Cylinder Height Check 3.3L cyl

[reclino]

Car is 1986 930 3.3 L 97mm cylinders
Finally getting to use the Surface Plate I bought used a couple years ago. Checking each cylinder with an indicator that is accurate to .0005". Need to match up 2 sets of 3 cylinders that are the same height. The first one I check varied .003" from side to side. I rotated the cylinder 90 deg under the indicator 4 times, taking a reading at each spot. See the pictures below.

The setup



First measure. (Top of cylinder when installed, smooth side without fins)




Second Measure rotated 90 deg



Third measure rotated 90 deg (bottom of cylinder, notice cooling fins)



Fourth measure rotated 90 deg



So using the second and fourth measure "side to side" this cylinder measures almost the same height.
However "top to bottom" the first and third measure are different by .0035". What the heck is going on here.
At this point I took a break and went to get a beer.

[reclino]

While sliding another cylinder into position I noticed I could rock it back and forth. Went back to first cylinder and it was the same. I could turn it 90 deg and the rocking would follow the part so it's not a low spot on the surface plate. .... Decided to measure the rocking.

[ahh911]

81sc 3.0 alusil.

Same test setup as you.

Off the car, within 0.2 thou/inch around for cyl 4,5,6. Didn't measure 1,2,3 right off the car.

Here's the thing that went wrong at first. I wiped the cylinder head against 1000 grit paper after finding that the sealing surface wasn't mating perfectly to the heads using blueing (was really the outer edge past the sealing ring, but on some heads this is sloped so this bluing test without clamping the heads would not work). What I didn't know was the the intake side needs a little more pressure than the exhaust side. I ended up with around 0.8 thou/inch on a particular 90 degree worst case delta between all measurement of the three cylinders. and around 0.4thou/inch on the 90s on any single cylinder (you measure ~3). Only then did I figure out how to measure it, I corrected the worst ones a little to the numbers given, had in the end removed 1.5thou/inch from original height. For the next 3 cylinders, knowing how to measure, I cleaned up the surface and achieved good height uniformity and removed hardly any surface material. After figuiring out the swiping technique and measuring after a few swipes, I was able to control the uniformity. The 3.0l has the sealing ring as mentioned.

Maybe someone tried to clean up the cylinder head by swiping across sandpaper and overdid it in one direction. Also, for the turbo, if the heads bang against the cylinders because of too much boost, that might wear the exhaust side having the extra heat, just guessing.

Come to think of it, mine never rocked, they were simply "cleaned up" off angle, so very different from what you are seeing.

Phil

[reclino]

I am trying to understand what you mean by 0.2 thou/inch
Do you mean .0002". this is usually called 2 tenths in my government machine shops. ???
There is more information in my "barn find" 1986 911 turbo thread on the turbo forum. These cylinders went to Craig Garrett with my heads for checking. He cut 0.25mm off the top of each cylinder, and 0.11mm off the head sealing surface of each cylinder head. The cut to the cylinder sealing surface was at the proper angle. These cylinders then took a ride to another machinist in NJ, further measurement of the bore determined that they were not round enough, so decided at that point to have the nickel plating redone by Power Seal USA. It looks to me like the plating on the ID may be higher in a few spots. I am using the trick head gaskets on this build PN 93010438101, that are 0.50mm thick. So I could probably afford to cut a bit more off the top of the cylinders if needed.
Fred Apgar who I have consulted a number of times on this build suggested lightly lapping a cylinder against one of the heads just to visualize what is really happening. As soon as I get some lapping compound I will report back. I am motivated to put this together correctly, so a little more waiting will be just fine.
Long story short, as I had cut both the cylinder and head side of this sealing surface I assumed everything would be just fine. It's always better to find the problems early, before they get loud and expensive.
David

[ahh911]

Yes, 0.2thou/inch would be .0002", since the tool is marked at .0005" increments, this would mean that it's most likely within error limits, but if I repeated the test several times, this is roughly the delta. So basically the cylinder height was very nearly perfect for cyl 4,5,6 as pulled off the car.

I did this sep 2020, so I'm looking at my notes, and am pretty sure that measuements are correct, but they do lie between the 0005" marks and needed judgement. I'll triple check that the notes make sense.
Phil

[reclino]

Yeah I was hoping for less than .001 difference between.cylinders. I wish my indicator was better like a starrett last word, but you can only prepare for so long
Wiith parts so far out of tolerance it's easy to measure.

[ahh911]

That is correct, the clasification for my tolerance group 5 was 1 thous/inch height. I tried to keep it there for cyl 1,2,3 that I had messed up a bit. Actually looking at my data, I see that cyl 3 had 0.0006um of delta for height on the 90 degrees, not 0.0004 mentioned earlier. So that put the delta between any points of the 3 cylinders to 1.2 thous/inch. (.0012), not the .0008 mentioned earlier. It was a judgment call, once I knew the clean up was off angle, I tried to bring it back and balance total height removed vs the height delta's between cylinders, of any one cylinder, and total height. This was the compromise. Like I said, the other side was nailed down to near perfection.

You think it's just localized irregularities in the surface? I would start with some machinists bluing compound, applying a thin amount on the cylinder top side flip it against the flat surface and rotate 5 degrees back and forth with even pressure. The low spots will show up (3 thou), guaranteed. As a matter of fact, I have a 1 thous/inch feeler gauge, that would show it very easily as well. Anyhow, the blue will get you started in the solution. I also applied it to the cylinder and tested it on mating to the heads using the same technique, then verified it was indeed the cylinder and not the heads and took it from there, but in my case I was trying for perfection and messed it up at first, with 3 thou off, you'd better figure it out.

Should mention with machinists blue, you are looking for the mirror imprint on the flat table that was left behind, that was the best way for me to look at the contact areas.

Phil

[reclino]

Well now I can see what I was feeling with the indicator. This looks like they plated right over the gasket surface, and then tried to bead blast it off. Leaving flecks of plating behind for me to deal with.
To lap off the marker I just rubbed some diamond buffing compound on a sheet of paper, lapped the cylinder using the surface plate to keep the paper flat.
Trust, but verify.
So I lapped this a bit more last night the bead blasted surface is making it almost impossible to just rub the sharpie off like it does on a smooth machined surface. The areas circled in red are definitely high sport, but it does not look so clear cut with more lapping and cleaning.
More to follow once I get some proper lapping compound.

[reclino]

This is the surface of one of the cylinders when I got it back from Craig Garrett, I was very careful not to disturb this surface, as I paid for it.

[reclino]

I talked to the machinist today about this, I have a plan to move forward. I am going to proceed with the bottom end assembly, get to the deck height measuring step. I can determine what the shortest good cylinder would be, base shims will allow for plenty of adjustments, then just cut the cylinders to set deck height as close to perfect as we can.
Easy