Maximum variation in head heights prior to installing cam tower?

[85RedCarrera]

I am building up my 3.2 to 3.4 and cannot find a reference regarding the maximum permissible difference in head heights (cam tower mounting surfaces- top of the heads) prior to installing the cam towers..... does anyone know?

I understand that the heads are to be grouped according to their heights, but what would be the max difference allowable between heads in the same group??

Any help appreciated!

Dave

[JoeMag]

How would you measure? If you’re trying to measure with the head on top of the cylinder, I’m not sure you can get within the accuracy of what’s required. Did you measure each cylinder and head individually? …do not know spec.

[brighton911]

Dave, I can say that a 8 thou difference in assembled height on one cylinder meant the cam bore was distorted enough to cause some binding. So yes, it matters a lot. I determined the high one by bolting just the cylinders down and using a straight edge across the tops. The cylinders were part of a 3.4 package that I bought so not original to my car. Once I adjusted the height of the tall one, all was well and the cam turned effortlessly.

[Henry Schmidt]

Your best chance at success is to minimize cascading tolerances.
Deck the case spigots. Try to hold .0005". Equalize the cylinder height and machine the heads all within .0005" to each other. Surface the cam tower to guarantee a level sealing surface.
Make notes about each process and think about using those notes to recreate a standard stack dimension. This may effect your deck height slightly but generally not a significant amount.

[4sd911]

If the heads come back from the machine shop and they are not within 0.0005” I assume they need to go back and recut? Or can the height be adjusted?

[85RedCarrera]

Brighton - Wow, 8 thou is a lot- I am not surprised that the cam was binding!

Henry- thank you for the professional approach to this.

I have one group of heads that vary .0001" from tallest to shortest.

The other group of heads has one head that is .0012" taller than the shortest head. The other two heads in this group are within .0003"

So, I have one outlier........

I mic'd the heads at 4 corners and took the average. I am going to measure the second group again to check my measurements- after my crazy cc'ing debacle, I realize that I need to be very careful with my technique, when dealing with such small tolerances/measurements. That said, I may need to have one thousandth taken off of the top side on that tallest head.

I finally have all of my chambers within 0.5cc, so I don't want to shave the chamber side of the head.

All of this is me just measuring the "thickness" of the heads. I still haven't placed the cylinders on the spigots and measured the stack all the way up to the cam tower. We'll see where I'm at when I start stacking things and pull out the machinist straight-edge!

I really appreciate the informative responses & advice- thank you! (I couldn't find anything in Wayne's book or the Bentley manual)

Considering all of the work that goes into doing this the right way, I am curious as to what the budget Porsche engine rebuild house(s) do- just skip it??

[4sd911]

Originally Posted by Henry Schmidt

Your best chance at success is to minimize cascading tolerances.
Deck the case spigots. Try to hold .0005". Equalize the cylinder height and machine the heads all within .0005" to each other. Surface the cam tower to guarantee a level sealing surface.
Make notes about each process and think about using those notes to recreate a standard stack dimension. This may effect your deck height slightly but generally not a significant amount.

Not hijack this tread but I may have a problem similar to this.

If we are keeping the head, the cylinder height and the case all within 0.0005", if all measurements cascade in the same direction can the torqued heads have a difference of 0.0015" and be OK?

My car after about 20,000 miles on the rebuild has started show a little oil leak at the head to cylinder. I plan on pulling the engine and removing the right bank to see whats going on.

[Henry Schmidt]

Originally Posted by 4sd911

Not hijack this tread but I may have a problem similar to this.

If we are keeping the head, the cylinder height and the case all within 0.0005", if all measurements cascade in the same direction can the torqued heads have a difference of 0.0015" and be OK?

My car after about 20,000 miles on the rebuild has started show a little oil leak at the head to cylinder. I plan on pulling the engine and removing the right bank to see whats going on.

If you were lucky enough to hold a 1.5 thousands on an entire stack you win. Those kinds of results only exist by shear luck. Of course 1.5 thousands between stack is acceptable
You target unrealistic numbers to reduce the cascade. I think you'll find deviations in the 2-2.5 thousands acceptable. Remember, when you torque down your head studs everything changes.

A word of caution: As cool as precision feels good, you can waste a tremendous amount of time, money and serenity searching for perfection.
Just for fun a few months ago I went searching for .1 gram balance on a set of steel 911 rods. Porsche calls out 4.0 gram but our normal goal is .5 gram but I was feeling froggy.
2 or 3 hours later I tapped out.

[mikedsilva]

This is one of those things I obsess over. I do a dry assembly with no torque and no base gaskets... if the straight edge shows no gaps, then I do a light torque and check again across the heads. No gaps and then I install the cam tower and install a cam with oil and turn it. If it turns freely, I'm good to go.

[stownsen914]

One thing I noticed during my current motor assembly process is that without torque applied, I saw some variation of a few thou in cylinder heights before I put the heads on. Initially this concerned me. Since everything was checked and/or machined at a respected machine shop, I went ahead with the assembly and checked carefully for camshaft binding as I torqued the heads and cam towers. In my situation, not even a hint of binding. Happy days ... I assume things may just not seat fully without being torqued down.