3300 Flywheel spacer mod

[gregv]

I've asked plenty of questions of late, and added little. Thanks for the helpful comments from members of this board. All have been very helpful in the top end rebuild process I am now near completing.

 

I thought I'd post a few photos of the flywheel spacer modification I have used.

 

In short, my assessment is that flywheel bolts break because the 'friction joint' between the crankshaft, the ali spacer and ali 'flywheel' gets loose and the bolts subsequently get stressed and break. Maintain the friction between the parts by maintaining bolt preload and the problem is solved. Dowels won't stop this from happening. The new steel centred 'starfish' flywheel on newer engines removes the relatively soft aluminium centre which the flywheel bolts tended to recess into, losing their preload and loosening the sandwiched parts.

 

Old flywheel centre is pictured below with Unbrako 5/16" 1-1/4" long cap screws tighened up against the ali flywheel. Two were loose, and all were recessed into the flywheel by up to 0.3mm.

 

 

I machined up a steel replacement spacer, but dispensed with the vacuum drive extension. I also machined a 4.5mm steel disc, and replaced the bolts with ARP brand 5/16" 1-1/2" long bolt (to counter the new compression disc and provide a few more threads into the crankshaft). The crank bolt holes were cleaned out and checked for adequate depth.

 

 

Replacement steel spacer with compression disc and new bolts

 

 

The aluminium flywheel is now sandwiched between two steel discs. The bolts are spec'ed to to tightened up to 50+ft/lbs into steel. We settled on 45ft/lb as a compromise figure.

 

 

With the vacuum pump drive no longer sitting in the hole in the alternator spider, I can use a mirror to check the witness paint on the bolt heads during regular inspections.

 

This idea was developed by a local Jab owner who generously spent many hours with me going through the rationale for why these bolts break and how he thought one could prevent it - basically get rid of most of the soft aluminium in the sandwich of parts. We'll see how things go...

 

Greg v

 

 

[facthunter]

Looks worth a try. Nev

 

 

[kgwilson]

The rationale appears sound and as the later models are modified to steel (it would appear) to overcome this problem I reckon you are on to a winner here. Time and well documented regular maintenance checks will be the judge.

 

Kevin

 

 

[ianboag]

The rationale appears sound and as the later models are modified to steel (it would appear) to overcome this problem I reckon you are on to a winner here. Time and well documented regular maintenance checks will be the judge.Kevin

And a few months ago, a 230 owned by the NZ Jab agent blew its bolts. It was about a year old. Engine had the new starfish thingo. That was not long after his 120 broke a through bolt, threw a pot and trashed the engine. These little things are sent to try us ......

 

 

[gregv]

And a few months ago, a 230 owned by the NZ Jab agent blew its bolts. It was about a year old. Engine had the new starfish thingo. That was not long after his 120 broke a through bolt, threw a pot and trashed the engine. These little things are sent to try us ......

Hmm. The 230 breaking its bolts in Tauranga was a new one wasn't it? Well, lets hope the ARP bolts live up to their advertised specifications...(ARP #741-1500, nominal tensile strength 180,000psi). Grade 8 bolts are supposed to be minimum 150,000psi tensile strength, with something like 60% of that used to compute shear strength...

 

I've done what I reasonably can to minimise the risk of flywheel breakage without replacing the entire flywheel assembly. I'll post any problems that either engine with this mod develop - call it a study where n=2.

 

Greg v

 

 

[Jaba-who]

Though Jabiru's supposition that the problem is due predominantly to propellor issues would make this mod not applicable to some engines.

 

Will the same issue of transmitted forces still have the same potentially destructive effects even with the plates?

 

 

[facthunter]

It's a clamping problem. There will be bad torsional vibration "periods" (Rev related). ALL engines have them. Why doesn't somebody do some tests?It's not rocket science. Nev..

 

 

[Jaba-who]

Nev,

 

I understand you are not a Jabiru engine owner so maybe you are unaware of the 10 paged document (JSB 012-2 ) that Jabiru have released that includes discussion about the causes of the bolt fractures. They outline 10 causes of bolt fracture of which only two are related to the flywheel itself. All the other 8 are related to transmitted vibration/shock through the crankshaft from the propeller.

 

What tests did you have in mind, and do you have any evidence they have not already done them?

 

I can't say how they arrived at this information but one has to assume they haven't just dreamt it up.

 

 

[ianboag]

I can't say how they arrived at this information but one has to assume they haven't just dreamt it up.

 

Wish I shared your faith on that one.

 

When bolts fail in a clamp joint it's because the clamp is not enough to stop the mating surfaces from moving back and forth and the bolts fail from fatigue. That's for ANY clamp joint, not just Jabiru flywheels. Google something like "clamp joint failure" to learn more. So whatever is going on with torsional vibrations (or phases of the moon or whatever) is enough to overcome the clamp joint friction. The answer is relatively simple - more clamp = more friction - so long as you don't start pulling threads out of the crank.

 

The occasional set of flywheel bolts still croak, so there's your tests. However some basic sums say that you can torque the bolts up to 45 (not 30) with grease (not Loctite). That's what gregv is doing. Nothing will break from the torque and the bits won't move .....

 

Cheers IB

 

 

[facthunter]

IF they had done it, they would have a few recommendations about rev settings to avoid surely? The crticality of the props is a good indicator that things are not right. You can really only run a wood prop on jabiru motors. Wooden props are the most forgiving in this area. The Sensenich composite prop only has a problem with Jabiru. When you have a substantial flywheel at the opposite end to the prop. ( which is another flywheel ) they oppose each other at times, and work together at other times. That is the nature of harmonics.

 

All you have to do is strobe the motor and if you mark references (flywheel/crankcase and prop hub/crankcase) you will see things happening with crankshaft flexing as you move through the rev range. You would have to do this sort of thing when certifying propellers to each individual engine type.

 

The situation is that there is virtually no torque load driven through the flywheel bolts. It is ALL torsional (rotational) speed variation. The bolts fail when the flywheel slips on its driving face and they become loaded in shear.. in opposite directions till they fail..Nev

 

 

[ianboag]

The Tauranga J230 that popped its bolts at 900 hrs TTIS had a Jab wooden prop and the starfish thingo and LAME maintenance every 100 hours (monthly) because it was an LSA being used in a flying school.

 

When you have a substantial flywheel at the opposite end to the prop. ( which is another flywheel ) they oppose each other at times, and work together at other times.

 

The aluminium disc on a Jabiru engine hardly qualifies as a "substantial flywheel" and neither (as you said) does a wooden Jab prop.

 

All you have to do is strobe the motor and if you mark references (flywheel/crankcase and prop hub/crankcase) you will see things happening with crankshaft flexing as you move through the rev range.

 

And if you clamp the prop end well enough and the "flywheel" end well enough then all will be tickety-boo. The crankshaft flex will be followed exactly by the things bolted to the ends of it and the clamp bolts will happily stay clamping ...... no movement of bolted on things = no bolt breakage.

 

Interesting I suppose that (AFAIK) it's never the bolts at the prop end that break ..

 

 

[facthunter]

"Substantial" is a relative term. Most aero engines have NO flywheel. Most of the larger 6 cyl engines have weights attached to each end of the crankshaft that are allowed to move around and absorb the torsional vibrations. Radial engines have monster weights that tumble about when the engines start up and sound like the engine is coming apart ..They don't do it for nothing. It's to address the very problem I'm talking about here. Nev

 

 

[Jaba-who]

I think Nev's use of "substantial" is reasonable, even though they are, in static terms, pretty minimal, they would load up pretty heavily at high revs.

 

I'd be interested to see the bolts that have fractured and where they have fractured. Wonder if they have all fractured at the level of the flywheel/crankshaft interface (which would suggest shearing forces at the flywheel as Nev suggests.)

 

But if the fractures are occurring deep in the threaded sockets then you might think that the forces were following a different path, especially if the holes (either in the flywheel or the crankshaft) show no deformity.

 

I can only assume the reason the prop bolts don't break is the timber allows for some deforming to absorb the vibrations. But also the bolts themselves are further encased in the drive bushes where they go through the metal flange. So I guess there is a fair bit of protection of the bolts in the forward end.

 

 

[facthunter]

JH the crankshaft threads elongate ( go oval) there would then be a combination of bending and shear loads. Once movement takes place it's only a matter of time for failure to occur . The drive should take place between the faces held tightly together rather than through the bolts themselves. Same as a keyed taper. The key gives an indexed alignment but is never able to take the drive for long. The well fitted taper fit takes the load. Nev

 

 

[ianboag]

I'd be interested to see the bolts that have fractured and where they have fractured. Wonder if they have all fractured at the level of the flywheel/crankshaft interface (which would suggest shearing forces at the flywheel as Nev suggests.)

They break at the interface. Fatigue - not shear - bending back and forth - not enough clamp.

 

 

[Jaba-who]

If the fly wheel always turns in the same direction you wouldn't expect bending back and forth would you?

 

 

[facthunter]

The engine turns in "pulses" changing rotational velocity. That is the purpose of a flywheel (to even it out). It absorbs energy, and gives energy back as well, at a high frequency. Flywheels coming loose are not uncommon as they do a lot of work.. Nev

 

 

[peibo]

Hello guys

 

I joined a coupla years ago, and have been a happy camper up to now with my wonderful Jabiru JODEL. Reading JSB012-2 got my attention tho, and led me to disassemble the back end of the engine, where I discovered fretting between the "flywheel" and the VP coupling flange, and between the flange and the cshaft gear end. Now I am in the middle of the "flywheel" bolt saga on my solid lifter 2200 SN1757 with no dowel pins nor starfish, 200hrs on the Hobbs, 400+flights.

 

I am not inclined to remove the engine for the crankshaft dowel pin mod and associated starfish "flywheel" and thicker gear replacement thinking that the problem is mainly due to not enough clamp as you say.

 

Had not read your thread and intelligent analyses and approaches, but came up with a very similar approach as gregv's, except for the ARP bolts (can't find em in France- do you have a link to a web store please???), and machined hard steel 3mm X 16,5mm O.D. individual washers instead of the big washer. I also did the VP drive replacement spacer out of steel 5mm X 52mm O.D. but added qty 6, 5mm dowel holes between them on dia.44,5mm drill circle which will be reamed for ground pins driven into the reamed 5mm holes in the alloy "flywheel" and new steel spacer. The clamped sandwich will be between steel parts this way with a pinned drive of the alloy "flywheel".

 

Could you please give me your opinions, as well as the bolt link if possible, and some news of the mod you have implemented. I'm hoping to be back in the air next week.

 

Also, any comments on crank gear cracking and failure with subsequent wall thickening by Jabiru?..

 

Seems to me that root cause of all of this is perhaps prop related as Jab states in their litany of reasons. My prop fretted on the prop flange despite Belleville washers, and regular maintenance. (evidenced by brick red powder -iron oxide?-coming out between prop and flange, and fretting marks on flange. My prop maker says Jab prop bushings are too short, and end up crushing the holes in a circular direction, ovalizing them and permitting alternating circular movement evidenced by the fretting. This means the FLYWHEEL effect of the prop is incomplete, and there is a deficiency in rotational smoothing which then acts at the other end of the crank upon the friction drive of the " flywheel" AND produces circular shock loading of the gear which is driving camshaft and distributor shafts with their intrinsic inertias. Perhaps this causes pulsing rotation of the distributer rotor shafts, destroying the rotors, and requiring glueing- huh? suggested by Jabiru...

 

My prop guy suggested reinforced prop driver holes, longer prop bushings,and going from 1/4 24. Bolts to 5/16 or 8mm, and forget the Belleville washers...

 

Could this be the answer??

 

Your thoughts

 

Pierre

 

 

[Bruce Tuncks]

Hi Pierre,

 

Yep I reckon your prop was moving at each power pulse.

 

If you don't want to buy a new prop and drive-bushes, you could apply epoxy-flock to the bushes and holes in the prop and reassemble. There will be squeeze-out on the driver-flange as well. Don't allow the bolts to be glued in.

 

Then reassemble with the Belleville washers and all the correct torques. So your prop will be epoxied to the driver flange and still clamped with the bolts. It won't move anymore and so it won't put shock-waves through the crankshaft back to the flywheel. But you will need to use heat to remove the bushes one day. So do a good balance and tracking job at this time.

 

I did this years ago and with all the bolts out, (they need re-torquing at intervals) the prop can't be moved.

 

Maybe it shouldn't have been necessary, but it sure can't do any harm compared with a loose prop.

 

good luck,

 

Bruce

 

 

[peibo]

Hello Bruce

 

Thanks for your input.

 

Went ahead and got my prop fixed up. Received it and the longer bushings Saturday. The manufacturer says Jab prop flange surface area is very small. Should be 85cm2 for the power..... but calculates out to 64cm2 or so..... Hope the prop will stop fretting. If not, might try to bond it like you did, or get a new flange machined with more surface area to increase friction drive. Flywheel is assembled and torqued to 40ftlb on grease. 45 seemed high...what do you think?

 

 

[Bruce Tuncks]

Bon Jour Pierre,

 

I reckon that 40 ft-lb should be enough. There is about 25% more clamping force for the same torque with greased threads compared with dry. It is much more than the 24 ft-lb test torque in the service bulletin JSB012-2, and heaps more than the 15ft-lb on my older engine with the 1/4 inch bolts. I reckon that loctite , not on the bolts but on the mating faces couldn't do any harm and may help prevent movement starting.

 

If you wanted to see how much torque a bolt could take, you could also do a test on a spare bolt....see the thread on through-bolts where this was discussed.

 

The Jabiru prop relies on the fit of the bushes as well as the friction of the flange, so I reckon your prop manufacturer has a formula which doesn't really apply.

 

Now I need to decide what to do with my own engine. I did the check in the service bulletin, but with the loctite present, the bolts could have the problem ianboag described so well. I'm impressed with the job that you and gregv have done, but I don't know if I should do this with only 1/4 inch bolts.

 

...Bruce

 

 

[ianboag]

Bon Jour Pierre,I reckon that 40 ft-lb should be enough. There is about 25% more clamping force for the same torque with greased threads compared with dry. It is much more than the 24 ft-lb test torque in the service bulletin JSB012-2, and heaps more than the 15ft-lb on my older engine with the 1/4 inch bolts.

.Bruce

For what it's worth, the expert on our field (who has done quite a bit of Jab research) gets a bit worried about whether that sort of torque level would run the risk of bending the treads in the crank. I did some sums that suggested it would not. Enough other people have reefed them up the 30 (ft-lb - lubed even) and higher with no ill effects that it sems OK.

 

Not for 1/4" bolts though - one would expect to "derate" the figures by the ratio of x-sectional areas - so (1/4) divided by (5/16) and squared is 16/25 = 0.65. So whatever feels OK for the 5/16" bolt torque-wise - 65% of that should work for the 1/4" ones. That will give the same STRESS in both bolts, but of course the bigger ones will have more clamp force because force=(stress) x (x-sectional area).

 

 

[facthunter]

Small point on that logic Ian. The actual cross sectional area is the root diameter of the threads, not the nominal diameter of the bolt. It's easy to check whether the bolt has stretched beyond its elastic limit. Measure it ( the overall length) and torque it and remove and remeasure. Some modern applications torque to permanent stretch. ( single use bolts). I can't see the logic in that but there you are.. Nev

 

 

[ianboag]

True. A conservative version of the area ratio (at thread root) might be diameters of 3/16 and 1/4. That makes the area ratio (3/4) squared ... = 0.5 near as matters. So put 0.5 where I had 0.65. That says 1/4" bolts torqued to 15 ft-lb have much the same bolt stress as 5/16" ones torqued to 30. The bigger bolts will have twice the clamp force of course.

 

 

[facthunter]

Yes , the smaller ones are not very capable of exerting much force. Oil gets to all of this area doesn't it?( From the inside). That makes it harder to stay in one position relying on a clamp force. Nev