Weissach911, in your opinion, what would be the material of choice for this application and why?

At the top of my consideration is that the stud should have a similar coefficient of thermal expansion to the cylinders. Do you believe this is important? I bought a set of 993tt studs from Steve W because for this reason.

On the other hand, how much stud pre-load does it take to appreciably stress and deform a cylinder wall to the point where it affects cylinder or piston ring life? This is the SuperTech and ARP side of the argument where they believe that thermal coefficient is not a top consideration, at least from what I understand.

Just a couple of simple questions, ehh.

Thanks.

Tricky question and I am not sure that any answer I can give will not get me shot but from an engineering perspective there must be some factors to consider in arriving at a judgement.

I think that the basic design of the cylinder head joint has been fairly close to the mark for some time now and the issues of leaking heads are a problem and maybe a fact of life.

As has been suggested by Fleiger, I would agree that the minimum preload needed to keep the stud free from fatigue loading is desirable as this will have less impact on the basic casting and reduce the likelhood of stress relaxation taking place in this area.

Clearly a Dilavar stud will be the best option for this consideration as it will help to maintain the preload at a more consistent level throughout the range of operating temperatures.

The precipitation hardening Martensitic Stainless Steel studs that are offered by Supertec and Casper Labs (Carpenter Custom 630 seems to be a slightly modified 17-4PH) will cause an increase in preload with increasing temperature and if the head is lifting off due to peak cylinder pressure will be the preferred choice.

I am not sure about Raceware materials but possibly they are similar to the the 17-4PH steel. These studs also have a lifetime warranty.

In general terms a standard steel stud will provide broadly the same increase in preload with temperature and at $14 for one that is black coated seems to be a very cost effective. If they were used with good quality nuts and ground washers I would find it difficult to understand why there is a significant difference in performance as the expansion rate is similar to 17-4PH and the modulus is almost identical.

I can believe that they could suffer more from corrosion but I think most of the 911s we are discussing now lead fairly sheltered lives and are much better manitained than in the Eighties and Ninties.

My decision would be to chose between a standard stud and a Dilavar stud and this comes down to what you believe causes heads to leak and I have to say that on balance I tend to favour case problems rather than stud problems and if pressed would think Dilavar gives this part of the engine an easier time.

I would worry about damage to the resin coating as this could lead to problems in salt laden environments - Buffalo etc in Winter but apart from this I would have few other concerns.

In the limit I would tend to side with Steve W and go for the Dilavar.

I am sure many people will disagree and it would be great to have enough time and money to make real measuremets of studs and heads and rely on data instead of opinion.

That is why the Supertec stud comes with specific instructions to clean the threaded holes completely and then install the studs to bottom then back them out a 1/4 turn.

I was unaware of this when I made my last post; thanks for clarifying and apologies for the mis-statement!

Thanks, Jim - I thought I remembered Henry's installation instruction saying this , but you saved me digging through my papers.

To me, the installation height issue (to the extent it is an issue) is solved entirely with the external wrenching nuts. You won't be bottoming out your hex wrench and having it pushed out of its hole in the stock barrel nut. Particularly useful if you have been manipulating stock dimensions for compression adjustment or head/valve clearance. Flange nuts for stock threads aren't that hard to come by, and, in my personal sample of one motor, work fine. For that matter, I think you could find reduced wrench (15mm) regular nuts for this application. With the thick hardened steel washers, wouldn't the loading distribution on the aluminum of the head be about the same? And the nuts would be strong enough themselves?

Still, you can get in trouble if you run stock length studs in too far - you can end up with not enough thread engagement (or, at least, less than you would want - a full diameter's worth). Easy enough, though, to measure installed height and hit the mark. Easy, at least, for someone who has 135mm seared into his brain as a result of not being careful about this the first bare case build he did.

Having basically slightly longer studs should be a help if you are building a longer than normal rod motor, with concomitant increase in head to case distance.

I understand (from hearing it from engineers like Weissach) that the first couple of threads take most of the strain, but in dealing with used (sometimes very well used) cases, might not that also mean those first few threads in a case have lost some of their strength?

I credit Steve and Henry both with a high level of practical experience, which has led to hard won knowledge. It is great that they and some of the others in the business will share some of that with the likes of most of the rest of us, who have usually only worked on our own motors.

For me, a stud decision came down to economics. I gave up on mag cases for race motors quite a while ago. My 2.7 alu cased race motor has a somewhat motley collection of not so special studs holding things together well enough, but it is only in the 250 flywheel hp range. I am hoping for 300 hp from a 2.8, so I took Steve's advice that the stock steels (of which I have several engine's worth in fine shape) were not enough. I have had a good experience (which means no bad experiences) with Raceware rod bolts, so those studs ought to do the job.

Henry's Supertecs have a distinct price advantage, and having the hardware included was also a plus for me. If I could afford to have Steve build a motor for me, I could afford the studs he favors in the bargain.

Can you clarify your question? Are you under the impression that steel head studs place too much stress on the cylinders during normal operating conditions?

Head stud torque is engineered to have the appropriate preload @ operating temps. That is why you only torque to 26ft/lbs. So in reality, thermal coefficient of expansion has been considered.

The only time Dilivar would be theoretically nice is during transient temps(warm up/cool down) and you feel the need to drive the car. If you let the engine idle up to temp you have placed no load on any of the components you are worried about. Porsche felt the need to have Dilivars because owners do not always treat their cars the way they should...especially under warranty.

Since an engine spends most of it's @ operating temps the argument for the need for Dilivar is moot. It is even more so when you figure the price for them.

Excellent question...I did assume that the preload would be calculated by the desired clamping load when cold. This would mean that when warm, the aluminum cylinder would grow more than the 17-4PH (in the case of SuperTec) stud and would cause stress and possibly some amount of deformation/ distortion of the cylinder bore.

However, if the pre-load is based on the engine being warm, then wouldn't the pre-load be too low when cold and cause cylinder leaks & walking?

The expansion rates are close enough to regular steel studs that it does not cause issues. We see no bore wear issues on both race and street motors we dissassemble that use the Supertec studs.


Yes, All studs, including Dilivar will be too low. That is why we should warm them up before we drive.

Is the start up phase where the blow by is likely occurring (as pictured by Henry a few pages back?)

D.

Warm up to some extent. But the more common area of possible concern would be a cooling off period, then get in the car and drive the car aggressively. The cylinders have contracted but the studs retain more heat for longer periods of time. This transitional state is where I believe Porsche was focused on.

Keep in mind the transient temperatures are a relatively short in duration so the issues are minimal.

As far as the turbo in the pic I could only speculate that it is something other than the transitional temps as it looks more advanced than what I would expect to see.

First let me say this is an awesome thread. Thanks to all contributing!

Ok, setting the stud preload at temperature makes logical sense, but I don't follow why, for a purposly under torqued fastener, an elliptical all metal lock nut isn't used?

Especially given that over torqued lasts more fatigue cycles than under.... Seems like cheap insurance aginst things coming undone during the transient warm up/cool down phases.

t

Tadd

To me, the fact that head stud nuts on properly built motors don't come loose is reason enough not to complicate torque settings with lock nuts. I'd never want to use something like that on those studs, which occasionally will back out anyway.

The residual frictions created by the stretch 26 lbs/ft creates are empirically (or so I posit) adequate to prevent the nuts backing off. At operating temperatures the strain apparently is increased, increasing the friction. But since the elastic limit has not been reached, there is no relaxation phase where there is less friction and an opportunity for the nut to rotate.

This would apply as well, I'd think, to the case through bolts. They don't back off either.

To me, any kind of lock really has a place only where, for some reason, a real torque (some percentage of the fastener's capacity) should not be applied. Like where you have crush washers on valve covers - those 8mm studs can take a whole lot more than is needed to squish the washers or warp the covers, so the nylocks help prevent vibration loosening nuts which are only minimally torqued. Or perhaps where there is rotary motion, as with axle nuts.

Walt

In general I would agree with the comments about not inreasing prevailing torque due the friction and to try to control preload as accurately as possible.

I always worry about Kaynar self-locking nuts and if you buy them with an 'Aircraft Release' certificate the cost can be bewlidering - Symmonds Precision in Europe quote $14 each for an M10 in quantities of 200 pieces.

When you buy then for a Dollar from a race shop and without certificates then I wouldn't use then in torque/tension critical areas as the variation in prevailing torque could be an issue.

There is an old Lockheed report which details around 40% tension variation due to the oval head form.

I am not 100% convinced that this system has been designed to give a 'hot' preload as this wouldn't be best practice for a road engine although I could understand this approach being taken with high end race cars.

It must be worth looking at some basic data.

The preload in an M10 x 1.5 fastener torqued to 26 lbsft would be around 5250 lbs assuming that threads were lubricated.

This would result in a clamping force of 21000 lbs when cold and would mean that the joint could stand a peak cylinder pressure of about 1900 psi before it lifted.

This would be a very high pressure for a normally aspirated standard engine.

If we now look at the clamp force for a 'hot' engine and assuming a mean temperature increase of 125 degC the increase in clamping force can be approximated.

In this approximation I have assumed that the room temperature modulus of all the stud materials has reduced at 1% per 10 degC temperature rise, which is reasonably typical.

I have not made any allowance for the elastic compression of the head/cylinder which would reduce these values but difficult to calculate without spendig too much time. I have also assumed that the aluminium parts and the studs are at the same temperature.

The strain produced in each of the fasteners caused by the expansion of the cylinder and head would be approximately:

17-4PH 1.5 x 10^-3
Steel 1.0 x 10^-3
Dilavar 3.74 10^-4

This results in the following preload increase per stud when hot:

17-4PH 3375lbs Total 8625lbs
Steel 2250lbs Total 7500lbs
Dilavar 17lbs Total 5267lbs

Clamp Force increase

17-4PH 13500lbs Total 34500lbs
Steel 9000lbs Total 30000lbs
Dilavar 68lbs Total 21068lbs

If we then try to calculate the force required to pull out the stud from the case assuming a shear strength of about 40ksi for the AluSil material we end up with a value of about 26000psi.

The yield strength of the thread would be significantly lower than and some deformation could start to occur at about 18000lbs.

If cylinder temperature is higher than 125 degC then the above forces will increase in a reasonably linear fashion. The Alusil may weaken so stress relaxation may be more likely to occur.

The other question is how to deal with initial stress relaxation. I would imagine all these studs could lose about 10% of their preload within a coupls of hours or a few thermal cycles and it would be interesting to see if anyone has experience of this phenomenon.

The next logical place top look for issues is the behaviour of the CE ring and I think I may try a Wills Ring as an alternative.

Great post! But your numbers seem very high.

I measured 3989 lbs which would give you 1125 psi cold so it would appear Porsche Engineers calculated for both cold and hot.

engines with light load are 300 psi

N/A production engines at wide open throttle can get up to 900 to 1000psi

racing engines can get up to 1500 psi

I'm not saying my numbers are correct, but they do fit the scenario better as cylinder distortion, pulled head studs and other complications have not materialized as your number would imply.

Also, cylinder temp and the resulting expansion is not consistent. I would expect the lower cylinder area to be close to 90 degC with approx 110 deg as you come closer to the combustion chamber.


Just to give an example, a pushrod V8 can have around 85 ftlbs and up to 100 ftlbs of torque for cylinder head torque(approx 3278 psi). Completely different model I realize but it does show what is required to retain the integrity of the joint.

Aaron,

I have tried to take the 'best' approach to the torque/tension relationship and I agree that these figures may be high as I have assumed that the nut was very well lubricated.

I could easily understand a 25% reduction in these loads depending on how well the thread is formed, whether or not the nut/stud has been burnished prior to final tightening.

At 1500psi peak cylinder pressure I could start to worry about head shuffling and would think about increasing the preload.

I would then have to worry about the stud pulling out or suffering from dynamic problems and suffering from vibration induced stress relaxation.

The easiest to overcome these dynamic issues is to increase preload and this leads us into the argument about Timeserts vs Case Savers.

I also wondered in there could be a resonance issue and has any one tried 'lumpy'head bolts. (I see a post about RSR through bolts which made me think of this)

Could you guys educate those of us in the Holiday Inn Express crowd just a little here.

How do you calculate the preload on a fastener at a given torque value?

I understand that it varies with lubrication, properties of the two metals which are threaded, and burnishing. With tensile strength and cross section area of the fastener in the equation, does it come down to Archimedian inclined plane calculations? So many degrees of rotation = a stretch of X = a preload stress or force of Y?

I can see that the clamping force per head is just 4X individual preload. To calculate what that clamping force will hold do you just divide preload by the cylinder area?

Can peak cylinder pressure be calculated? Nominal static compression seems straightforward, but a running motor seems tricky. With some approximations and assumptions you can get a ballpark figure? As opposed to actual measuring it (which is probably way beyond what any of us shade tree guys are ever going to contemplate)? Or do you backcalculate from torque measured on a dyno? A factor to approximate peak pressure from BMEP?

Gasketing for a valve job on a 993.
 

I am about to pull he heads from my 98 993 motor to have a full valve job done. My problem is trying to decide what gasketing I need to order from our host. There are partial gasket sets avaialbe all the way up to a gasket set for close to $600 for a head job. So, I don't really know what I really need.

I am pulling the heads which I will send off to Steve Weiner for a full rebuild. So, what gasketing do I need to order from out host to reinstall the heads on my motor?

Not sure yet, but I may elect to pull the P&C's to check the rings and to change out the con rod studs. Does all of that dictate different gasketiing?

Any help would be appreciated.

Walt:
Cylinder pressure is, as best case, measured directly. Pressure is way too variable on how much fuel/air charge you load (VE).

These guys have a good write up on how it's done... It's literally adding a high speed readout to a piezo transducer and mapping to crank angle (or several other arrangements)

Untitled Document

tadd

Walt - preload to axial force data is quite well established and is commonly published and available from most good fastener suppliers.

Fastenal for example provide good charts for most Metric Fasteners and they are OK for first order approximations.

There is a good MIL spec that goes into great detail.

MIL-HDBK-60 - THREADED FASTENERS - TIGHTENING TO PROPER TENSION

You will need to download to view but the data is helpful.

As Tadd says peak cylinder pressures are best measured with piezzo transducers but again there is good published data so it is possible to make some estimates.

The problem with estimating peak pressures is that serious 'pinking' or pre-igniton can be much higher than typical values and either pull studs or cause piston damage.

I realise my calulations are estimates, but they do indicate differences between the various options and allow some judgements to be made but it would be better to have 'real' information based on detailed testing, which was, I thought, the point of the initial question and I was only trying to establish some basline.

Hi There
For my top end I ordered a Wrightwood racing set, but there will be gaskets not used. For example you will not need the stem oil seals if you are having the heads rebuilt.

People keep mentioning 26ft lbs of torque for the head studs, but with my steel ones the manual says 15ft lbs and then 90 degrees, which pretty much equals 40ft lbs as far as I can tell. That is what I used last night on my (lubricated with engine lube) head studs.

That felt pretty tight.

Berni