Flywheel attachment failure

[willowbrook]

Some of you will be aware that over the last two weeks a flight training school in New Zealand has suffered two Jabiru catastrophic engine failures. One 2200 engine broke a through bolt that resulted in a cylinder being destroyed but no other damage. Both this and the 3300 engine had flown approx 880 hours, all since Feb this year.

 

The second failure was much more serious and opens a significant question for us all as regulators and administrators of sport aviation involving Jabiru aircraft. The 3300 engine as mentioned had a total of 880 hours and was fitted with the doweled crankshaft to flywheel attachment. I am sure we had all hoped that the dowels had resolved the flywheel attachment failures but after four years of trouble free operations in NZ this engine has suffered fractured cap screws resulting in loss of ignition and valve timing. The pistons have struck the valves during the event. In the forced landing that followed extensive damage was inflicted on the airframe but the pilot was uninjured. Had the failure occurred during other phases of this flight the aircraft would have been lost at sea.

 

CAA have contacted Jabiru and CASA and further follow up joint work will follow.

 

At this stage CAA has drafted an Airworthiness Directive that will be released on 25 November 2010 which proposes a finite life of 500 hours on the subject cap screws.

 

The AD in New Zealand will apply to their members. I believe CASA will issue a similar AD and RAAus may wish to take similar action.

 

CAA considers the flywheel failure a critical safety of flight issue which becomes even more significant as both of the aircraft involved in these failures were LSA’s operating full time on flight training. This is the tenth flywheel attachment failure we have experienced in NZ with one aircraft suffering the failure twice. These latest failures like most of the others occurred on standard Jabiru installations ie Jabiru props.

 

I suspect there will be some reaction from your members who operate the jabiru and who may not have heard of the accidents and may approach you to find out the background to the AD. I hope this message goes some way to explaining why this action has been necessary. CAA

 

 

[facthunter]

Flywheel Attachment.

 

The flywheel problems obviously point to a problem with torsional vibration.. When this happens the crankshaft can twist excessively due to harmonics within it. The operator is not aware of it because it is happening within the engine. You could expect the flywheel and propeller to be affected, and often the crankshaft itself will fail.

 

All engines experience this to some extent ( even turbines, but it is very high frequency there).

 

Since it is due to HARMONICS it will be worse at some RPMs than others.

 

BIG engines are more critical, but they ALL do it.. It is an extremely destructive force that can be generated. I would suspect that there are CRITICAL RPM settings that should be avoided with these engines. ( that is the way it is handled in most situations).

 

It is a relatively simple matter to STROBE an engine on a test bed, and find out just how much deflection is occurring at what RPM.

 

Regarding the through bolt/stud failure and the resultant damage to the cylinder, I would doubt that nothing else is affected. I am sure if you were in a position to photograph the engine when it happened , you wouldn't want to use much of it subsequently. Nev

 

 

[airangel]

Prop strikes, No matter how minor also seem to play a part in this type of failure.

 

 

[Vev]

Hi Nev,

 

Re Flywheel bolt problem.

 

Thanks for the quality input as usual.

 

Would you expect the 4 cylinder to be more of a problem than the 6 owing the more aggressive firing pattern?

 

Cheers

 

Jack

 

 

[Alan]

I have a Jab J230 and have vibration analysis experience through work. Next time I am home I will attempt to measure the static natural torsional resonce of the propellor/crank/flywheel assembly. It will not be excactly the same as the dynamic torsional resonance but will give some guidance. I am intrigued enough now to put inthe effort.

 

Regs Alan

 

 

[facthunter]

4 Cylinder?

 

This type of problem is not related to firing impulses alone. It is also caused by the inertia of the pistons as they stop and start . This gives them a variable "flywheel" effect. ( A maximum at the position where the piston is near half stroke or just above & zero at TDC and BDC.) The "four" would be the worst. The six should be better, but the trouble is that the longer the shaft the more the torsional "wind up". It is accentuated in the case of the Jabiru by having a prop at one end and a flywheel at the other. Nev

 

 

[Guest ozzie]

Closing the throttle rapidly can also contribute to this sort of failure.

 

 

[Tomo]

The V8 Mack engines were/are a bit renowned to break a crank if you sit there and rev them - nil load.

 

 

[facthunter]

Harmonic.

 

IF you hit a harmonic and hold it there, it will break on NO LOAD. We are talking of seriously large deflections here as a possibility. Nev

 

 

[gregv]

New Zealand CAA AD for Jab 2200/3300 engines

 

This is the AD released today by NZ CAA giving 5/16" flywheel attachment screws in the 2200/3300 a 500 hour finite life. Apologies if this has been posted elsewhere.

 

Greg v

 

JABENG.pdf

 

JABENG.pdf

 

JABENG.pdf

[Guest Maj Millard]

That's pretty impressive stuff from the NZ CAA, very well written, researched and referenced. And super quick, direct and to the point, no BS...with no messing around on such important subjects.

 

Ok, now lets see how our RAAus and CASA do in comparision...to start with they will have to rewrite and probabily change everything..............Maj...

 

 

[jetboy]

They did act swiftly and I have no problem with compliance (doesnt take much more work to replace the capscrews when doing the required checks anyway)

 

However some complications do arise:

 

1/ the subject aircraft a J230 was new this year I understand, and operated as certified aircraft rules ie. LAME mtce schedule with all the enhancements that provides.....and the engine a certified ASTM compliant version with all the extra care and attention that attracts...with the continued airworthiness instructions provided by the manufacturer....so I'm waiting to hear of a SB from said outfit

 

2/ the subject 3300 presumably had the most recent flywheel attachment system - substantially changed pre- July 2008 (Jaba chat refers) which really means the old attachment itself was OK - just the original set - torques have been altered - but still not up to the torque specified by the bolt manufacturer and our RAANZ advice which as a microlight operator I can follow. this is the same advice being used in UK and Europe.

 

3/ I'm not trying to critique or re-engineer the problem, but its not technically really a flywheel, just an aluminium disc for mounting the starter ringgear on and a few magnets. The only failure that results in "uncommanded engine shutdown" is the ignition, which could be eliminated by changing to dual electronic or mags, thus also eliminating the distributors and their drivegears, a reliability risk area in themselves.

 

I'll shutup for now just needed to vent some. I suspect my engine will wear out in other areas before I need to replace those capscrews more than once

 

Ralph

 

 

[Guest Maj Millard]

Jetboy, The starter-drive, flywheel, or whatever you want to call it is a critical component on the engine. As you know if it lets go, or even moves slightly, things could stop pretty quickly, either through an ignition failure, or by mechanical jamming.

 

There have been plenty of problems in the past with direct-drive cranks with props attached to them. the prop when spinning is one huge gyroscope creating massive gyroscopic forces evertime there is movement around the aircrafts' axises. One way to eliminate some of the gyroscopic/torsional forces transmitted to the crank from the prop and pistons, is to attach a similiar force at the other end. Even though in this case it is multi-functional and aluminum, it is still technically somewhat of a flywheel, and probabily acts as such to a degree.

 

On an auto crank/flywheel set-up for instance, you've got a steel flywheel attached to a steel crank, via steel dowels and large steel bolts. Temperature wise you've got even, and regular heat expansion throughout the whole assembly.

 

With the Jab set-up you've got a steel crank with an alum flywheel attached with steel bolts and dowels. Aluminimium which has the highest coefficient of heat transfer and expansion, moves at a different rate to the steel. Some movement therefore is to be expected somewhere, unless everything stays super-tight.

 

Regardless, it is a critical part of the engine and should stay attached !!

 

yes we can only stay tuned to see what Jab comes up with on this one................Maj...

 

 

[eightyknots]

The flywheel problems obviously point to a problem with torsional vibration.. When this happens the crankshaft can twist excessively due to harmonics within it. The operator is not aware of it because it is happening within the engine. You could expect the flywheel and propeller to be affected, and often the crankshaft itself will fail.All engines experience this to some extent ( even turbines, but it is very high frequency there).

Since it is due to HARMONICS it will be worse at some RPMs than others.

 

BIG engines are more critical, but they ALL do it.. It is an extremely destructive force that can be generated. I would suspect that there are CRITICAL RPM settings that should be avoided with these engines. ( that is the way it is handled in most situations).

 

It is a relatively simple matter to STROBE an engine on a test bed, and find out just how much deflection is occurring at what RPM.

I wonder why Rotax, Jabiru, Lycoming, etc, don't publish figures as to which harmonic rev ranges to avoid to prevent engine failure?

 

 

[facthunter]

Why Rotax Jabiru Lycomingetc...?

 

Rotax has a complicated clutch in the reduction drive that allows movement of the crankshaft relative to the prop driveshaft and DAMPS it, so it is taken care of.

 

Jabiru DON'T but should investigate it.

 

Varous Lycomong/ Continental " flat" engines have internal vibration dampers on the crank webs and DO have RPM ranges recommended to be avoided.

 

They also run propellers that are specifically approved for use on each particular engine. Non-approved or "uncertified" propellers are not permitted on Certified engines. These are subject to inspections at mandated intervals and have a finite service life.

 

Large radials have great big counterweights attached "loosely" to the flywheels so that they "rock" around as the engine turns and absorb some of the torsional vibrations.

 

Nonethe less they will have at least TWO RPM settings that are PROHIBITED. Nev.

 

 

[Guest Maj Millard]

QUOTE=facthunter;Rotax has a complicated clutch in the reduction drive that allows movement of the crankshaft relative to the prop driveshaft and DAMPS it, so it is taken care of.

 

Jabiru DON'T but should investigate it.

 

Correct facthunter, except the rotax one is not that complicated. On the 2 stroke gearboxes they solved the problem by using a large rubber donut dampener. In the 912 gearbox the problem is taken care of with proper selection of gears and gear ratios, and the use of a multiple steel disc clutch arrangement, which also acts as a slipper clutch and protects things should a prop-strike occur.

 

This arrangement works well and is only complicated after 900 hrs of reliable use, when it should be removed and sent to the dealer for a check and adjustment, which runs around A$900.00. By far a lot cheaper than the repairs required after a jab flywheel lets go . .................................................................maj...

 

 

[ianboag]

Another 10c worth

 

The bolts/dowels broke.

 

In previous cases, the bolts have broken just where they emerge from the crankshaft. I believe this latest case is the same. When the bolts break like this, the crank gear is no longer connected to the crankshaft. The engine goes round a couple more times (camshaft stationary) and there is clattering and banging as valves hit pistons until it stops.

 

The bolts break there because the flywheel slips and bends them back and forth. The failure comes from fatigue from all that bending. It means there is not enough clamp to hold the flywheel. The dowels were never meant to be drive pins taking the full load - the role of dowels in a clamp joint is just to carry the (relatively) small load that might happen if the frictional clamp forces are exceeded slightly.

 

The bolts break because they are loose. End of story. This is all pretty standard clamp joint stuff not restricted to Jabiru flywheels. Bolts break from fatigue when a joint is not sufficiently well clamped.

 

Assuming the sandwich was all torqued up properly in the first place, this means that either the bolts have unscrewed a bit or the sandwich they compress has relaxed ie the bolt heads have "imprinted" a little bit.

 

We will never know ... because .....

 

The factory specify the use of the mean green 620 Loctite. This stuff is so good that even if the sandwich has relaxed a little you will not find out by putting a torque wrench on the screws. "Unscrew" is not going to happen because the bolts are firmly glued into the bottom of the hole.

 

If the bolt tension REALLY needs watching, the Loctite has to go. Then the screws can be meaningfully checked (untorqued/retorqued) at whatever interval floats your boat. I've heard opinions from non-ignorant people who believe that grease is a better thing to put on than Loctite. I did hear it said (rumours from years ago are always worth quoting) that back when all the bolt/dowel hoohah first happened, the man in Hamilton who became the default Jabiru doweller for NZ would only use grease on the bolts.

 

If an untorque/retorque sequence is not possible then nobody will ever know how, when or if the bolts lose tension over time. We will only know after the fact if they break .......

 

Replacing the bolts every 500 (or 200 or 600 or whatever) hours seems harmless enough. Inspecting them every 20 or 50 or 100 or whatever hours with a torque wrench is a waste of time. For us private bugsmasher people the 500 hour replacement thing is every four or five years. It's not a big job and the bolts are not expensive.

 

Go figure.

 

Ian B

 

 

[Guest Maj Millard]

Have to agree with you Ian. I don't believe the bolts come loose either. I feel it's more of an imprint problem with the aluminum flywheel. Bolts may stretch a little if they're under tension over time, but shouldn't do much in this application. Proper application of dowling should eliminate any radial movement if the bolts become a little loose, to a point.

 

I agree with you 100%, the crush effect is being lost (as it does on a prop hub when wood props shrink). So what's causing the looseness ??....I believe it is the use of steel and alum, if there is any movement at all the steel is going to win over the alum everytime, the alum yeilds and wears, and the looseness starts to do it's thing, further pounding the alum into submission. Maybe steel inserts in the alum flywheel may solve the problem ??..And what about the positive -locking devices commonly used on auto flywheel bolts to ensure they stay put, a simple solution that works on everything from small tractors to Mack trucks.

 

This is not a new problem we are tackling here, solutions have been around for years. Jab now needs to work out which one to use...........................................................................................maj...

 

 

[ianboag]

Going back a bit ....

 

Worth reading ....

 

Contrails ! Jabiru flywheel bolts

 

 

[Guest Maj Millard]

Yes that all makes good sense to me Ian. I did not see any date on those recommendations, would the engines that suffered the recent failures have been modified thus ?.

 

My simple solution to what is really a simple problem would be as follows: if they want to continue to use alum for the flywheel I would suggest at least 1" (25ml) steel inserts pressed into the flywheel around the bolt attachment holes. These would be an interferrence fit, and pressed in with the flywheel heated to assure a tight bond when cooled. Replace the current screws with hex-head high-tensile bolts with a large dia steel washer under them. The washer would be large enough in dia to extend beyond the steel inserts (say 30-35 ml dia). the hex bolts should have positive locking either by drilling the heads and lockwiring ,or by the use of a thin steel locking tab as used on auto flywheels. By using the steel inserts you you have increased the surface area between steel and alum considerably and therefore decreased the wear potential by a similiar amount.

 

All steel to alum contacts should be coated with a light film of grease or similiar, such as a thin film of Duralac, to eliminate dissimiliar corrosion between the two. The increased steel to alum bearing area compared to what is present now should make for a lot less wearing of the mount holes due to any movement, that may present itself over time.................................................Maj...

 

 

[Yenn]

I have seen it written somewhere by Jabiru that 2400rpm should be avoided for harmonic balance reasons. Can't find the reference at the moment but will try to find it later.

 

 

[ianboag]

Yes that all makes good sense to me Ian. I did not see any date on those recommendations, would the engines that suffered the recent failures have been modified thus ?.

This was all from about 2007. What he says is this is a clamp joint and there is heaps of scope to increase the clamp force. Hard to argue with. At the time Jab were saying 18 ft-lb for the bolt torque - they upped it to 24 and he is saying 30 would be fine based on normal sums about the bolt strength. More clamp = less ability to wiggle and chance of breaking bolts.

 

My simple solution to what is really a simple problem would be as follows: if they want to continue to use alum for the flywheel I would suggest at least 1" (25ml) steel inserts pressed into the flywheel around the bolt attachment holes. these would be an interferrence fit and pressed in with the flywheel heated to assure a tight bond when cooled. replace the screws with hex head high-tensile bolts with a large dia steel washer under them and large enough in dia to extend beyond the steel inserts (say 30-35 ml dia). the hex bolts should have positive locking either by drilling the heads and lockwiring or by the use of a thin steel locking tab as used on auto flywheels. By using the steel inserts you you have increased the surface area between steel and alum considerably and therefore decreased the wear potential by a similiar amount.All steel to alum contact should be coated with a light film of grease or similiar, such as a thim film of Duralac, to eliminate dissimiliar corrosion between the two.

I think all one needs to do is crank the bolt torque up to about 30 ft-lb (lubricated) and forget about Loctite. Then you can (meaningfully) check the bolts as often as you want to. There are no locking issues - the bolt is not going to unwind from that level of tension. If you are worried about that, just put a paint blob on it - if it unwinds at all you can see. If it relaxes, one will be able to detect that and either re-torque it or (more likely) replace the bolts ... This is pretty much what was done by Assessco and as far as I know there have been no issues with any of those engines.

 

Go figure.

 

IB

 

 

[Guest Maj Millard]

My solution offers positive locking through the whole life of the engine, and would eliminate any need to go in and periodically check bolt tension, or replace bolts, which can be eliminated with smarter engineering.

 

You don't need to peroidically check or replace the flywheel bolts on your automobile do you ?.

 

Personally I'd rather use the time for flying, not tearing the back off my engine every 200 hours !............................................................................Maj...

 

 

[ianboag]

Problems and solutions

 

Major. You are suggesting a design change and it seems a reasonable one. There have been other suggestions along those lines from time to time eg using a taper/key approach like lots of auto engines, counterboring the crank etc etc.

 

Today's problem though is what - if anything - should be done to the current configuration to make it safer.

 

IB

 

 

[Guest Maj Millard]

Thanks Ian, The solution is there somewhere !!...........................................maj...