Another Jab bites the dust.

[slartibartfast]

Just ignoring the crap for a second, I'd like to post something on topic and from personal experience.

 

2 weeks ago I began a 50 hourly, being 25 hours after putting in K Liner valve guide inserts. Engine time now 350 hours.

 

During the rebuild, one of the through bolt nuts stripped. I discussed (disgust?) with Jabiru and asked them to send me a new set of nuts and a new bolt.

 

They sent the original nuts. 1 week later they released the AD demanding the new nuts. Not happy.

 

When I pulled the cowl off for this service, I noticed there had been leakage at the cylinder base on the 2 front cylinders - 1 either side.

 

I checked the torque (which was the plan anyway), and one of the through bolts wouldn't torque to me. The bolt pulled through the nut on the far side. The nut on one of the studs stripped the thread too. These all torqued up to spec during the rebuild 25 hours ago, so they let go on their own. I must have been a Jabiru profit margin away from an engine failure.

 

I rang Jabiru last week to discuss, and they said they would send a new set of 12 point nuts and the longer bolts, and that there would be no charge.

 

Happy with that, I asked for express postage and paid extra so it would arrive by the weekend. This was Tuesday morning before Anzac day. They didn't turn up. They had sent them on Thursday evening. When they did arrive this week, I found I had been charged for the lot.

 

When I rang, they couldn't help me because that guy was away. He rang next day and said "sorry if I told you that - I don't remember". I may get the money back for the nuts only.

 

He also dismissed the Nowra failure saying they hadn't done the AD.

 

Thanks for all the info in this thread too. Well the helpful bits anyway. The overhaul manual advises the use of Loctite Coppermax (silicone) on the bases, but not the surfaces themselves. Here's what it says:

 

Apply a small amount of Coppermax sealant to the cylinder base O ring. Wipe excess sealant off the base face of the cylinder before fitting – the joint between cylinder and case as clean as possible with no sealants etc between. Then offer the cylinder / piston assembly up to the engine.

and later (curiously):

 

Excess sealant can now be wiped off the cylinder bases and the crankcase join – this is most easily done once the sealants have dried.

This bit is interesting too:

 

Technicians reading this must understand the difference between “bolt tension” and “nut torque”. Nut torque is the setting on the torque wrench & defines how much rotational force it takes to move the nut. Bolt tension is the actual tensile load in the bolt – the force holding the parts together. Tuning a guitar is a good illustration of the difference between bolt tension and nut torque – for a given torque on the guitar string adjustor the tone of the string might be too high or too low – in that case the musician can set the string tension directly by turning the adjuster until the string tone is right. For a through-bolt we cannot measure the tension directly and we are forced to make assumptions of the relationship between nut torque and bolt tension – in effect we are trying to set the guitar string tone indirectly by assuming that a given torque on the string adjuster will give a certain string tone. Testing has shown that assembling these parts dry apart from Loctite 620 gives good, repeatable results for bolt tension. Altering these parameters has unpredictable effects and is not recommended.

Just adding to the story. I'll be fitting new bolts, studs and nuts tomorrow - after recalibrating my torque wrench.

 

 

[Gibbo]

Ok.. I suspect it's a mixture of three differant cause's..

 

- Uneven mixtures producing differing EGT's etc.

 

- All other designs have a 'solid' gasket not a flexible silicon based type.. (call me if wrong - but my EA81 etc had a non flexible type etc etc etc)

 

- Offset little ends on the conrods - may have been changed in later designs (old 1600's did have the offset on rear pistons from memory). Pressure is not being distrubed evenly across piston surface. Something has to give..

 

Had one let go on a hire aircraft. 760hours in circuits. Top end I think had been done a couple of times.

 

 

[pudestcon]

after recalibrating my torque wrench.

How you gonna do that Slarti? Just interested is all.

 

Pud

 

 

[rankamateur]

How you gonna do that Slarti? Just interested is all.Pud

His answer would make an interesting thread, count me in!

 

 

[Tomo]

How you gonna do that Slarti? Just interested is all.Pud

Best way is to calibrate it with another known calibrated wrench.

 

Or the old fish scales trick...

 

 

[Bigglesworth]

I've only just noticed this thread, and spent a lot of time reading.

 

I come from a background of repairing small engines (lawnmowers, chainsaws, stationary engines etc), and I've seen a LOT of interesting failures.

 

I also have done a full strip and rebuild on a jab 2200 which is in my Cheetah.

 

I have had the opportunity to see machines where the components are too weak and learn the symptoms. They seal around the bolt area but not in between. A LOT of Jabiru that I have seen (including mine) exhibit this at the cylinder base. The whole sleeve is weak, but the base is ridiculous.

 

Also can't anyone else recognise that the cylinder in the picture has been loose for a few hours before the snap?

 

Look at the wear pattern around the nut area.

 

If it was a clean break you'd get in indentation matching the nut, and maybe cracks/chips around it.

 

If it was a loose bolt under little tension/movement you get the 'blackening' that I'm sure you've all seen in loose spinner/cowling bolts.

 

That looks to me like places where something has been loose and experiencing high force for a reasonably short period of time (my guess would be a few hours)

 

If a through-bolt was loose, the force on the base of the barrel become uneven and invloves a bending moment, this could lead to the crack as an over stress crack.

 

Once that happened the uneven force on the bolt would snap the head off, and also the uneven force (because of the con rod) would make the other 'top' bolt the next to go. Once they snapped the cylinder was still 'working' but less compression/less force, and the bottom ones held in to stop the cylinder snapping off and releasing hell in a handbasket.

 

I will be contacting the Chad/Andrew to see if there was evidence of wear on the other side cylinder.

 

I see it that my life might hang on working out how to stop this happening to mine.

 

Oh, and after the rebuild I have once had an issue with old fuel (98 premium) causing the engine to 'knock' on full power on climbout after take-off. I immediately backed off the power and no further issue.

 

Once above the airfield I put the plane back in that scenario to see if it would happen again, which it did.

 

I actually cleaned out the carby and a lot of other stuff before accepting it was just fuel being old. Keep the fuel new and haven't had a problem since.

 

So if through bolts snap because of detonation, I should have hit it then.

 

I blame the weak bases, bad engineering tolerances in the crankcase, IIRC there are no dowels in the cylinder to crankcase join (your chainsaw might even have them) and the bolts are loose enough.

 

Add in the silicone seal. Big time asking for trouble, The o-ring is sufficient.

 

I have seen gaskets fail and loosen joins before.

 

 

[Bigglesworth]

Just re-read Slarti's post.

 

So it confirms that the nuts can loosen by themselves, whether due to silicone of just being too weak it intuitive felt to me when I did the rebuild.

 

As for all this sh*t about torque setting, I have had inexpensive motors to play with for a long time and work out some rules on torque setting.

 

For starters the through bolts are designed for multiple torquings, so they are old school, low tech.

 

In 99% of cases that actual torque setting is not too relevant (and often wrong. I have had rules of thumb to prevent gasket failure that gave a lot better life) but that they are torqued evenly and not incorrectly stressed.

 

Once the nut has touched the next bit of tension will compress out all gaps and excess sealant (hopefully) after that it can only stretch the bolt, or compress the material. If both are of high standard, upping the torque by, say 20%, normally only results in 1/8 a turn of a nut.

 

If properly designed, a motor will have a large difference between the force needed for an effective seal, and the force at which things distort.

 

And the average lawnmower can be this forgiving. Remember, these are engineered to be cheap, lightweight, and reliablilty is not crucial.

 

An aero-engine is meant to be reliable, reliable, reliable, and light secondary. And cheap for 3rd.

 

There are 6 bolts in the head to cylinder, and metal to metal. There are no failures here. Surprise, surprise.

 

 

[pudestcon]

Or the old fish scales trick...

Please explain... Got pictures?

Pud

 

 

[slartibartfast]

How you gonna do that Slarti? Just interested is all.Pud

Well Pud, there's a good explanation here - http://www.dieseldoctor.com/messageboard/data/229.html

 

Hey where the hell are the editing options? I've got "use rich editor" turned on.

 

Anyway, a good check for 30 ft/lbs is to put the square drive in a vice with the handle horizontal, measure 12 inches out from the center of the drive, mark it on the handle, and hang 30 lbs from there. If it's out and your wrench can be adjusted, adjust it.

 

Otherwise determine the error factor, and apply it to your wrench setting.

 

Now you're torquing.

 

There's a boring video here -

 

 

 

[turboplanner]

I was eating an orange when I read Slarti's post.

 

I'm not too sure how accessible the cylinder bases are out in the field, but since the statistics seem to show that an early warning of thru bolt failure is a black residue leaking out, I would be recommending at least a weekly pre-flight involving a torch, and mirror on an extension, or a $300.00 camera probe, rather than wait for a forced landing.

 

I would be interested in Jabiru's reason for the coppermax sealant to be added to the O Ring, because an O Ring is a perfect sealant (if diameter and groove dimensions are correct) and is designed to compress to allow metal to metal contact thus providing a permanent pressure seal. Any additional, or loose, silicon would be superfluous.

 

Black residue leakage would be an indicator that the faces had come apart far enough to de-compress the O Ring, so that is consistent with what's been happening in the failures.

 

Having decided to apply the Coppermax, I think what Jab's conflicting instructions refer to, is on the inside of the O Ring, wipe away any visible residue once the O Ring is sitting in the groove. On the outside of the O Ring any residue is only going to be squeezed to the outside of the engine, and that's easier to remove when dry. And of course when you are applying silicon you make sure there is no dirt on your hand, bits of grass etc. Licking a finger will ensure the silicon doesn't stick to it.

 

The story about the difference between bolt tension and nut torque is a bit of motherhood rambling , which of course is true, but belongs in an engineering design centre with test equipment which shows that a bolt of X material and Y diameter fails at Z tension, That a nut with B thread and a face to face diameter of C fails the bolt at D TORQUE, and so the recommended TORQUE setting in the workshop manual will be D minus a very generous safety margin.

 

"assembling the parts dry apart from Loctite 620" means either using new dry studs, or making sure there is no oil on the existing studs (which will involve washing with something like Handy Andy (and no, not 20 seconds under the oxy torch).

 

To the DRY stud, then add Loctite 620 and the manufacturers safety calculation will then be met.

 

If there is oil on the stud or nut thread you will get quite a lot more turn on the nut (maybe even more than 1 turn) due to the lower coefficient of friction, before the torque wrench triggers, so the bolt will be over tensioned.

 

Also, avoid using a torque wrench with loud rock music playing, six mates arguing over stubbies or anything else distracting. With light nuts you should be as alert as you are when coming in for a landing because it is very easy to run past the wrench signal point. I've actually seen some guys reach the signal point and give the wrench an extra half turn just to be sure, and they are the ones who frequently have "bad motors".

 

 

[David Isaac]

Well said Tubz. I had to laugh at your 'tongue in cheek' comments. I have heard and seen the same thing. Must be an age thing. If you live long enough you will have seen a lot.

 

 

[Camel]

How can you retension nuts that where loctite was used ?

 

I believe anyone retensioning these nuts as described would do damage.

 

Bolts stretch or twist then break.

 

Adding that extra bit of tension is a no no as well, the tension suits the size of the bolt and the task, in other words tight is tight but what is the amount of force the bolt will exert that it is clamping, maybe enough force to break or hold it together as intended.

 

There are a few basic principles involved here, tension to correct value, use correct loctite on clean surface, do not retension when loctite is set, do not re-use bolts that have been put under tension where stretch or twist may have occurred, nuts or bolts that have had loctite used on them should be seriously consider whether to reuse as the twisting motion to undo may have weakened the bolt, I wouldn't reuse them.

 

I wish I knew for sure that the Jabiru through bolts fitted correctly were up to the Job ? that is what it's about and of interest to me unfortunately if incorrectly tensioned or retensioned they become more likely to fail.

 

 

[damkia]

How can you retension nuts that where loctite was used ?I believe anyone retensioning these nuts as described would do damage.

Bolts stretch or twist then break.

 

Adding that extra bit of tension is a no no as well, the tension suits the size of the bolt and the task, in other words tight is tight but what is the amount of force the bolt will exert that it is clamping, maybe enough force to break or hold it together as intended.

 

There are a few basic principles involved here, tension to correct value, use correct loctite on clean surface, do not retension when loctite is set, do not re-use bolts that have been put under tension where stretch or twist may have occurred, nuts or bolts that have had loctite used on them should be seriously consider whether to reuse as the twisting motion to undo may have weakened the bolt, I wouldn't reuse them.

 

I wish I knew for sure that the Jabiru through bolts fitted correctly were up to the Job ? that is what it's about and of interest to me unfortunately if incorrectly tensioned or retensioned they become more likely to fail.

On such a critical component such as the through bolt I would have thought it prudent to have a maximum number of times it can be used (torqued up) before replacement (ie replaced every 5 or so re-torques). Whilst it may not be a directive of Jabiru, it may be worthwhile approaching them to do some real life trials of periodic replacement with new bolts, even with the new larger bolt solution they have started to use. The worst that can happen is the problem is fixed.

 

The big problem I see with these bolts would seem to be that they are retorqued too often coupled with a bit of resonance along their length ("guitar stringing"), stretching them each time until the metallurgical properties of the steel change from elastic deformation t0 plastic deformation and fatigue, and they fail. This would generally be seen as the bolts not being shiny, but have a slightly grainy appearance inthe shaft, especially at the stress riser of the junction of the thread to the shaft of the bolt and under the head of the bolt.

 

 

[Tomo]

How can you retension nuts that where loctite was used ?

You're on the money I reckon Mr Camel!!

 

Testing has shown that assembling these parts dry apart from Loctite 620 gives good, repeatable results for bolt tension.

If 620 was used, you cannot re-tension the nuts, (of that size anyway) end of story! Loctite 620 is a retaining compound, high temp, high strength, can take up a 0.010" gap (from memory) and hold under high pressure. In other words on a small nut and thread it's the end of those moving without some sort of damage. I didn't realise this was used until Mr Slartipants () posted that quote from the manual. I'm almost of the thought, when people are re-torquing these nuts, it's just twisting the through bolt... if that's the case it would make pretty good sense why they are breaking. Not to mention not actually clamping the barrels together - hence small leakage showing prior to breakage.

 

 

[Guest Andys@coffs]

How can you retension nuts that where loctite was used ?I believe anyone retensioning these nuts as described would do damage.

Bolts stretch or twist then break.

 

Adding that extra bit of tension is a no no as well, the tension suits the size of the bolt and the task, in other words tight is tight but what is the amount of force the bolt will exert that it is clamping, maybe enough force to break or hold it together as intended.

 

There are a few basic principles involved here, tension to correct value, use correct loctite on clean surface, do not retension when loctite is set, do not re-use bolts that have been put under tension where stretch or twist may have occurred, nuts or bolts that have had loctite used on them should be seriously consider whether to reuse as the twisting motion to undo may have weakened the bolt, I wouldn't reuse them.

 

I wish I knew for sure that the Jabiru through bolts fitted correctly were up to the Job ? that is what it's about and of interest to me unfortunately if incorrectly tensioned or retensioned they become more likely to fail.

Camel

There is another thread within the Jabiru section about the availablility from WA of alternate through bolts. These alternates have rolled rather than machined threads (although I have read posts that claim that the assertion that the Jabiru supplied bolts are machined is wrong, and they too are rolled).

 

That said, there is some testing documented in that thread where the required torque is compared to the point of failure and if I summarise (hopefully correctly as Im not a MechEng) the point of failure for the J supplied bolts/nuts was very close to the required torque (arguably too close), annecdotally at times below the required torque, where the WA sourced alternates point of failure is significantly above the required torque. That thread also suggests that interface between J supplied bolt and thread is not correct with the bolt appearing to those in the know, to be oversized leading to early failure.

 

If you havent yet read that thread you should its every bit as useful, perhaps more than this one, and as I recall almost all posts were relevant to the issue.....unlike a fiar chunk of this one.

 

Andy

 

 

[turboplanner]

I agree with you Camel, you can't retension nuts accurately where loctite has set.

 

Where bolts have proven to be adequate, the slight increase in torque to undo the nut will not be an issue, but the threads need to be cleaned, and that may require dis-assembly.

 

This is why some people don't like to use Loctite. I got my experience in racing where twice each lap the engine was pulling hard at 10,000 rpm, and Riley's Daughter's Law applied (anything that could come off did come off), and I got the benefit of what Loctite was all about.

 

Where bolts/nuts are marginal, I throw both bolts and nuts away and use new ones. You are certainly up for the extra time of disassembly and reassembly of the motor, but that's less than a rebuild after a blow up and possibly an engine write off, or in flying, a forced landing.

 

I don't doubt that a percentage of Jab engine thru bolts are being butchered by the operators, but that leaves a substantial percentage where correct procedures are followed.

 

I can't for the life of me understand how any self respecting designer would allow a bolt breakage pint at the thread, or use a "loose" thread design where the contact was not the full length from tip to base, but I know plenty of motorcross engines where you have to be extremely careful when tensioning heads, and in this discussion, while stripped threads are certainly a symptom, I'm starting to believe the underlying cause is somewhere else.

 

 

[Tomo]

Please explain...

Slarti gave the explanation below.. (or above, whatever tickles your fancy)

 

Anyway, a good check for 30 ft/lbs is to put the square drive in a vice with the handle horizontal, measure 12 inches out from the center of the drive, mark it on the handle, and hang 30 lbs from there. If it's out and your wrench can be adjusted, adjust it.

 

Got pictures?

Well I found this on the net, and though it's not exactly a purists way to do something, it gives you an idea what I was talking about... with the scales at least.

 

 

 

[facthunter]

Tension wrenches can be miles out and they generally have a long handle so you don't have to be Arnie to put too much torque on the nut. The little extra 1/4 or more, turn people should be kept at a safe distance, from your plane (and maybe your children too).

 

What about overtorqued wheel nuts? ( Most of the time) on your car.

 

Concerning the use of Loctite on the threads. Big end nuts used to be split pinned. These days they are just torqued and they only come apart if the bolts break. Usually over revving.

 

All aero cylinder base flanges look too thin, but surely they have been engineered properly? I like to see a good radius on them. A scratch or mark on them will just about ensure a crack starts.

 

Ideally you would have more than four studs but thes coincide with the through bolt ( actually, they are all studs) so have to correspond with the main bearing support sections of the case.

 

I find the need to retension these studs a bit of a concern. There is no gasket on any of the 3 surfaces involved and they are in virtual metal to metal contact. ( Unless you use the sealant wrongly, and let it set before assembly).

 

They do have to stretch during use, at least for the thermal expansion of the aluminium case, but I would think the extent of that should not be such as to require retensioning. If the stud is designed properly it will not be stretched to near it's elastic limit ( look up Young's modulus) and will return to it's original size each time the engine cools. Any other forces to cause stretch, would have to relax the load on some faces enough to cause fretting of the mating faces.

 

Evidence of this is often found in motors that have experienced stud breakages. ( between the crankcase mating faces).

 

Does this occur before or after the stud break?

 

Most engines do not run long after the stud breaks. Do they Nev

 

 

[turboplanner]

What about overtorqued wheel nuts? ( Most of the time) on your car.I find the need to retension these studs a bit of a concern. There is no gasket on any of the 3 surfaces involved and they are in virtual metal to metal contact. ( Unless you use the sealant wrongly, and let it set before assembly).

 

( look up Young's modulus)

 

Most engines do not run long after the stud breaks. Do they Nev

Line 1

 

I still have a scar on my leg where a day out of Coopers Creek in the outback, I couldn't budge a truck wheel nut even by standing on the long spring bar, and with no civilization around to help, decided to jump on the bar, and finished up on my back when the bar sprang back, to make matters worse landing in a bindii patch.

 

Line 2,3

 

This is why I'm leaning towards a detonation type force.

 

Line 4

 

See also Young's Daughter's modulus.

 

Line 5

 

Not sure if this was a special note to yourself.

 

I've run a couple of old engines on the farm after shearing a stud

 

I've also heard of older (pre mid 1940's), so old that studs sheared or nuts stripped during repair being held on by No 8 fencing wire, which when correctly twisted by a farmer has considerably more compression strength than what we've been talking about here.

 

 

[Guest ozzie]

Most if not all aircraft maintenance facilities have master torque wrenches that are used to test the workshop torque wrenches.

 

How long since those working on these engines have had their torque wrenches checked and calibrated?

 

Question in relation to the thru bolts. Are these taken straight up to their tension values or are they stepped up in several rounds?

 

 

[turboplanner]

I'll go out on a limb here and suggest the Jab thru bolt issue has nothing to to with all your torque wrenches being out of calibration.

 

 

[facthunter]

If the studs stretch permanently at close to the recommended torque figures it could.. This matter is complicated by poor thread fits, loctite use. effect of such ( as a thread lubricant) on achieved strain in the stud.. The idea of torqueing the stud is to stretch it by an amount that will impose the required load to hold the pieces together. This is what they are trying to say in starti's post about guitar strings. Nev

 

 

[Guest ozzie]

I'll go out on a limb here and suggest the Jab thru bolt issue has nothing to to with all your torque wrenches being out of calibration.

It eliminates one questionable factor and sets a positive base line to move on from.

So the question still remains. How long, if ever, since your torque wrench has been calibrated and tagged OK?

 

 

[turboplanner]

Mine had its weekly calibration (no I only made that up Oz) half an hour ago.

 

Set to trigger at 20 lbs/ft

 

I attached a scale 16" exactly out from the drive centreline

 

The scale showed 12.25 lbs consistently to trigger the torque wrench

 

Someone like to work out how accurate the Torque wrench is?

 

 

[damkia]

Mine had its weekly calibration (no I only made that up Oz) half an hour ago.Set to trigger at 20 lbs/ft

 

I attached a scale 16" exactly out from the drive centreline

 

The scale showed 12.25 lbs consistently to trigger the torque wrench

 

Someone like to work out how accurate the Torque wrench is?

16" = 4/3ft

 

@12.25lb

 

So for 1 ft it would be 12.25 x (4/3) = 16.333 ft.lb IN THIS CASE (easy way of remembering is to look at what a "ft.lb" represents - how many pounds at how many feet, or simply remember that in maths a "." can also mean "multiplied by")

 

Be aware that there are two types of main errors - zero offset, and scaling. Zero offset means that the scales do not start at a true zero point ( ie a needle with or two either side of zero). Scaling is is a linear inaccuracy (ie as if you were measuring things on the moon - less gravity affects all readings by an equal percentage).

 

The best idea if you are serious is to do three series of three different settings on the torque wrench, ie, 10 ft.lb (x3), 30 ft.lb (x3) and 50 ft.lb (x3) and post up the figures you get from your scales. From this you can draw a "line of best fit" on a piece of graph paper and extrapolate all points in between.

 

There are more obscure errors but not really appropriate in this discussion (parallax errors in viewing scale from an angle, etc), and of course the obvious of the accuracy of the scales you are pulling on (with their zero offset and scaling errors...).