Cylinder Head Flatness - How Flat is Flat?

[mgmshar]

In the middle of doing head gaskets on my GN. After cleaning the block and head sealing surfaces, I used a straight edge and feeler gauges to measure flatness.

The question is - how flat do they need to be using this measurement method?

In other words, how thick of a feeler gage would you feel comfortable allowing to fit under the straight edge when it's laid across the head surface (or the block surface)?

I have this recollection that .003" maximum is the spec, but I can't find confirmation anywhere. For what it's worth, I'm using Fel-Pro 9441 PT's with ARP studs.

Thanks,

 

[SHADOWII]

Yep!

.003" sounds right to my recollection. Personally I prefer.001". With the 9441, It may not be a problem unless due to cavities,etc. A slight twist will pull down.

 

[Nick Micale]

Since we are talking about a PERFORMANCE engine here, a fresh head with .003" distortion would be unacceptable.

We have surfaced heads and removed less than that total.

 

[bison]

I had 2 sets of aluminum heads cut by my step father in the last 8 months. They were fixtured and fly cut. The most variance between any 2 points was .0003. A straight edge is not an accurate way to assess flatness IMO. A fixtured head in a milling machine will give you much more precise numbers. You can easily compare many different points on the surface. More than .001 would be unacceptable for anything with any cylinder pressure.

 

[mgmshar]

After consideration, I think I'm going to get these re-surfaced, for two reasons...

1. They're not completely flat.

2. Because of this screw-up...
http://www.turbobuick.com/forums/engine-tech/348936-emergency-request-am-i-screwed-head-problem.html

Thanks for the replies - much appreciated.

 

[Tom Walker]

straight edge.

good call on the straight edge. Milling mach. is better if it is checked regularly. Dove tails and gibs wear mostly in the middle and need to be wear checked often. Preferred is a scraped in surface plate or a qualified granite plate. Years ago I lapped mine in on lapping table and never could get a .0005 feeler under anywhere. Ah, for the good old days....

 

[turbodave231]

The "Rule of Thumb" when a flatness spec is not available is .001" per cylinder for the block and deck surface. But in this instance, when head sealing is critical, I would not want to see more than .001" in any one direction.

Bison is right. I've measured plenty of heads that look flat with a precision straight edge. Fixture them up on the mill and run a dial indicator over the surface and you can see that they are not flat.

A good machine shop also "Trams" the mill on a regular basis to be sure the machine spindle is exactly perpendicular to the mill table. Even a small error in the spindle will dish the surface from side to side as it cuts. Any "crash" of the spindle can throw the best machine out of tram.

Dave