This is an intresting topic, that I have read a lot about, but don't know that much about. I was recently in the mall waitng for the wife to buy somthing or other, and I picked up an issue of hot rodder, or somthing like that and in there they had an article on setting "torques for hi- performance engines" and in this article, they said that the only true and correct way to "torque"is by mesuring the stretch with a dial indicater, and then they went into examples of oem head bolts as well as arp headsutds were they torqued them to the "spec" as per instructoins pased on lube or oil or dry, and none of them came even close to what they should have been via the streach numbers, due to the incosistincy in the friction level between the block and the stud. Moral of the story if it isn't set via strech number it isnt set right. This is how it is done in F1, indy, cart, baja, nascar, nhra, ihra, ect... The other thing that I read was from a builder or custom copper head gaskets, in regards to torquing and HP that you make the most HP with the least possible torque that still seals the head. so ? well I am confused, but I would probabley be inclined to go with the strech version myself :wink:
Generally speaking, the way I understand the "pros" IE: (F1 engineers) set the torques of their fasteners is they determine the torque at which the fastener yields and torque to a certain percentage, IE: 80-90%. The way they determine this is either by looking up on a table and/or calculations based on size, thread pitc, materials, and lubrication, or they can actually buy extra (they've got enough money, right?) and torque a fastener until it yields, then remove it, replace it with a new one of the same type, and torque it to a little bit less.
I think I understand the train of thought of the guy who said the least amount of torque is best for power, but I think I disagree with his recommendation. He is probably saying that because of more clamp load, there is more strain (movement/distortion) of the head/block assembly, and that a block with cylinders that is strained will have the cylinder-shape bores change into oval-shape bores. The piston will rub unevenly and seal less well on an oval bore and therefore, will make less power when the bores aren't perfectly cylindrical. I think I agree with all that (although I'm not sure how big of a factor it really is.) But undertorquing goes against the other good practice rules with ensuring a good head gasket seal, IMO. Based on what I've read, more bolt load is generally better; and the VW head/block seems plenty stiff. The recommended practice in eliminating the "oval bore" syndrome is rather than reducing bolt load, is to machine the bores with the stresses of the head and head fasteners at full torque simulated. Very good machinists will have fixtures called "torque plates" that bolt onto the engine block, sort of like a "simulated head", but with big holes in them for the boring/honing tool to reach through and machine the bores. They will bring the simulated head fasteners up to full torque before doing machining the final bore/honing operations on the cylinder. Then the perfectly machined cylinders will in theory remain cylindrical when the real head is bolted and torqued up on top of there. Using a torque plate is one of those little machining tricks that makes the difference between an econo-rebuild suitable for street use, vs a high quality blueprinted type of engine rebuild. Some machinists have torque plates and use them for rebuilding VW motors, others don't.
A danger with undertorquing is if the pressure on the cylinder head exceeds the bolt load, the bolts can then see a dynamic loading where they stretch and relax a little bit with each cylinder fire. This can lead to failure of the fasteners, even though they may be installed at a load far under their yield or tensile strength! You might think, there is less chance of failure of the fastener if it's torqued less? Well it sounds weird but it is not necessarily true. Over time small amounts of stretching from dynamic oscillations of an undertorqued bolt can lead to metal fatigue of the fastener material. (Not to mention the head gasket will have a tough time sealing when the surfaces pressing against it are moving!) As a result of this, the general practice is to specify fastener torques that comes as close as possible to the yield strengh of the fastener, assuming the material being fastened is adequately rigid and there is adequate thread engagement, etc.
I think you are right about lubricant effecting torque readings drastically. When I torqued head bolts with only oil lubricant as specified by bentley, it can be felt how much friction there is and the lack of "smoothness" when torquing the bolt because of all the friction. But when I lubricated the threads with moly lube, boy did they torque up smoothly. It would be very difficult to measure the lengths of installed head bolts/studs as they are threaded into blind holes! A bolt that had both ends exposed, that wasn't threaded into a blind hole, might be measurable however. What you could do is measure the length before/after installation of the bolt though, to see if it has permanently stretched at all. If it has permenently stretched, it was torqued to above it's yield strength.