Topics

ThroughBolts.


jabcamit@...
 

 

 

I'm going to strip, upgrade and rebuild my early big-fin engine before running it. I'm making a list of upgrades to incorporate.

 

If my through-bolts have undersize threads (nuts rattle on thread), I'll replace them (probably with ARP through bolts if I can get suitable ones). If the bolts are OK, I 'm thinking about getting the fat central section ground down (shortened or even removed) to make them more stretchy (and so maintain more constant pressure between the crankcase halves).

 

I think Jabiru identified through-bolt resonance as a failure mechanism in a report to CASA (in response to them imposing flight restrictions due to through bolt failures?). I can't visualise any circumstances where resonance could cause through-bolt failure, does anyone have ideas as to how this might happen?

 

BobP


Mick O'Connor
 

Bob,
What’s your serial number ?
I’ve just refitted my pistons/cylinders and replaced all through bolts in my late solid lifter with new items from Jab. They are same thread but are longer to accommodate bigger 12 pt ARP nuts and have a rubber damper ring at the centre of the bolt.

Mick


On 13 Feb 2021, at 13:10, jabcamit@... wrote:

 

 

I'm going to strip, upgrade and rebuild my early big-fin engine before running it. I'm making a list of upgrades to incorporate.

 

If my through-bolts have undersize threads (nuts rattle on thread), I'll replace them (probably with ARP through bolts if I can get suitable ones). If the bolts are OK, I 'm thinking about getting the fat central section ground down (shortened or even removed) to make them more stretchy (and so maintain more constant pressure between the crankcase halves).

 

I think Jabiru identified through-bolt resonance as a failure mechanism in a report to CASA (in response to them imposing flight restrictions due to through bolt failures?). I can't visualise any circumstances where resonance could cause through-bolt failure, does anyone have ideas as to how this might happen?

 

BobP


bob noffs
 

i have a head from a big fin 3300. it will fit any position. just sayin' if you run into that. 
 bob noffs

On Sat, Feb 13, 2021 at 7:10 AM <jabcamit@...> wrote:

 

 

I'm going to strip, upgrade and rebuild my early big-fin engine before running it. I'm making a list of upgrades to incorporate.

 

If my through-bolts have undersize threads (nuts rattle on thread), I'll replace them (probably with ARP through bolts if I can get suitable ones). If the bolts are OK, I 'm thinking about getting the fat central section ground down (shortened or even removed) to make them more stretchy (and so maintain more constant pressure between the crankcase halves).

 

I think Jabiru identified through-bolt resonance as a failure mechanism in a report to CASA (in response to them imposing flight restrictions due to through bolt failures?). I can't visualise any circumstances where resonance could cause through-bolt failure, does anyone have ideas as to how this might happen?

 

BobP


Raymond Henrie
 

ARP would not have through bolts, they are specially made by Jabiru for their engine.  There is a through bolt required replacement that has a new design to combat breaking.  JabiruUS had them.  If memory serves me they were about $35 each.  The nuts are available through ARP but need modification as required to clear the cylinder fillet.  


Mick O'Connor
 

No necessarily.  My ARP nuts as supplied by a Jab were not modified 
However, I had to grind a 45 degree angle on the bottom of the supplied washers to clear the fillet. 
Mick

On 13 Feb 2021, at 14:19, Raymond Henrie via groups.io <original_dr_doright@...> wrote:

ARP would not have through bolts, they are specially made by Jabiru for their engine.  There is a through bolt required replacement that has a new design to combat breaking.  JabiruUS had them.  If memory serves me they were about $35 each.  The nuts are available through ARP but need modification as required to clear the cylinder fillet.  


David Amsler
 

Bob,
Ian's advice was to never reuse the old 3/8" through bolts.  If the original nuts were of the distorted thread locking type, I would strongly recommend not re-using the bolts.   I think the current Jabiru replacement through bolts now have a shortened thick section in center, but would need to verify.  If doing total teardown, why not just put in the 7/16" bolts?

David A.

On Sat, Feb 13, 2021 at 8:10 AM <jabcamit@...> wrote:

 

 

I'm going to strip, upgrade and rebuild my early big-fin engine before running it. I'm making a list of upgrades to incorporate.

 

If my through-bolts have undersize threads (nuts rattle on thread), I'll replace them (probably with ARP through bolts if I can get suitable ones). If the bolts are OK, I 'm thinking about getting the fat central section ground down (shortened or even removed) to make them more stretchy (and so maintain more constant pressure between the crankcase halves).

 

I think Jabiru identified through-bolt resonance as a failure mechanism in a report to CASA (in response to them imposing flight restrictions due to through bolt failures?). I can't visualise any circumstances where resonance could cause through-bolt failure, does anyone have ideas as to how this might happen?

 

BobP


jabcamit@...
 

Hi Mick,
SN: is 7xx (can't remember exact number).  Its one of the  first engines with big (wide) finned heads and the bigger carb with swept plenum. I should have asked before you fitted them - don't suppose you know the profile  (lengths and widths of the fat and thin bits) of the  new through-bolts ? or have a photo? 

I have a used but good, and a  failed (sheared) throughbolt that Clive kindly sent to me.  Even without being tensioned (torqued up) its resonant frequency is several times higher than the engine's 6000 per minute explosions. Also it will withstand a bending force (load in the centre  supported both ends) of at least 80kg without permanent distortion, so I really can't see how  resonance relates to through bolt failure.  FWIW the  bolt  failed part way along the threaded section (I guess where the nut thread applies tension) not where the thread meets the waisted section - even though it looks like the  start of the thread might act as a stress riser on the waisted section.


Mick what is your SN? and do you have a list of any mods you may done during your rebuild? (grooves to lubricate the camshaft bearing points or whatever).  

Bob


jabcamit@...
 

Thanks for that infomation Raymond. 

I paddled through the ARP website last time I loooked at through-bolt issues and I was impressed their attention to design for maximum strength, reliability (avoiding stress risers), and tension control (specification of purpose thread lubricants to better control the torque-tension relationship).  Jabiru have yet to grasp this basic principal, despite throughbolt failures resulting in CASA imposing flight restriction on their engines, so I really don't have much (any) faith in Jabirus engineering abilities.

 I'm also concerned by the Jabiru "radius the underside of the nuts" solution to prevent nut-fillet fouling. So my plan is to fit a  standard (ARP?) nut but with a thick hardened steel washer under the nut that has a flat filed on one side to ensure good clearance from the fillet.  

I'm curious as to what (if any) usefull  function the fat central section of the through-bolts performs? If no useful function then  a full length waisted bolt will be better (more stretchy so more constant clamping forces) and I guess ARP will have an off the shelf bolt that does the job. 

BobP


David Amsler
 

Bob,

From my notes on bolts from a few years ago:

Old style 3/8” bolts:  Thread root Dia. 0.324”;    Pitch Dia. 0.348”;  Necked portion: 8.15 mm (0.321”)

New style 7/16” bolts: Thread root Dia. 0.376”;   Pitch Dia. 0.405”;  Necked portion: 8.50 mm (0.335”)

 

Root diameter of the threads on the 3/8” bolt is almost exactly the same as the necked down portion of the bolts.  When bolts are stretched to the yield point, the stress per sq. in of area at thread root will be the same as at necked down portion, and can be expected to yield there.  The sharp V at bottom of threads is a stress concentrator and prescription for failed bolts. When used with the original distorted shape locking nuts forces were further concentrated at the two opposing points where the nuts were tightest.   Then add in the very imprecise tensioning from Jabiru's "dry" torque specs, and the result is not a prescription for possible failure, but for probable failure. 

Only some cylinder barrels require chamfering the base of the nuts.  Make sure yours is one of those in the limited range of serial numbers before doing so.  (mine was just 3 serial numbers before they found and fixed the machining error, the notice to chamfer the nuts had been issued about 9 months prior to my purchasing that engine, and I was never specifically notified of the problem.)   In order to evenly distribute clamping forces to the cylinder flanges.  Personally see little problem from chamfering the nuts. to the degree called for.  I vaguely recall some notice not to use washers for this, but  if using washers,  I would not use a hardened washer, but a good quality AN spec washer.  FWIW: I do know Ian's bolts and nuts were 4130 steel.

David A.

On Sun, Feb 14, 2021 at 7:55 AM <jabcamit@...> wrote:
Thanks for that infomation Raymond. 

I paddled through the ARP website last time I loooked at through-bolt issues and I was impressed their attention to design for maximum strength, reliability (avoiding stress risers), and tension control (specification of purpose thread lubricants to better control the torque-tension relationship).  Jabiru have yet to grasp this basic principal, despite throughbolt failures resulting in CASA imposing flight restriction on their engines, so I really don't have much (any) faith in Jabirus engineering abilities.

 I'm also concerned by the Jabiru "radius the underside of the nuts" solution to prevent nut-fillet fouling. So my plan is to fit a  standard (ARP?) nut but with a thick hardened steel washer under the nut that has a flat filed on one side to ensure good clearance from the fillet.  

I'm curious as to what (if any) usefull  function the fat central section of the through-bolts performs? If no useful function then  a full length waisted bolt will be better (more stretchy so more constant clamping forces) and I guess ARP will have an off the shelf bolt that does the job. 

BobP


Mick O'Connor
 

Bob,
That’s exactly what I did. I ground a 45 deg chamfer on the underside of the Jab supplied through bolt washers. Checked before & after clearance with an eye glass 

Mick

On 14 Feb 2021, at 13:17, David Amsler <EnergyHawk@...> wrote:

Bob,

From my notes on bolts from a few years ago:

Old style 3/8” bolts:  Thread root Dia. 0.324”;    Pitch Dia. 0.348”;  Necked portion: 8.15 mm (0.321”)

New style 7/16” bolts: Thread root Dia. 0.376”;   Pitch Dia. 0.405”;  Necked portion: 8.50 mm (0.335”)

 

Root diameter of the threads on the 3/8” bolt is almost exactly the same as the necked down portion of the bolts.  When bolts are stretched to the yield point, the stress per sq. in of area at thread root will be the same as at necked down portion, and can be expected to yield there.  The sharp V at bottom of threads is a stress concentrator and prescription for failed bolts. When used with the original distorted shape locking nuts forces were further concentrated at the two opposing points where the nuts were tightest.   Then add in the very imprecise tensioning from Jabiru's "dry" torque specs, and the result is not a prescription for possible failure, but for probable failure. 

Only some cylinder barrels require chamfering the base of the nuts.  Make sure yours is one of those in the limited range of serial numbers before doing so.  (mine was just 3 serial numbers before they found and fixed the machining error, the notice to chamfer the nuts had been issued about 9 months prior to my purchasing that engine, and I was never specifically notified of the problem.)   In order to evenly distribute clamping forces to the cylinder flanges.  Personally see little problem from chamfering the nuts. to the degree called for.  I vaguely recall some notice not to use washers for this, but  if using washers,  I would not use a hardened washer, but a good quality AN spec washer.  FWIW: I do know Ian's bolts and nuts were 4130 steel.

David A.

On Sun, Feb 14, 2021 at 7:55 AM <jabcamit@...> wrote:
Thanks for that infomation Raymond. 

I paddled through the ARP website last time I loooked at through-bolt issues and I was impressed their attention to design for maximum strength, reliability (avoiding stress risers), and tension control (specification of purpose thread lubricants to better control the torque-tension relationship).  Jabiru have yet to grasp this basic principal, despite throughbolt failures resulting in CASA imposing flight restriction on their engines, so I really don't have much (any) faith in Jabirus engineering abilities.

 I'm also concerned by the Jabiru "radius the underside of the nuts" solution to prevent nut-fillet fouling. So my plan is to fit a  standard (ARP?) nut but with a thick hardened steel washer under the nut that has a flat filed on one side to ensure good clearance from the fillet.  

I'm curious as to what (if any) usefull  function the fat central section of the through-bolts performs? If no useful function then  a full length waisted bolt will be better (more stretchy so more constant clamping forces) and I guess ARP will have an off the shelf bolt that does the job. 

BobP


jabcamit@...
 

Sounds good (shaping washers rather than nuts)  Mick. Do you think (I speculate) that the nut - cylinder clearance might reduce when the engine is hot? Perhaps its the nuts fouling the barrel that causes some engines to go tight when hot. Perhaps this is what caused the beach marks on Clive's failed bolt (see pic in my last  posting)

As (unlike the nuts) the washers don't need to rotate, my plan is to have a big chamfer  but just on side of the washer -  where it  faces towards  (hopefully doesn't touch) the cylinder flange.  If I do I think I'll bond the washer to the cylinder flange so it can't rotate.


Bob

PS Thanks for the new versus old bolt pics and dims.


Mick O'Connor
 

Bob,
I won’t even try to speculate. You’d have carefully measure before (cold) end after (hot) clearance, not an easy task. 
Why has nobody suggested carefully grinding the flange fillet to give adequate clearance ?

Mick

On 14 Feb 2021, at 15:38, jabcamit@... wrote:

Sounds good (shaping washers rather than nuts)  Mick. Do you think (I speculate) that the nut - cylinder clearance might reduce when the engine is hot? Perhaps its the nuts fouling the barrel that causes some engines to go tight when hot. Perhaps this is what caused the beach marks on Clive's failed bolt (see pic in my last  posting)

As (unlike the nuts) the washers don't need to rotate, my plan is to have a big chamfer  but just on side of the washer -  where it  faces towards  (hopefully doesn't touch) the cylinder flange.  If I do I think I'll bond the washer to the cylinder flange so it can't rotate.


Bob

PS Thanks for the new versus old bolt pics and dims.


jabcamit@...
 

I think locally filing away the fillet would cause a stress riser between cylinder and flange  (don't know whether big enough to cause a significant failure risk)  but a washer with a flat on one side gives the same effect with no downside that I can think of (as long as the bolts are long enough).  A perfect solution might be a  plain washer with a  flat on one side washers then suitably rated Nordloc X washers (Belville Nordlocs), to maintain bolt tension.

I remember (or perhaps I dreamed) that for a short period Jabiru tried an undercut fillet (U shaped groove)  but that raises the stress even more and the change was quickly abandoned.


glen english LIST
 

My 3300#2575 thru bolts have the bulge at one end, not in the middle.

I cannot see that on any of the assembly drawings.

Alas when I pulled them out, I did not write down whether the bulge was at the even or odd cylinder side of the engine, I made a mental note only .... so I fogot.

I think it was the odd # side.

Does anyone know ?

-glen


glen english LIST
 

I mean the even side, think it was 2,4,6 side.

On 2/15/2021 12:14 PM, glen english LIST wrote:
My 3300#2575 thru bolts have the bulge at one  end, not in the middle.
I cannot see that on any of the assembly drawings.
Alas when I pulled them out, I did not write down whether the bulge  was at the even or odd cylinder side of the engine, I made a mental note only .... so I fogot.
I think it was the odd # side.
Does anyone know ?
-glen


jabcamit@...
 

Hi David, and many thanks for all that info. Your suggestion to replace through-bolts when an engine is rebuilt makes perfect sense.

 

From the dimensions you gave me, Jabiru 5/16” bolts are around 8% less stretchy than the 3/8” originals due to the fatter waisted section. Insufficient stretch seems to be one of the main contributors to bolt/crankcase failure, so thinner through bolts will be more reliable (subject to there being no stress risers where the threaded section meets the waisted section).

 

Clive's failed through-bolt occurred well into the threaded section ( not where the end thread might cut into the waisted section creating a stress riser), suggesting there's something else wrong with the design. I suspect nut to cylinder interference (perhaps just when the cylinder is very hot) may be the cause, Obviously the original 3/8 thread gives more cylinder-nut clearance than Jabiru's 5/16 solution. In either event I think through-bolts made by a specialist maker (ARP or whoever) will be much more reliable.

 

I don't think professionally made “squashing” lock nuts compromise reliability. Typically its just the first three turns on a nut thread that transfers 95% of the load onto the bolt. Clive's nuts spin down those three turns under just finger pressure, so anything further towards the end of the thread isn't going to have much effect. Locking is effected by the (free end) final turn of the thread being ovalled (I guess as a post operation after a “perfect” thread has been formed). I doubt if the loads when the bolt restores the nut's threadform will compromise reliability .

 

With regard to Jabiru's suggestion NOT to fit washers under the nuts, I don't think Jabiru have the faintest understanding of how to maximize throughbolt (engine) reliability. Mick reports that his new throughbolt kit came with washers; - and they weren't radiused (but had to be to prevent fouling). Also Jabiru remain (despite repeated poking) vague with respect to whether/what thread lubrication. The attached PDF explains the fundamental principals of nut- bolt fixtures. Two observations directly relevant to the Jabiru engine are:

 

  1. when torquing a bolt 50% of the torque applied applied is lost in overcoming the friction between the nut and the item being clamped. A further 40% is lost overcoming friction between the nut and the bolt. So its only the remaining 10% of the torque applied that actually tensions the bolt. They point out that the nut undersurface friction and the nut thread friction both vary widely eg due to lubricant choice and surface roughness . If the combined effect increases friction by just 5% then for a given (torque wrench) torque, the bolt tension (what we want to meet tight limits) is reduced by 50% !!!!!!!!!!!!!!!! (5% of torque applied instead of 10%)..

 

But it gets worse. In the case of the Jabiru engine, radiussing the under- surface of the nuts (to clear the cylinder flange) will significantly reduce nut-cylinder friction because that friction has less mechanical advantage, conversely if the edge of the nut even marginally touches the cylinder flange, friction will significantly increase. Application of an effective lubricant on the thread and/or under the nut will significantly increase bolt tension for a given torque – perhaps to the point where the elastic limit is exceeded.

 

  1. the article explains how by tightening a joint with just sufficient torque to align the parts and snug the joint together, and then counting turns (nut flats) to tension the joint, the uncertainty in bolt tension can be reduced by a factor of 4 (in their example from 40% to 10%). In our case the improvement is likely to be much better.

Mick if you have any surplus old through bolts (& nuts) I'd be pleased to stretch some bolts against an uncompressible load) and then plot graphs of applied toque versus extension e bolt tension) and rotation degrees with different thread lubrications. I think it should be possible to determine a more accurate and repeatable bolt tightening regime. And even (by routing the observations through the UK light aircraft association) to get Jabiru adoption/endorsement.

 

 

 

 

 

 

I'm sure your right with regard to any end thread cutting into the waisted section acting as a stress riser. This tends to confirm (at least for Clive's failed through-bolt which failed part way along the thread) that the failure mechanism isn't a simple case of repeated stretching when close to the bolt's elastic limit. ARP throughbots are designed specifically to avoid stress risers.

 

I looked which part of a nuts thread transfers load from nut to bolt. I read that the load on each turn decreases as the “free end” of the bolt is reached with the first 3 turns collectively providing almost all of the load transfer. So I wouldn't expect a professional locking nut (free end ovalled to induce binding on the bolt after the thread has been formed) to cause any reliability issue. Also I think the forces required to restore the nut to its original perfect thread form will be trivial compared with those required to tension the joint. But I think it makes good sense to always replace through bolts when rebuilding an engine ts probably accademic as if an engine is stripped and because of the history of failed through bolts it makes sense to change them anyway.

he distribution

 

within the body of the

I think 3/8 bolts are better.  This is how I see it:

1) The waisted section should be as thin and long as possible, subject to being able to provide enough tension (results in a more constant clamping force).

2) The unavoidable design compromise is  clearance between the nut and  the cylinder (and cylinder fillet).  Bigger  nuts7/16 instead of 3/8) result in less clearance. 

I suspect  Jabiru went  to bigger bolts in desparation when they couldn't work out why bolts were failing, despite their several innecttive/ counter-productive fixes. I thik its because they didn't (perhaps still don't)) understand joint theory. (They had the same confusion with flywheel attachment - bigger and bigger bolts didn't fix the problem).  But once they  determined that bigger bolts were necessary its difficult to back-track  without admitting previous incompetence.  I think Jabiru's "resonance" claim is a "fig leaf" smoke screen  justification  to hide previous incompetence from the  regulators  (Jabiru/ Mr Stiff if this assertion is incorrect could you please explain in detail how you think the resonance failure mode works). I think they also claimed resonance was the cause of flywheel's falling off.

In my student apprentice days (Marconi UK) I was given the job of devising a fix for a radar unit battery charger that wouldn't charge if the battery went too flat. I devised 2 solutions, one solution added yet more compenents, the other reconfigured some compents, deleted several others and the circuit worked better. My supervisor made it very clear that it was only acceptable to devise a more complex solution!

BobP


Mick O'Connor
 

Bob,
Where did you get the ‘do not use washers’ from ?
Manual says you must use them and use Jab steel items.
I got my washers with my new through bolts and ARP nuts. 
Mick

On 16 Feb 2021, at 14:26, jabcamit@... wrote:

Hi David, and many thanks for all that info. Your suggestion to replace through-bolts when an engine is rebuilt makes perfect sense.

 

From the dimensions you gave me, Jabiru 5/16” bolts are around 8% less stretchy than the 3/8” originals due to the fatter waisted section. Insufficient stretch seems to be one of the main contributors to bolt/crankcase failure, so thinner through bolts will be more reliable (subject to there being no stress risers where the threaded section meets the waisted section).

 

Clive's failed through-bolt occurred well into the threaded section ( not where the end thread might cut into the waisted section creating a stress riser), suggesting there's something else wrong with the design. I suspect nut to cylinder interference (perhaps just when the cylinder is very hot) may be the cause, Obviously the original 3/8 thread gives more cylinder-nut clearance than Jabiru's 5/16 solution. In either event I think through-bolts made by a specialist maker (ARP or whoever) will be much more reliable.

 

I don't think professionally made “squashing” lock nuts compromise reliability. Typically its just the first three turns on a nut thread that transfers 95% of the load onto the bolt. Clive's nuts spin down those three turns under just finger pressure, so anything further towards the end of the thread isn't going to have much effect. Locking is effected by the (free end) final turn of the thread being ovalled (I guess as a post operation after a “perfect” thread has been formed). I doubt if the loads when the bolt restores the nut's threadform will compromise reliability .

 

With regard to Jabiru's suggestion NOT to fit washers under the nuts, I don't think Jabiru have the faintest understanding of how to maximize throughbolt (engine) reliability. Mick reports that his new throughbolt kit came with washers; - and they weren't radiused (but had to be to prevent fouling). Also Jabiru remain (despite repeated poking) vague with respect to whether/what thread lubrication. The attached PDF explains the fundamental principals of nut- bolt fixtures. Two observations directly relevant to the Jabiru engine are:

 

  1. when torquing a bolt 50% of the torque applied applied is lost in overcoming the friction between the nut and the item being clamped. A further 40% is lost overcoming friction between the nut and the bolt. So its only the remaining 10% of the torque applied that actually tensions the bolt. They point out that the nut undersurface friction and the nut thread friction both vary widely eg due to lubricant choice and surface roughness . If the combined effect increases friction by just 5% then for a given (torque wrench) torque, the bolt tension (what we want to meet tight limits) is reduced by 50% !!!!!!!!!!!!!!!! (5% of torque applied instead of 10%)..

 

But it gets worse. In the case of the Jabiru engine, radiussing the under- surface of the nuts (to clear the cylinder flange) will significantly reduce nut-cylinder friction because that friction has less mechanical advantage, conversely if the edge of the nut even marginally touches the cylinder flange, friction will significantly increase. Application of an effective lubricant on the thread and/or under the nut will significantly increase bolt tension for a given torque – perhaps to the point where the elastic limit is exceeded.

 

  1. the article explains how by tightening a joint with just sufficient torque to align the parts and snug the joint together, and then counting turns (nut flats) to tension the joint, the uncertainty in bolt tension can be reduced by a factor of 4 (in their example from 40% to 10%). In our case the improvement is likely to be much better.

Mick if you have any surplus old through bolts (& nuts) I'd be pleased to stretch some bolts against an uncompressible load) and then plot graphs of applied toque versus extension e bolt tension) and rotation degrees with different thread lubrications. I think it should be possible to determine a more accurate and repeatable bolt tightening regime. And even (by routing the observations through the UK light aircraft association) to get Jabiru adoption/endorsement.

 

 

 

 

 

 

I'm sure your right with regard to any end thread cutting into the waisted section acting as a stress riser. This tends to confirm (at least for Clive's failed through-bolt which failed part way along the thread) that the failure mechanism isn't a simple case of repeated stretching when close to the bolt's elastic limit. ARP throughbots are designed specifically to avoid stress risers.

 

I looked which part of a nuts thread transfers load from nut to bolt. I read that the load on each turn decreases as the “free end” of the bolt is reached with the first 3 turns collectively providing almost all of the load transfer. So I wouldn't expect a professional locking nut (free end ovalled to induce binding on the bolt after the thread has been formed) to cause any reliability issue. Also I think the forces required to restore the nut to its original perfect thread form will be trivial compared with those required to tension the joint. But I think it makes good sense to always replace through bolts when rebuilding an engine ts probably accademic as if an engine is stripped and because of the history of failed through bolts it makes sense to change them anyway.

he distribution

 

within the body of the

I think 3/8 bolts are better.  This is how I see it:

1) The waisted section should be as thin and long as possible, subject to being able to provide enough tension (results in a more constant clamping force).

2) The unavoidable design compromise is  clearance between the nut and  the cylinder (and cylinder fillet).  Bigger  nuts7/16 instead of 3/8) result in less clearance. 

I suspect  Jabiru went  to bigger bolts in desparation when they couldn't work out why bolts were failing, despite their several innecttive/ counter-productive fixes. I thik its because they didn't (perhaps still don't)) understand joint theory. (They had the same confusion with flywheel attachment - bigger and bigger bolts didn't fix the problem).  But once they  determined that bigger bolts were necessary its difficult to back-track  without admitting previous incompetence.  I think Jabiru's "resonance" claim is a "fig leaf" smoke screen  justification  to hide previous incompetence from the  regulators  (Jabiru/ Mr Stiff if this assertion is incorrect could you please explain in detail how you think the resonance failure mode works). I think they also claimed resonance was the cause of flywheel's falling off.

In my student apprentice days (Marconi UK) I was given the job of devising a fix for a radar unit battery charger that wouldn't charge if the battery went too flat. I devised 2 solutions, one solution added yet more compenents, the other reconfigured some compents, deleted several others and the circuit worked better. My supervisor made it very clear that it was only acceptable to devise a more complex solution!

BobP

<engineering fundamentals.pdf>


David Amsler
 

Bob,

Assuming you meant 7/16" rather than 5/16" in opening paragraph, we are in basic agreement but would add the following to consider.

The problem with the MS21042 nuts, is that they are tight on two sides, this concentrates the stress at those two point, and the weak part of the assembly is the root of those threads, thus the concentrating of stress to a smaller area makes it even easier to start a crack at the root of the last thread in the nut, and once it starts to crack. ----

I think, but am not sure that CAMit's 7/16" bolts have a shorter fat section in the middle, more like the new Jabiru bolts, so ability to stretch may be about the same.

One potential issue I see with washers chamfered in just one spot is,  What keeps the washer from turning so that part of the un-chamfered part contacts and tries to climb the fillet on cylinder base.

David A.

On Tue, Feb 16, 2021 at 9:26 AM <jabcamit@...> wrote:

Hi David, and many thanks for all that info. Your suggestion to replace through-bolts when an engine is rebuilt makes perfect sense.

 

From the dimensions you gave me, Jabiru 5/16” bolts are around 8% less stretchy than the 3/8” originals due to the fatter waisted section. Insufficient stretch seems to be one of the main contributors to bolt/crankcase failure, so thinner through bolts will be more reliable (subject to there being no stress risers where the threaded section meets the waisted section).

 

Clive's failed through-bolt occurred well into the threaded section ( not where the end thread might cut into the waisted section creating a stress riser), suggesting there's something else wrong with the design. I suspect nut to cylinder interference (perhaps just when the cylinder is very hot) may be the cause, Obviously the original 3/8 thread gives more cylinder-nut clearance than Jabiru's 5/16 solution. In either event I think through-bolts made by a specialist maker (ARP or whoever) will be much more reliable.

 

I don't think professionally made “squashing” lock nuts compromise reliability. Typically its just the first three turns on a nut thread that transfers 95% of the load onto the bolt. Clive's nuts spin down those three turns under just finger pressure, so anything further towards the end of the thread isn't going to have much effect. Locking is effected by the (free end) final turn of the thread being ovalled (I guess as a post operation after a “perfect” thread has been formed). I doubt if the loads when the bolt restores the nut's threadform will compromise reliability .

 

With regard to Jabiru's suggestion NOT to fit washers under the nuts, I don't think Jabiru have the faintest understanding of how to maximize throughbolt (engine) reliability. Mick reports that his new throughbolt kit came with washers; - and they weren't radiused (but had to be to prevent fouling). Also Jabiru remain (despite repeated poking) vague with respect to whether/what thread lubrication. The attached PDF explains the fundamental principals of nut- bolt fixtures. Two observations directly relevant to the Jabiru engine are:

 

  1. when torquing a bolt 50% of the torque applied applied is lost in overcoming the friction between the nut and the item being clamped. A further 40% is lost overcoming friction between the nut and the bolt. So its only the remaining 10% of the torque applied that actually tensions the bolt. They point out that the nut undersurface friction and the nut thread friction both vary widely eg due to lubricant choice and surface roughness . If the combined effect increases friction by just 5% then for a given (torque wrench) torque, the bolt tension (what we want to meet tight limits) is reduced by 50% !!!!!!!!!!!!!!!! (5% of torque applied instead of 10%)..

 

But it gets worse. In the case of the Jabiru engine, radiussing the under- surface of the nuts (to clear the cylinder flange) will significantly reduce nut-cylinder friction because that friction has less mechanical advantage, conversely if the edge of the nut even marginally touches the cylinder flange, friction will significantly increase. Application of an effective lubricant on the thread and/or under the nut will significantly increase bolt tension for a given torque – perhaps to the point where the elastic limit is exceeded.

 

  1. the article explains how by tightening a joint with just sufficient torque to align the parts and snug the joint together, and then counting turns (nut flats) to tension the joint, the uncertainty in bolt tension can be reduced by a factor of 4 (in their example from 40% to 10%). In our case the improvement is likely to be much better.

Mick if you have any surplus old through bolts (& nuts) I'd be pleased to stretch some bolts against an uncompressible load) and then plot graphs of applied toque versus extension e bolt tension) and rotation degrees with different thread lubrications. I think it should be possible to determine a more accurate and repeatable bolt tightening regime. And even (by routing the observations through the UK light aircraft association) to get Jabiru adoption/endorsement.

 

 

 

 

 

 

I'm sure your right with regard to any end thread cutting into the waisted section acting as a stress riser. This tends to confirm (at least for Clive's failed through-bolt which failed part way along the thread) that the failure mechanism isn't a simple case of repeated stretching when close to the bolt's elastic limit. ARP throughbots are designed specifically to avoid stress risers.

 

I looked which part of a nuts thread transfers load from nut to bolt. I read that the load on each turn decreases as the “free end” of the bolt is reached with the first 3 turns collectively providing almost all of the load transfer. So I wouldn't expect a professional locking nut (free end ovalled to induce binding on the bolt after the thread has been formed) to cause any reliability issue. Also I think the forces required to restore the nut to its original perfect thread form will be trivial compared with those required to tension the joint. But I think it makes good sense to always replace through bolts when rebuilding an engine ts probably accademic as if an engine is stripped and because of the history of failed through bolts it makes sense to change them anyway.

he distribution

 

within the body of the

I think 3/8 bolts are better.  This is how I see it:

1) The waisted section should be as thin and long as possible, subject to being able to provide enough tension (results in a more constant clamping force).

2) The unavoidable design compromise is  clearance between the nut and  the cylinder (and cylinder fillet).  Bigger  nuts7/16 instead of 3/8) result in less clearance. 

I suspect  Jabiru went  to bigger bolts in desparation when they couldn't work out why bolts were failing, despite their several innecttive/ counter-productive fixes. I thik its because they didn't (perhaps still don't)) understand joint theory. (They had the same confusion with flywheel attachment - bigger and bigger bolts didn't fix the problem).  But once they  determined that bigger bolts were necessary its difficult to back-track  without admitting previous incompetence.  I think Jabiru's "resonance" claim is a "fig leaf" smoke screen  justification  to hide previous incompetence from the  regulators  (Jabiru/ Mr Stiff if this assertion is incorrect could you please explain in detail how you think the resonance failure mode works). I think they also claimed resonance was the cause of flywheel's falling off.

In my student apprentice days (Marconi UK) I was given the job of devising a fix for a radar unit battery charger that wouldn't charge if the battery went too flat. I devised 2 solutions, one solution added yet more compenents, the other reconfigured some compents, deleted several others and the circuit worked better. My supervisor made it very clear that it was only acceptable to devise a more complex solution!

BobP


David Amsler
 

I was told back in the dark ages when the service letter to radius the base of the nuts when used on those cylinders with improperly radiused base flanges first came out, to not use washers, but that was in the very early stages, before the longer bolts were available which might have been the reason,  Having a CAMit engine, have not followed all of Jabiru's twists and turns on the issue since.

David A.

On Tue, Feb 16, 2021 at 10:20 AM Mick O'Connor via groups.io <mickoconnor909=btinternet.com@groups.io> wrote:
Bob,
Where did you get the ‘do not use washers’ from ?
Manual says you must use them and use Jab steel items.
I got my washers with my new through bolts and ARP nuts. 
Mick

On 16 Feb 2021, at 14:26, jabcamit@... wrote:

Hi David, and many thanks for all that info. Your suggestion to replace through-bolts when an engine is rebuilt makes perfect sense.

 

From the dimensions you gave me, Jabiru 5/16” bolts are around 8% less stretchy than the 3/8” originals due to the fatter waisted section. Insufficient stretch seems to be one of the main contributors to bolt/crankcase failure, so thinner through bolts will be more reliable (subject to there being no stress risers where the threaded section meets the waisted section).

 

Clive's failed through-bolt occurred well into the threaded section ( not where the end thread might cut into the waisted section creating a stress riser), suggesting there's something else wrong with the design. I suspect nut to cylinder interference (perhaps just when the cylinder is very hot) may be the cause, Obviously the original 3/8 thread gives more cylinder-nut clearance than Jabiru's 5/16 solution. In either event I think through-bolts made by a specialist maker (ARP or whoever) will be much more reliable.

 

I don't think professionally made “squashing” lock nuts compromise reliability. Typically its just the first three turns on a nut thread that transfers 95% of the load onto the bolt. Clive's nuts spin down those three turns under just finger pressure, so anything further towards the end of the thread isn't going to have much effect. Locking is effected by the (free end) final turn of the thread being ovalled (I guess as a post operation after a “perfect” thread has been formed). I doubt if the loads when the bolt restores the nut's threadform will compromise reliability .

 

With regard to Jabiru's suggestion NOT to fit washers under the nuts, I don't think Jabiru have the faintest understanding of how to maximize throughbolt (engine) reliability. Mick reports that his new throughbolt kit came with washers; - and they weren't radiused (but had to be to prevent fouling). Also Jabiru remain (despite repeated poking) vague with respect to whether/what thread lubrication. The attached PDF explains the fundamental principals of nut- bolt fixtures. Two observations directly relevant to the Jabiru engine are:

 

  1. when torquing a bolt 50% of the torque applied applied is lost in overcoming the friction between the nut and the item being clamped. A further 40% is lost overcoming friction between the nut and the bolt. So its only the remaining 10% of the torque applied that actually tensions the bolt. They point out that the nut undersurface friction and the nut thread friction both vary widely eg due to lubricant choice and surface roughness . If the combined effect increases friction by just 5% then for a given (torque wrench) torque, the bolt tension (what we want to meet tight limits) is reduced by 50% !!!!!!!!!!!!!!!! (5% of torque applied instead of 10%)..

 

But it gets worse. In the case of the Jabiru engine, radiussing the under- surface of the nuts (to clear the cylinder flange) will significantly reduce nut-cylinder friction because that friction has less mechanical advantage, conversely if the edge of the nut even marginally touches the cylinder flange, friction will significantly increase. Application of an effective lubricant on the thread and/or under the nut will significantly increase bolt tension for a given torque – perhaps to the point where the elastic limit is exceeded.

 

  1. the article explains how by tightening a joint with just sufficient torque to align the parts and snug the joint together, and then counting turns (nut flats) to tension the joint, the uncertainty in bolt tension can be reduced by a factor of 4 (in their example from 40% to 10%). In our case the improvement is likely to be much better.

Mick if you have any surplus old through bolts (& nuts) I'd be pleased to stretch some bolts against an uncompressible load) and then plot graphs of applied toque versus extension e bolt tension) and rotation degrees with different thread lubrications. I think it should be possible to determine a more accurate and repeatable bolt tightening regime. And even (by routing the observations through the UK light aircraft association) to get Jabiru adoption/endorsement.

 

 

 

 

 

 

I'm sure your right with regard to any end thread cutting into the waisted section acting as a stress riser. This tends to confirm (at least for Clive's failed through-bolt which failed part way along the thread) that the failure mechanism isn't a simple case of repeated stretching when close to the bolt's elastic limit. ARP throughbots are designed specifically to avoid stress risers.

 

I looked which part of a nuts thread transfers load from nut to bolt. I read that the load on each turn decreases as the “free end” of the bolt is reached with the first 3 turns collectively providing almost all of the load transfer. So I wouldn't expect a professional locking nut (free end ovalled to induce binding on the bolt after the thread has been formed) to cause any reliability issue. Also I think the forces required to restore the nut to its original perfect thread form will be trivial compared with those required to tension the joint. But I think it makes good sense to always replace through bolts when rebuilding an engine ts probably accademic as if an engine is stripped and because of the history of failed through bolts it makes sense to change them anyway.

he distribution

 

within the body of the

I think 3/8 bolts are better.  This is how I see it:

1) The waisted section should be as thin and long as possible, subject to being able to provide enough tension (results in a more constant clamping force).

2) The unavoidable design compromise is  clearance between the nut and  the cylinder (and cylinder fillet).  Bigger  nuts7/16 instead of 3/8) result in less clearance. 

I suspect  Jabiru went  to bigger bolts in desparation when they couldn't work out why bolts were failing, despite their several innecttive/ counter-productive fixes. I thik its because they didn't (perhaps still don't)) understand joint theory. (They had the same confusion with flywheel attachment - bigger and bigger bolts didn't fix the problem).  But once they  determined that bigger bolts were necessary its difficult to back-track  without admitting previous incompetence.  I think Jabiru's "resonance" claim is a "fig leaf" smoke screen  justification  to hide previous incompetence from the  regulators  (Jabiru/ Mr Stiff if this assertion is incorrect could you please explain in detail how you think the resonance failure mode works). I think they also claimed resonance was the cause of flywheel's falling off.

In my student apprentice days (Marconi UK) I was given the job of devising a fix for a radar unit battery charger that wouldn't charge if the battery went too flat. I devised 2 solutions, one solution added yet more compenents, the other reconfigured some compents, deleted several others and the circuit worked better. My supervisor made it very clear that it was only acceptable to devise a more complex solution!

BobP

<engineering fundamentals.pdf>


David Amsler
 

Bob,

Thanks for the Engineering Fundamentals for Threaded Fasteners.  Is that an insomnia cure, or to keep me awake at night worrying about what will let go next :-)

David A.

On Tue, Feb 16, 2021 at 10:35 AM David Amsler via groups.io <EnergyHawk=gmail.com@groups.io> wrote:
Bob,

Assuming you meant 7/16" rather than 5/16" in opening paragraph, we are in basic agreement but would add the following to consider.

The problem with the MS21042 nuts, is that they are tight on two sides, this concentrates the stress at those two point, and the weak part of the assembly is the root of those threads, thus the concentrating of stress to a smaller area makes it even easier to start a crack at the root of the last thread in the nut, and once it starts to crack. ----

I think, but am not sure that CAMit's 7/16" bolts have a shorter fat section in the middle, more like the new Jabiru bolts, so ability to stretch may be about the same.

One potential issue I see with washers chamfered in just one spot is,  What keeps the washer from turning so that part of the un-chamfered part contacts and tries to climb the fillet on cylinder base.

David A.

On Tue, Feb 16, 2021 at 9:26 AM <jabcamit@...> wrote:

Hi David, and many thanks for all that info. Your suggestion to replace through-bolts when an engine is rebuilt makes perfect sense.

 

From the dimensions you gave me, Jabiru 5/16” bolts are around 8% less stretchy than the 3/8” originals due to the fatter waisted section. Insufficient stretch seems to be one of the main contributors to bolt/crankcase failure, so thinner through bolts will be more reliable (subject to there being no stress risers where the threaded section meets the waisted section).

 

Clive's failed through-bolt occurred well into the threaded section ( not where the end thread might cut into the waisted section creating a stress riser), suggesting there's something else wrong with the design. I suspect nut to cylinder interference (perhaps just when the cylinder is very hot) may be the cause, Obviously the original 3/8 thread gives more cylinder-nut clearance than Jabiru's 5/16 solution. In either event I think through-bolts made by a specialist maker (ARP or whoever) will be much more reliable.

 

I don't think professionally made “squashing” lock nuts compromise reliability. Typically its just the first three turns on a nut thread that transfers 95% of the load onto the bolt. Clive's nuts spin down those three turns under just finger pressure, so anything further towards the end of the thread isn't going to have much effect. Locking is effected by the (free end) final turn of the thread being ovalled (I guess as a post operation after a “perfect” thread has been formed). I doubt if the loads when the bolt restores the nut's threadform will compromise reliability .

 

With regard to Jabiru's suggestion NOT to fit washers under the nuts, I don't think Jabiru have the faintest understanding of how to maximize throughbolt (engine) reliability. Mick reports that his new throughbolt kit came with washers; - and they weren't radiused (but had to be to prevent fouling). Also Jabiru remain (despite repeated poking) vague with respect to whether/what thread lubrication. The attached PDF explains the fundamental principals of nut- bolt fixtures. Two observations directly relevant to the Jabiru engine are:

 

  1. when torquing a bolt 50% of the torque applied applied is lost in overcoming the friction between the nut and the item being clamped. A further 40% is lost overcoming friction between the nut and the bolt. So its only the remaining 10% of the torque applied that actually tensions the bolt. They point out that the nut undersurface friction and the nut thread friction both vary widely eg due to lubricant choice and surface roughness . If the combined effect increases friction by just 5% then for a given (torque wrench) torque, the bolt tension (what we want to meet tight limits) is reduced by 50% !!!!!!!!!!!!!!!! (5% of torque applied instead of 10%)..

 

But it gets worse. In the case of the Jabiru engine, radiussing the under- surface of the nuts (to clear the cylinder flange) will significantly reduce nut-cylinder friction because that friction has less mechanical advantage, conversely if the edge of the nut even marginally touches the cylinder flange, friction will significantly increase. Application of an effective lubricant on the thread and/or under the nut will significantly increase bolt tension for a given torque – perhaps to the point where the elastic limit is exceeded.

 

  1. the article explains how by tightening a joint with just sufficient torque to align the parts and snug the joint together, and then counting turns (nut flats) to tension the joint, the uncertainty in bolt tension can be reduced by a factor of 4 (in their example from 40% to 10%). In our case the improvement is likely to be much better.

Mick if you have any surplus old through bolts (& nuts) I'd be pleased to stretch some bolts against an uncompressible load) and then plot graphs of applied toque versus extension e bolt tension) and rotation degrees with different thread lubrications. I think it should be possible to determine a more accurate and repeatable bolt tightening regime. And even (by routing the observations through the UK light aircraft association) to get Jabiru adoption/endorsement.

 

 

 

 

 

 

I'm sure your right with regard to any end thread cutting into the waisted section acting as a stress riser. This tends to confirm (at least for Clive's failed through-bolt which failed part way along the thread) that the failure mechanism isn't a simple case of repeated stretching when close to the bolt's elastic limit. ARP throughbots are designed specifically to avoid stress risers.

 

I looked which part of a nuts thread transfers load from nut to bolt. I read that the load on each turn decreases as the “free end” of the bolt is reached with the first 3 turns collectively providing almost all of the load transfer. So I wouldn't expect a professional locking nut (free end ovalled to induce binding on the bolt after the thread has been formed) to cause any reliability issue. Also I think the forces required to restore the nut to its original perfect thread form will be trivial compared with those required to tension the joint. But I think it makes good sense to always replace through bolts when rebuilding an engine ts probably accademic as if an engine is stripped and because of the history of failed through bolts it makes sense to change them anyway.

he distribution

 

within the body of the

I think 3/8 bolts are better.  This is how I see it:

1) The waisted section should be as thin and long as possible, subject to being able to provide enough tension (results in a more constant clamping force).

2) The unavoidable design compromise is  clearance between the nut and  the cylinder (and cylinder fillet).  Bigger  nuts7/16 instead of 3/8) result in less clearance. 

I suspect  Jabiru went  to bigger bolts in desparation when they couldn't work out why bolts were failing, despite their several innecttive/ counter-productive fixes. I thik its because they didn't (perhaps still don't)) understand joint theory. (They had the same confusion with flywheel attachment - bigger and bigger bolts didn't fix the problem).  But once they  determined that bigger bolts were necessary its difficult to back-track  without admitting previous incompetence.  I think Jabiru's "resonance" claim is a "fig leaf" smoke screen  justification  to hide previous incompetence from the  regulators  (Jabiru/ Mr Stiff if this assertion is incorrect could you please explain in detail how you think the resonance failure mode works). I think they also claimed resonance was the cause of flywheel's falling off.

In my student apprentice days (Marconi UK) I was given the job of devising a fix for a radar unit battery charger that wouldn't charge if the battery went too flat. I devised 2 solutions, one solution added yet more compenents, the other reconfigured some compents, deleted several others and the circuit worked better. My supervisor made it very clear that it was only acceptable to devise a more complex solution!

BobP