  If the max values are exceeded the part will not fit!

The zerk grease fitting is screwed into a fitting that is made from 17 mm 6061 aluminum hex rod. I found that it is only available in Europe!

One rocker arm per cylinder must be removed to install the “bridge”, and, as was mentioned earlier, this installation can be done on the airplane without removing the cylinder. It is, however, much easier to do this when the cylinders are off the engine.

 

Fig. 24.  This shows the “O” Ring-sealed “bridge” between the 2 rocker axles.

The 2 screws (shown at top) position both axles rigidly in the cylinder, so the aluminum “bridge” must fit very tightly with the “O” Rings compressed (but not destroyed!). The grease gun can exert high pressure. If the “O”rings are loose, the grease will force them out of position, the cavity will not take the pressure, and the axles will not get grease.

Fig. 25. The rocker arm “bridge”.  Do not exceed max and min values. Radius 0.040” is for a 2 mm diameter “O” ring with an I.D. of 11 mm.

MAIN CRANKSHAFT NUT

The main crankshaft-retaining nut at the front of the engine can be EXTREMELY difficult to remove! I got a set of Tigre tools from Germany many years ago, and the first time I used the special tool on an engine I destroyed the tool!  (See Figure 26)

We ended up making a very thick new tool – very heavy duty – that worked by using a long arm as well as a big hammer! It is a right hand thread, but be advised some of these nuts do not want to come off! It will take a LOT of force…

Fig. 26. Special tools with VERY long moments! Still not enough force!

TACHOMETER CABLE

These too, are old, and the outer sheath often starts to come apart, which allows oil to spray into the engine compartment as well as the cockpit. They can be repaired by unsoldering the ends, removing the internal drive cable and re-twisting the outer sheath so that it forms a continuous surface again. Then the outer sheath can be covered in heat shrink tubing to try and keep any oil inside the cable. Re-solder the ends and then lubricate the cable with LUBIPLATE 630AA (Aircraft Spruce). See Figures 27, 27a and 27b.

Fig. 27 Tachometer cable repaired and covered with heat shrink tubing.

Fig. 27a.  Tach cable taken apart and ready for repair.

Fig 27b.  Repaired and ready for re-assembly.

CYLINDER BASE NUTS

There are many “horror” stories circulating about the Tigre cylinder base nuts: they are always loosening; you need pal-nuts or some sort of locking device or they will come off; the studs are weak…

Our experience has been positive – nothing to report. We have taken apart some spare engines, though, and on one we found Imperial Dimension nuts! One half-inch nuts  are close to the 13 mm metric size but obviously they cannot be used. Perhaps this is how one of those rumors got started….

There is one problem with them, however:    actually getting the torque from the wrench to the nut! A straight line between the ratchet and the socket (for the extension) is not possible because parts of the cylinder get in the way. So an offset extension must be used. Then there is very little room to get the “crow-foot” type socket on the nut. And if you do get it on, it seems you cannot rotate it very far because the cylinder barrel gets in the way. So again, we made new tools (see figures 28 and 28a).

It appears that if the nuts are torqued to the correct value (and no paint interferes with the mating surfaces) the old Spanish nuts will work fine.  No “locking” device (such as a lock washer or pal nut) is used on the nuts of one of our engines. The other engine uses the standard metric locking tab washer.   It would, however, be nice to find a source of new nuts.

Fig.28. Tools made for tightening the cylinder head nuts. Made from ½” thick tool steel.

Figure 28a.  The offset wrenches. Essentially, they are 10cm and a 15cm offsets.

ENGINE ANALYZER

These units are amazing tools and have helped us with our initiation into the world of Tigres. We are each using a different type – one is an EI UBG-16 (see Figure 29), and the other is the GRT EIS 4000 (see Figure 30).

As we have the same airframe/engine/carburetor/mags/propeller/empty weight, we would expect similar readings. And indeed we have. The installations are similar in that probes need to be installed and wires run, but there are a few differences. The EI unit in the Jungmann has a remote head because the full unit is too long to be installed behind the instrument panel. The only available space I could find that would enable all wires to reach both the probes and the modules was in the left wing root area. There are a LOT of wires and cables now in that space (see Figure 29a). I have since found out that the manufacturer will make longer cables if you ask for them before purchase, so a different location is possible.

Fig.29. The EI UBG-16 remote head as installed in the Jungmann.

.

Fig. 29a. The wing root area filled with cables and modules for the EI UBG-16 Analyzer.

Fig 30.  The GRT Engine Information System.

CASE CORROSION

The crankcase and accessory cases are made of an aluminum alloy that has varying percentages of magnesium. We have found the Spanish did not properly prepare the engines for paint and, as a result, this alloy tends to eventually shed all paint. It then starts to corrode – a white powdery type of corrosion. The only way to prevent this is to get rid of the old paint, chemically etch and alodine the surfaces then prime and paint. Preferably with an epoxy primer and a good topcoat (polyurethane or similar).

Fig.31. This rear accessory case had no paint left on it!

MOISTURE ABSORBERS

To reduce corrosion damage on the valves, desiccant (silica gel) canisters are placed in the exhaust stacks after flight. (See Figure 32) The silica gel absorbs moisture, and it’s color changes from dark blue to pink as this happens. When they are all pink, put them in a microwave oven for ten minutes at 20% power. They turn blue (by heating up) and then they are placed back in the exhaust stacks to absorb more moisture.

The canisters are made of cheap PVC tube, glued to a plug end, and turned on a lathe to reduce the diameter. A nylon screen is glued inside over the small hole, then they are filled with silica gel (Aircraft Spruce sells this in a one pound size can), and lastly more screen is pulled over the end and secured with rib-stitching cord. The desiccant tube is put in the exhaust stack, then the rubber plug is put in and expanded by turning the screw. The rubber plugs can be found at any swimming pool supply store.

Fig. 32. Desiccant (Silica gel) ‘Canister” and plug.

ROCKER BOX COVERS

These covers can leak grease –mainly because the covers themselves are often deformed, the screws come loose, or the gaskets leak. The orange, sort-of soft rubber gaskets are good, but if you tighten them just a bit too much they will break and leak. Again, don’t ask how I know this…

We have found a thin layer of the black gasket goo that mechanics apply to the surfaces (to prevent the gasket from moving during assembly) works very well in stopping leaks.  So does a thin layer of silicone sealant. It seems just about any type of engine gasket material works well.

The grease can also leak out through the bolt holes if the retaining screws get loose. There is typically no lock washer under these screws. We have found a new type of lock washer called the “Nord-Lock” that works great! However they are not made in the size we need. So you buy the 8mm I.D. washers (they come in pairs) see Figure 33, and make a jig to hold the washers for drilling the I.D. to 0.358” (see Figures 34 and 35). A pair of washers are drilled together in the jig. They are stainless steel but drill speed should only be about 300 rpm.

Fig. 33.  The new “Nord-Lock” lockwashers.

They are used in pairs: The serrated edges go against the bolt face and rocker covers, while the scalloped edges mate with each other. When the screw is tightened the scalloped surfaces come together and resist any loosening rotation. None of our rocker cover screws have loosened since using these new washers.

Fig. 34.  The jig that holds the washers.  It is held in a lathe chuck.

Fig. 35.  Steel plate that holds the washers in place while drilling a bigger (0.358” dia.) hole.

BRASS MANIFOLD NUTS

In North America it is common to see brass nuts on a cylinder’s exhaust port studs in air-cooled engines. The rational is that because of the high temperatures there will be more corrosion, and if the nut was made of steel it will be very difficult to remove the nut for routine servicing. These brass nuts are supposed to be only used once then discarded.  A few inches away on most cylinders, the intake port studs have re-useable steel nuts with lock washers. I have rarely had difficulty removing the intake nuts.

The Tigre, on the other hand, is supposed to have brass nuts for both the exhaust and intake port studs. Brass is much softer than steel and can deform much more easily. If a brass nut is over-tightened or re-used several times, the threads can become very “sloppy”.

I have had the intake manifold of my Tigre off several times and have re-used the nuts that came on the engine when I got it. I may have noticed they were brass, but no red flag popped into my brain - and I carried on using them over and over. Last winter I took the engine off the plane and disassembled it for measurement and paint. I reassembled it using new gaskets and appropriate hardware but again re-used the intake manifold nuts – torqued to the value given in the Tigre manual.

After a few hours the idle started becoming a little rough and the idle EGT’s were all increasing. I no longer had to lean at idle. As I had sent the TBI to Ellison for overhaul while the engine was apart, I assumed that perhaps they had done something to the idle circuit. All 4 EGT’s were high and with each flight getting higher. Full and cruise power were no problem – all indications were normal but it seemed I could not move the mixture lever as far back as I could before the TBI’s overhaul. It HAD to be the overhaul of the TBI. I thought for sure they had screwed something up.

My normal approach to land is with throttle closed (idle), and there started to be a lot of backfiring during descent. This engine hadn’t done that in years. Fiddling with the mixture did not help. Finally, during a pre-flight inspection, I noticed blue (fuel) stains underneath the intake manifold where it mates with cylinder #1. How in the world can fuel get in that area?

Answer: EVERY intake manifold brass nut had come loose! All 12! A huge intake manifold leak. Hence the high EGT on all 4 cylinders, and the backfiring, and the over-lean idle mixture.

I thought perhaps the torque value was incorrect because I had used new, better (?) intake gaskets and maybe a different type of gasket material needs a different torque?? Certainly the brass nuts are part – if not the main reason – of the failure. I am now using steel nuts with a flat washer and a lock washer under each nut (as in North American engines). I also added a bit of the Green Loc-Tite thread sealant to each stud. This stuff lubricates the threads but does not “weld” the nut to the stud as the Blue and Red Loc-Tite sealants do. With 100 hours now they have not moved.

CONCLUSION

Finally, if there is one thing we have learned in many years of working on old, European engines, it is that when something fails to work correctly, your first action should be to ensure that it has been correctly manufactured, installed, serviced, and operated. Next, measure it to verify that it is still within limits, and only if it is not should you replace it. All too often people's first reaction to a failure is to assume that the builders did not know what they were doing, and then to seek some alternative part from the local farm supply store.

We have also learned that there is no such thing as “new”. An exhaust valves may not have been installed or used in an engine, but that is far from making it new. 60 years on a shelf in someone's garage is not without its consequences.

Certainly improved materials have become available, but even then, caution is needed. Buta-N may work better than leather in some instances, but in others definitely not. We should be VERY careful when attempting to “re-invent” the wheel! On the other hand, we know that the designers and builders of these engines would have given their right arm for an electronic engine analyzer.

If you've read all of this report, you are part of a small group of people who are hopelessly addicted to the smell, feel, sight and sound of “older technology” aircraft engines. If you have information that would enhance this report, please send it to us, and after editing we will add it. Let's keep sharing information and working together in order to keep these motors running safely.

---