Not a stupid question and worth asking.

Rod bolts always have to be stretched, the question is how much :)

ARP suggest a Porsche Rod Bolt needs to be stretched by around 0.0095"

ARP also suggest that if a rod bolt has plastically deformed more then 0.001" it should be discarded.

Porsche now suggest 20Nm + 90 degrees which is a stretch of 0.0125" but the 'use once' approach seems to suggest that bolts plastically deform.


It has been suggested that dial calipers can be used to 0.0005" but in inexperienced hands only to 0.003" which is clearly poor.

There is a common misunderstanding that repeatability and accuracy are the same thing but this isn't true.

You can consistently miss a target and still hit the same spot over and over again.

The result may be precise but it is not accurate.

The uncertainty of any measuring system is the really crucial point and if the uncertainty is greater than the difference being measured there is a problem and you just don't know where you are.

As the NPL (National Physical Laboratory) in the UK will only calibrate the standard digital caliper to an uncertainty of +/-0.02mm this is where I place my trust.

The real issue is that we are trying to develop a preload in a rod bolt and this preload needs to be greater than the applied alternating stress that is developed in the bolt.

Assuming the bolt is of sufficient strength the tighter we can fasten it the better the service factor - again assuming that we don't deform the rod.

The need for the bolt to remain elastic is arguable but as the preload up to yield point is linear with respect to stretch the uncertainty of this measurement results in an uncertainty in preload.

It seems to me that the ARP spec may result in a lower preload than that used by Porsche but I don't have an ARP bolt to measure so I am not 100% sure that this is true.

As I said in an earlier post if we have a calibrated caliper that is correctly used then the uncertainty of preload would be around +/- 10%.

I would believe that in general terms this is optimistic and real results could be much worse.

This is why I suggest that torque is likely to be as effective as using digital calipers.

If we use a stretch gauge such as an ARP then the preload uncertainty is likely to be around +/- 2.5% and the correct use of the devices is much more easily achieved than using a caliper. It may not, however, be much better than torque + angle

In terms of material behaviour most elastic bolts are loaded to 80% of yield but I believe rod bolts are generally used closer to their yield point than a 'commercial' fastener.

The preload difference between yield and ultimate tensile strength is probably around 10% and this is why I believe it is important to monitor torque during tightening even when measuring stretch.

Where does that take us ?

If you don't want to buy a good gauge use torque + angle and if possible uses a digital readout torque wrench to monitor torque.


If you really want to measure preload accurately you could buy one of these :)

https://www.skidmore-wilhelm.com/collections/bolt-testers/products/model-hs

^ I feel like I said "use torque" in many of the the other threads and it was more or less ignored. The short version here is that if you do not have a good stretch gauge, using torque plus angle is the way to go.

Which is what I said.

Does ARP give a method for torque plus angle? Haven't seen it mentioned on their site... I have seen one for Porsche bolts.

Also, ARP says to use their special grease to get the right "torque," yet many on this site and Waynes book specify to use Loctite.

Which is it? I would think you would get different results torquing with grease VS Loctite...

Of note, in the assembly videos from Porsche shops on Youtube, I don't see anyone using Loctite on the rod bolts...

It would be interesting to compare torque + angle and stretch - I am not sure there would be much practical difference.

I have a colleague with a Norbar Ultrasonic Bolt Meter and I will try to steal it away for a weekend.

I think using Loctite is an issue. If the Loctite spreads to the faying surface of the nut it will make a mess of any torque readings as it starts to 'set'.

Stretch won't be affected and torque +angle may be OK as well.

The coefficient of friction of Loctite is similar to a light oil but I wouldn't generally use thread lockers on rod bolts.

The idea of an Ultralube is to try to reduce the scatter in preload due to variation in friction whist torque tightening a bolt.

It isn't much of a benefit if you measure stretch or torque + angle.

If you torque tighten to a figure that has been specified for a 'dry' or lightly oiled bolt than the reduction in friction may well result in overtightening and excessive preload.

To stretch an ARP bolt to 0.0095" you would need an angle of around 70 degrees and the 'snugging' torque would be quite easy to determine but around 20Nm seems sensible.

When you use ARP products you follow their instructions 100%. The instructions are simple....if you don't have a stretch gauge, which ARP recommends, you torque to 50 foot pounds. Using their supplied grease of course. No angle needed or given.

It is *really* important to not over think these things.

If you think Chris in this forum knows a lot about bolts, you should talk to Chris Brown, the engineer at ARP. He will make your head spin.

I don't know much about bolts but I do have a basic knowledge of the tensile deformation of metals which governs the manner in which the bolt deforms and how it responds to being tightened.

http://forums.pelicanparts.com/support/smileys/nuts.gif Understatement of the year!

Interesting discussion - Along the lines of using an ultralube to reduce errors due to preload, I'm sure I've also read that tightening once or twice before the final tightening is also beneficial, effectively burnishing the threads beforehand? Correct?