Can I reuse ARP rod bolts?
 

My 964 3.6 engine had a fresh rebuild done and then a rod bearing failed after several track days at DE events. Going to attempt rebuild myself and have many questions coming. I had new ARP rod bolts used but I don't have any measurements of sizes of the bolts and don't know how much they were torqued or stretched. Advice? thanks.

I just did a 3.2, uses same rods, trying to remember, I took several steps up to 35#ft.
They're supposed to be able to reuse but I don't know anyone that has, I know I have a set in the drawer.

Really? I thought the aftermarket rod bolts were safe to reuse, similar to bolts in aftermarket rods like Pauter and Carrillo. Maybe with new nuts at least. Not a good idea?

these were not in the engine long, like I said maybe 4 short track DE days and a few dyno runs. I'll buy new if its recommended but Ive heard different recommendations on reuse. I'll buy new if I have to but I want to be prudent with expenditures.

ARP rod bolts should be elastic and good to be re-used unless they have been overtightened.

The standard torque sequence for an M9 Porsche rod bolt is to torque to 11 lbsft and then 2 x 90degree turns using lightly oiled heads and threads.

ARP recommend that the bolts are tightened using either a stretch gauge or torqued to 40 lbsft using ARP Ultra Torque assembly lubricant.

This will produce a much higher preload than 40 lbsft using a light oil.

The stretch recommended by ARP is 0.010" to 0.0105".

ARP recommend that bolts should be measured prior to installation and that if they have permanently deformed by more than 0.001" after use they should be replaced.

This is a very conservative approach and if you don't have the original data makes it difficult to know what to do next and most of us would buy new bolts. Good news for ARP but costly and maybe there is no need for replacement.

To explain what is happening is a little tricky but there is a simple way to assess if the bolts are still fit for purpose.

When a rod bolt is installed at the correct preload then the load applied when the engine is operating will always be below the level of this preload.

Due to this bolts will never be subjected to any fatigue loading and can be simply selected from a 'strength of materials' criteria which is much easier than trying to consider fatigue ratings.

This is an important factor as it also means that the bolt will not elongate further in service than the length applied during tightening.

If, therefore, a bolt has permanently deformed it is due to it having been overtightened during assembly.

Where does that take us next?

If the bolt has permanently deformed it may still be capable of operating safely and elastically we just need an effective way to make a decision.

It is also important to know that the majority of engineering steels will work harden when plastically deformed.

http://i197.photobucket.com/albums/a...ane/Yield2.jpg

This means that when a material yields (Ơe) it will start to increase in strength following what is generally known as a flow curve and this curve is a material characteristic.

If we deform the material significantly to Ơ1 and then unload the sample or component then it will have deformed slightly (Ɛp).

If you then reload the part it will now not yield until it reaches Ơ1 again which will effectively become its new yield point and the bolt will still be elastic and operate normally. All of the plastic deformation at this point will be uniform and there will be a small reduction in diameter to compensate for the small increase in length. These changes should be insignificant.

This is just fine until we reach Ơb which is normally referred to as the Ultimate Tensile Strength of a material.

At this point the components ability to sustain a load begins to fall. The increase in strength is insufficient to compensate for the reduction in area.

Deformation becomes non-uniform the bolt would start to 'neck' and then break.

The question is what is a safe level of uniform deformation and when have we become too close to the UTS of the material to take a risk.

The difference between an elastically loaded bolt at its correct operating preload and yield stress may only be 10-15% and in the case of a rod bolt the closer to the yield stress the better.

IMHO a 0.001" elongation od an M9 rod bolt is a plastic strain of around 0.0005 which is very low (Strain is always a non-dimensional number) and I wouldn't be concerned.

If you have no history of the bolt I would use the following procedure:

1. Accurately measure the length of the bolt in the unloaded condition.

Use a stretch gauge if you have one if not an accurate micrometer. A vernier is just not good enough.

I would as a matter of routine record the ambient temperature as this may have an impact if you compare results from measurements taken on a different day.

2. I would then tighten the bolt using whichever method you plan to use when building the engine and lubricate in the same manner being careful not to overtighten.

3. Undo the bolt and remove from the con rod.

4. Re-measure the length of the bolt.

If is has not changed than the bolt has either (a) never been overtightened and you are good to re-use or (b) it may have been slightly overtightened but is still operating elastically and is still good to re-use.

If the bolt increases in length then its Ultimate Tensile Strength has been exceeded and is should be junked.

I apologise for being long winded but believed some explanation of basic behaviour of engineering materials may help with what is a potentially difficult decision.

Wow Chris I cant thank you enough for the detailed explanation. I'll follow your process and record/report my findings. I do have the ARP assembly lubricant.

Chris, this is the best single explanation of the stress/strain curve I have ever read. Well done Sir.

I believe you can download the specs on the bolts. I might have an old hard copy I can take a pic of. Anyway, measure the bolts to see if they stretched from the original tolerance. Get yourself a stretch gauge and the ARP grease with the money you save from not buying new bolts.

A rod bearing failure brings up the question of what caused it, especially on a new built engine. The below thread details rod bearing issues I found when assembling my 3.6.

http://forums.pelicanparts.com/911-engine-rebuilding-forum/856995-glyco-vs-clevite-rod-bearings-data.html

My thoughts exactly..

onevoice thanks for that link. I have a set of clevite 77's ready to install after the failed set was Glyco. I didn't install them nor do I have measurements from them. Hard to tell if they were the cause with no evidence. With new current build I have had the crank sized to the exact oil clearances based on measurements with the rods and bearings installed so my bottom end should be solid from the start.

I would doubt that new bolts are made to a sufficiently accurate length tolerance for this to be a good approach.

ARP recommend replacement if the bolts have extended permanently by more then 0.001" but that the initial length needs to be recorded.

http://i197.photobucket.com/albums/a...pstkum02nt.jpg

Whilst I believe that this is very conservative it would mean that you would typically have to make the length of the bolt to a precision of 0.0001" for this approach to work.

I am sure we could agree that this is unlikely even on a bolt to bolt basis.

I have tried to suggest a procedure that eliminates the need for having access to the original measurement to overcome this problem whilst ensuring that the bolts are still within their elastic operating range.

Chris, you're right. The bolts vary in length a fair amount, so you wouldn't be able to tell if they were already deformed just by measuring their current length.

When I installed mine, I measured the unloaded bolt with some good calipers, torqued them with the assistance of the ARP stretch gauge, then verified that stretch measurement with my calipers.
Mine needed about 45 ft-lbs to reach the suggested stretch, but I didn't do the load/unload burnishing process that some people suggest (sounds like a PITA).
Also, I had to buy a torque adapter so i could fit both the torque wrench and the stretch gauge on the bolt at the same time. I'm not sure how other people manage this without a helper... I needed three or four hands in order to hold everything.

Graham,

If you are installing the bolts to stretch then you will have the correct preload.

There is no need to 'burnish'. This procedure is only used to make the coefficient of friction consistent when using new bolts.

This process does improve the torque/preload relationship which isn't relevant when you measure stretch. :)

Right. I just thought I'd mention it for the benefit of others, since 40 ft lbs (suggested torque without a stretch gauge) vs 45 ft lbs (actual torque required to achieve proper stretch) is a quite a difference, in my opinion.

Interesting that for the M9 rod bolts for the water cooled motors they spec stretch at .095-.100 or torque to 45/lbs.

If memory serves me correctly, a crap shoot mind you, years ago they specc'd higher stretch and lower torque on the 3.2/3.3/3.6 bolts. We found it always took much higher torque to reach the desired stretch.

Cheers

The tightening method suggested by Porsche in the 964 manual is 11ftlbs followed by 2 x 90 degree turns.

This will probably produce a more consistent result that simply tightening to a torque value.

If you took a batch of around 50 good quality Grade 12.9 bolts such as Unbrako or Holo-Krome and carried out a simple test of stretch vs torque I would expect that a scatter of 30% (34ftlbs to 46ftlbs) would be quite normal assuming that you exercised care in both lubricating the bolts and in applying the torque.

Under 'normal' commercial workshop conditions this would likely be much worse.

The Hand Book of Bolts and Bolted Joints (ISBN: 9780824799779) suggest that variations in preload/torque relationships can be as much as +/- 30% but I think this is extreme and is probably based on using low cost fasteners of lower grades rather than the quality of Rod Bolts supplied by both Porsche and ARP.

The torque + angle method should reduce scatter to around 20% and stretch should generally be better than 10%.

The torque measured when a stretch gauge is being used doesn't have any real value, it only serves to indicate that the correlation between torque and preload is generally quite poor.

The figures that I used for the stretch recommended by ARP came from their instructions for Part No. 204-6005.

The fact that the water cooled bolts use different values suggest that the basic bolt must be stiffer for the given thread size.

Get a bolt stretch gauge, it's the only way IMO.

It's way too critical not to.

Chris,

Once again great data.

Thanks!

So I ended up ordering an ARP billet bolt stretch guage. I figured I should learn how to use it. Thanks for the help.

Graham - I've always torqued, measured with stretch gauge, torqued some more, measured some more, etc. until I got the stretch within the range specified. Is there some other way to do this? Measuring simultaneously with twisting would save time, to be sure.

I used a torque adapter like this:
NAPA AUTO PARTS

You attach it to your torque wrench at a 90 degree angle so it doesn't affect the amount of torque measured/applied. I'll admit that I needed a helper in order to hold the crank, rod, torque wrench, and stretch gauge all at the same time.

I actually found that the old school beam type wrench made the most sense for this, so I didn't have to continually change the setting on my click type torque wrench. You probably can't do this with one of the ratcheting torque wrenches, since it needs to stay perpendicular.

So you applied a factor like this: •{Desired Torque to Fastener * Length of Torque Wrench} / {Length of Torque Wrench * Adapter Center to Center Distance (2")}

And how does this square with the 90 degree/perpendicular you mention?

•This Calculation Assumes That the Attachment is Aligned with Torque Wrench Handle When Torqueing Fastener

But I see the light - you have to account for the extra length of your lever arm. Perhaps that is why the old beam style was simpler? No adjusting of click stops when you had done the math?

You just don't need a torque wrench when you are using a stretch gauge just use a simple bar with a socket.:)

Making both measurements is trying to confirm a precise measurement with an imprecise measurement which doesn't seem helpful.

The 'ruling' measurement must be stretch as this is a far more accurate and repeatable value than torque - especially when torque is determined by a wrench with a 'click' adjuster which is a really nasty device.

A dial indicating wrench is far better but the industry has done a great marketing job persuading us that wrenches that click are accurate which is just not true.

If you measure stretch and torque and there are discrepancies I would guarantee that the error will be the torque measurement not the stretch.

Young's Modulus is a very basic material property and it is very, very stable.

thanks for all the tips everyone. I bought another set of ARP rod bolts as I just didn't want the uncertainty of getting this critical clamp wrong this time around. I will practice installing and measuring stretch with my 'old' set of bolts so I'm ready when its 'go time' with the new ones. I'll also measure the new bolts so I'll have a starting length value for each if I ever need it again. Much appreciated everyone. Heading to Disneyworld with the fam this weekend so will hopefully get started after the following weekend.