Best race valve material

[blwilliams]

Lots of good information and opinions here and I thank all that have contributed.
While I believe an aftermarket BBC head (Made in the USA) is probably a bit tougher than an Alfa head all of these basics apply.
I will no longer be using the PBM valves in my higher end racing heads.
I should listen to my own sales pitch of you get what you pay for.
PBM valve is under $15 the Ferea and others are about double that.
I few extra $100 bucks on exhaust valves may have saved $thousands in repair/replacement bills.

[digger]

N20mike is on the money. Usually not concentricity.

It's usually worn valve guides or cams that have design flaws around max lift. Causes valve to whip

Many years of testing by major automotive manufacturers and valve manufacturers has resulted in non-concetric valve face to valve seat as being the blame for most valve stem breakage. Just a couple of thousandths of mis-alignment will result in significant bending as the valve goes through many thousands of cycles per minute in a running engine, even a bone-stock engine. Add in the concept that the valve is rotating, which results in back and forth bending throughout the circumference of the valve stem. It WILL break.

here is a plot of a 7mm valve with the seat insert 0.002" offset from the valve seat

the nominal seat pressure induces a stress of 25MPA, bending increases that to 54Mpa when it has found its natural settling position which is nowhere near enough of an alternating stress to induces any kind of fatigue. a bigger stem diameter will be be worse as simplicity its an enforced displacment problem



the valve whips around more than 0.002" when its running at lift so while length of unsupported is higher i bet the bending stresses are still going to be a fair bit higher under those conditions, i'm not convinced a little bit non concentricity is a show stopper, also when the valve get slammed on the seat there is going to be way more stress

[MadBill]

...
The concentricity of the valve face to stem center line to valve guide to vale seat must be dead on.
I have seen many valve failures and from some of the very best and highest quality valve manufactures when used in heads that had less then optimum concentricity in the valve job...

What's the consensus on how close to zero it has to be to pass the bounce test?

[ProPower engines]

Depends on guide clearance.
If you have lots it can be way off concen. and still bounce.
With straight stems and honed fitted set up at ..0008 it needs to be less then .0005 to still bounce 2 -3 times.
That's not being generous the stem has .0002 runout that seems most common these days regardless.
Its acceptable runout for a stainless valve.

[williamsmotowerx]

here is a plot of a 7mm valve with the seat insert 0.002" offset from the valve seat

the nominal seat pressure induces a stress of 25MPA, bending increases that to 54Mpa when it has found its natural settling position which is nowhere near enough of an alternating stress to induces any kind of fatigue. a bigger stem diameter will be be worse as simplicity its an enforced displacment problem



the valve whips around more than 0.002" when its running at lift so while length of unsupported is higher i bet the bending stresses are still going to be a fair bit higher under those conditions, i'm not convinced a little bit non concentricity is a show stopper, also when the valve get slammed on the seat there is going to be way more stress
[/quote]

Do you have stem in a fixed position in your model? Add .0008" clearance for valve guide and stress will be even less.

[digger]

here is a plot of a 7mm valve with the seat insert 0.002" offset from the valve seat

the nominal seat pressure induces a stress of 25MPA, bending increases that to 54Mpa when it has found its natural settling position which is nowhere near enough of an alternating stress to induces any kind of fatigue. a bigger stem diameter will be be worse as simplicity its an enforced displacment problem



the valve whips around more than 0.002" when its running at lift so while length of unsupported is higher i bet the bending stresses are still going to be a fair bit higher under those conditions, i'm not convinced a little bit non concentricity is a show stopper, also when the valve get slammed on the seat there is going to be way more stress

Do you have stem in a fixed position in your model? Add .0008" clearance for valve guide and stress will be even less.
[/quote]

yes there is essentially no guide clearance, adding clearance would reduce the stress but add an additional complexity to the model that i didnt have time to play around with. many are using bigger stems which would increase the stress so the order of magnitude is probably fairly representative.