How flat should exhaust header flange be?

[tom h]

Is there any general agreement on how flat the exhaust header flanges should be?

My passenger side flange is very flat, well under 0.005".

The stock driver side header was a little warped to begin, about 0.045-0.050", but was sealing fine.

After welding a 'precautionary' reinforcing gusset between the #3 & #5 cyls, the D/S flange seems to have warped a bit more, even though it was clamped down (but not to an actual cyl head).

Now, the worst case deviation from flat is about 0.060", at the #5 cyl. The way I measure, is put a straightedge on the flange for #1 & #3 cyls ; those 2 flanges are flat against the straightedge. The #5 flange has air between it & the straight edge, ranging between .050" and .060".

I had been planning to JetHot coat the headers, but now am reluctant to do so if there is a chance the D/S header may not seal (do not want to use a gasket).

Do any of you know how much deviation from flatness can be tolerated?

And if .050 - .060 is too much, is "fly cutting" the flange to achieve flatness, the best way to deal with it?

If I do have to fly-cut the flange, hopefully the fit of the x-over pipe won't be affected

 

[tom h]

I think I have a way to estimate how "flat" is "flat enough" on the header flange ...

I will temporarily re-install the D/S header, torque the #1 & #3 cyls to spec (35 ft-lbs), then torque the #5 cyl until I cannot insert/remove a .001" feeler guage between the flange & cyl head (but not to exceed 35 ft-lbs, in any case).

If the flange flattens out at a relatively low torque (~20 or 25 ft-lbs), then I'll have more confidence that spec torque (35 ft-lbs) should seal leak tight.

Otherwise, I guess it will be time to find a good machinist for the header flange

 

[tom h]

good news.

I just finished the little "experiment" I described above.

By the time I had torqued the #5 D/S header flange to 15-20 ft-lbs, a .0015" feeler guage was stuck tight between the flange & head.

I am very confident that at spec 35 ft-lbs the entire header flange will be pulled completely flat.

So this a useful data point: at just 20 ft-lbs bolt torque, even a .060" flange warp can be pulled tight to the head.

I suppose .060" is a modest warpage. On another thread I had read that as much as 1/4" warp can occur if improper welding is done!

(note: The header bolts were lightly lubricated with engine oil for this test. This can affect the bolt torque readings.)

 

[Ken Cunningham]

Tom H:
I resurrected this thread because I am looking at the same problem, but more severe than yours. I now have 2 d/s headers in my possession: 1) the stock one, welded 1x and now leaking again, and 2) a recent Ebay purchase of an already-welded header. This header has the biggest weld bead all the way around the pipe and all that heat must have warped the flange, because that back cylinder flange is about 1/8" below line formed by the first two cylinder flanges.

I considered whether the bolt torque would seal up, but even if it did I think that the stress on that pipe would cause a very quick failure on the weld, or next to it.

One suggestion from a co-worker: put a thick gasket under just that last flange. Any merit to this?

 

[tom h]

Originally posted by Ken Cunningham
... One suggestion from a co-worker: put a thick gasket under just that last flange. Any merit to this?

I personally don't like that idea , it intentionally creates yet another bend or warp in the flange.

And if you visualize it, a gasket under just that one flange could well cause more stress on that region ... only a "tapered" gasket, if there were such a thing, would work.

With your 1/8" (.125") flange warp, I think that could still flatten out when torqued to the correct 35-40 ft-lb torque ... why don't you test it? If it flattens out your're probably OK. It might still crack later but that's a risk to be taken.

What many have reported doing , for extra insurance against leaks, is applying a narrow bead of ultra-copper silicone around the exhaust port (probably applying to the header is easier), let it skim over (20 minutes air exposure), then install and torque evenly & uniformly to 35-40 ft-lb.

Welding a bracing gusset between #3 & #5 tubes has also been reported to prevent against future cracking. I had a pre-emptive gusset welded in, even though my header is not cracked.

The forum moderator, TurboDave, has a photo of a welded gusset on his web site.

 

[Ken Cunningham]

Thanks for your input, Tom. After reviewing a couple of good discussions of this very topic on this board, I have read 3-4 different ways to handle this. For my case I think I will do exactly as you suggest. I do believe that it will stress the header, but I was already prepared for this to recrack. That's why I have 2 headers now: to be prepared every year or so to do the switch out. After doing the job a couple of times, it's really not so tough. Not much more than changing the plugs...