Valve Seat Concentricity Shocker

[David Vizard]

When it comes to valve sizes I often find the need to machine a set down to the required size from something just a little larger. This means being able to cut new seats.

I do this on my lathe with a dedicated precision collet set up that is installed on the lathe head and adjusted so as to hold a valve stem typically to 0.0002 or less in terms of concentricity. Every valve I machine I check out the seat runout on my Goodson seat checker. I don’t like to see any more than about 0.0003. The valves I use out of the box, mostly Ferrea, typically are zero to about a tenth of a thou runout so that, so to speak, is my model.

Well the collet setup suffered an as yet unresolved problem and just went to pot with runout errors as much as a thou but typically around 6 -8 tenths of a thou. I am now in a state of high concern. This prompts me to call a friend who runs the machine shop in a well-respected company. So over I go with a set of valves and my Goodson concentricity checker. A valve is ground and checked – 2 thousandths out!! Worse than the ones I feel are a problem. I go to another shop – 3 thou out then another 1.5 out.

Three shops that build hi-perf motor and have not seen any valve leakage when the assembled heads are vacuumed.
I am just wondering how many shops actually test their valve seat grinders for concentricity!!!

Right now I am back to searching for the problem with my lathe setup as I see I really need to do my own valve seats.
Any comments on what the limits of eccentricity may be before a leakage is seen???
DV

[PackardV8]

Hi, David,

Too many variables to give a "to-the-tenth" answer.

Diameter of valve head, diameter of stem, radius of head-to-stem, valve material, distance from guide-to-seat, stem-to-guide clearance and spring pressure, among others, can change the equation.

Maybe, ask a specific question, "On a typical SBC 2.020" 11/32" Ferrea valve, with 150# seat pressure, with .0015" guide clearance at how much run-out do you see leakage at 30" vacuum? How little run-out do you typically see and accept on a performance valve job?" Or design your own specificity. It would help those who answer and those of us who read the answers

[Schurkey]

I am a little confused. Are you having problems with valve faces (on the valve) or valve seats (on the head)? I get the sense you're talking about valve problems, not head problems. Is my terminology incorrect?

The valves I grind with my ancient Sioux typically test at less than half-a-thou, and often about 0.0003. I'm still dial-indicating each one--I don't do enough valve jobs to be comfortable with "trusting" the machine. This is "chucked in the machine, magnetic-base dial indicator", not tested with the lovely tool you've pictured. Clearly, my Sioux cannot provide the utmost precision.

I checked some Famous-Name stock-replacement valves for my DOHC Lumina, and the new valves were worse than the originals, and worse than my regrinds. About 0.0007, if I remember correctly, using the same procedure and equipment.

Valve seats in the head I can't make any claims for. I haven't used it enough to get comfortable, and it's ancient pilot-and-stones stuff that is probably not capable of real precision, at least without making a career out of each seat.

I don't think that concentricity problems with valve faces and head seats combined, totaling less than the valve guide clearance will show up on a sealing test. I'd expect concentricity to be more of a durability issue than a sealing issue--unless things are WAY goofy.

Have you verified that the problem is with the valves and therefore with the cutting equipment, and not some unresolved issue with your measurement tool? A spot of dreck on one of those ball-bearings would make all the difference...

[David Vizard]

I am a little confused. Are you having problems with valve faces (on the valve) or valve seats (on the head)? I get the sense you're talking about valve problems, not head problems. Is my terminology incorrect?

The valves I grind with my ancient Sioux typically test at less than half-a-thou, and often about 0.0003. I'm still dial-indicating each one--I don't do enough valve jobs to be comfortable with "trusting" the machine. This is "chucked in the machine, magnetic-base dial indicator", not tested with the lovely tool you've pictured. Clearly, my Sioux cannot provide the utmost precision.

I checked some Famous-Name stock-replacement valves for my DOHC Lumina, and the new valves were worse than the originals, and worse than my regrinds. About 0.0007, if I remember correctly, using the same procedure and equipment.

Valve seats in the head I can't make any claims for. I haven't used it enough to get comfortable, and it's ancient pilot-and-stones stuff that is probably not capable of real precision, at least without making a career out of each seat.

I don't think that concentricity problems with valve faces and head seats combined, totaling less than the valve guide clearance will show up on a sealing test. I'd expect concentricity to be more of a durability issue than a sealing issue--unless things are WAY goofy.

Have you verified that the problem is with the valves and therefore with the cutting equipment, and not some unresolved issue with your measurement tool? A spot of dreck on one of those ball-bearings would make all the difference...

Schurkey,

Thanks for your comprehensive reply.

I am talking about the valve seats here. Your 'old' Sioux is doing way better then most - suggest you hang on to that baby if it is clean and paid for.

I typically don't vacuum test my seats but I always indicate both those in the head and on the valves for concentricity.

I think the point I was trying to put over is how many home engine builders go into a machine shop to have the valve seats cleaned up only to get an eccentric job done. Could be something that they never check.

DV

[GARY C]

Then the next question would be, how does a cold head on a bench compare to a tq-ed head on an engine at operating temperature?

[ProPower engines]

When it comes to valve sizes I often find the need to machine a set down to the required size from something just a little larger. This means being able to cut new seats.

I do this on my lathe with a dedicated precision collet set up that is installed on the lathe head and adjusted so as to hold a valve stem typically to 0.0002 or less in terms of concentricity. Every valve I machine I check out the seat runout on my Goodson seat checker. I don’t like to see any more than about 0.0003. The valves I use out of the box, mostly Ferrea, typically are zero to about a tenth of a thou runout so that, so to speak, is my model.

Well the collet setup suffered an as yet unresolved problem and just went to pot with runout errors as much as a thou but typically around 6 -8 tenths of a thou. I am now in a state of high concern. This prompts me to call a friend who runs the machine shop in a well-respected company. So over I go with a set of valves and my Goodson concentricity checker. A valve is ground and checked – 2 thousandths out!! Worse than the ones I feel are a problem. I go to another shop – 3 thou out then another 1.5 out.

Three shops that build hi-perf motor and have not seen any valve leakage when the assembled heads are vacuumed.
I am just wondering how many shops actually test their valve seat grinders for concentricity!!!

Right now I am back to searching for the problem with my lathe setup as I see I really need to do my own valve seats.
Any comments on what the limits of eccentricity may be before a leakage is seen???
DV

The biggest issue I see here is the way the valve is supported in the checking fixture. it does not appear to hold the tip centered in the fixture while supporting the valve stem centered near the valve head at the end portion of the stem that would normally be supported by the guide.

When I cut down valves I will do just the same as you do but I never rely on a machined valve face as it seems you are doing here. unless the stem is dialed in for run out before grinding the valve face run out IMO will always be an unwanted issue.

I machine my valves to within .010-.015 of finished valve face to tip length to start with before grinding the face to final size.
Then with a set up valve I will set up the dial on the feed to "0" and then true the face to the desired length by grinding.
I then re "0" the feed dial and then check each and every valve in the chuck for stem run out at the valve head end of the stem and correct the run out by adjusting the chuck as needed till the run out is less then .0005 most times I can get to
.000-.0003 then using the feed dial as a guide grind the face till I get to "0" and spark out till I don;;t hear any stone contact.
I do 8 intakes valves and re-dress the stone then do the 8 exhausts every time for the best possible finish on the face.

I use Sioux valve grinding machines and I have not found more user friendly machines.
I have kwik Way machines as well but I use them for specific jobs but I do have 4 Sioux machines just because I have not found any other machine that allows the operator to grind a set of valves easily and quickly that will keep the valve tip to valve face length with in .001 valve to valve every time. And this comes in very handy when doing head work on non adjustable stuff keeping the tip as tall as possible with out the need to remove a bunch of material to keep the tips all the same height. And very helpful when using a +.050 valve lock and a lash cap as there is just enough to use then on a std. tip without and removal of the tip to correct the installed tip heights.

As for the valve seats in the heads I use a Sunnen VGS-20 and I get .0003 or less run out in the seats in the head.

I realize that valve stem clearance also has a lot to do with it but since I started honing guide I have stayed with the same practices and have seen seat and valve face wear reduced by 1000 times when doing a freshen up and guide wear also reduced by keeping the runout to an absolute minimum on both faces.

[Walter R. Malik]

Realistically ... as long as the valve is truly round then the concentricity can be out as much as the guide clearances will allow.

I see it everywhere with a lot of shops I visit. Even those who are supposed to be specialists with cylinder heads.
If you can not put a back-cut on a valve after removing it from the grinder and re-installing it because the seat width varies all over the place, that grinder certainly is not grinding concentric.
I would say that being a few tenths non-concentric is OK as long as it is round but .002" is not to be accepted.

[statsystems]

Realistically ... as long as the valve is truly round then the concentricity can be out as much as the guide clearances will allow.

I see it everywhere with a lot of shops I visit. Even those who are supposed to be specialists with cylinder heads.
If you can not put a back-cut on a valve after removing it from the grinder and re-installing it because the seat width varies all over the place, that grinder certainly is not grinding concentric.
I would say that being a few tenths non-concentric is OK as long as it is round but .002" is not to be accepted.

Yep. Most of the valve grinders I've seen won't put a back cut on top of a seat. I've only seen two I know of.

[Walter R. Malik]

Realistically ... as long as the valve is truly round then the concentricity can be out as much as the guide clearances will allow.

I see it everywhere with a lot of shops I visit. Even those who are supposed to be specialists with cylinder heads.
If you can not put a back-cut on a valve after removing it from the grinder and re-installing it because the seat width varies all over the place, that grinder certainly is not grinding concentric.
I would say that being a few tenths non-concentric is OK as long as it is round but .002" is not to be accepted.

Yep. Most of the valve grinders I've seen won't put a back cut on top of a seat. I've only seen two I know of.

I have a Sioux, 2 Quik-Ways, (one uses Lacey Williams collets), and a Black & Decker ... guess which one is the most concentric ?
The Black & Decker; every time.

[statsystems]

Realistically ... as long as the valve is truly round then the concentricity can be out as much as the guide clearances will allow.

I see it everywhere with a lot of shops I visit. Even those who are supposed to be specialists with cylinder heads.
If you can not put a back-cut on a valve after removing it from the grinder and re-installing it because the seat width varies all over the place, that grinder certainly is not grinding concentric.
I would say that being a few tenths non-concentric is OK as long as it is round but .002" is not to be accepted.

Yep. Most of the valve grinders I've seen won't put a back cut on top of a seat. I've only seen two I know of.

I have a Sioux, 2 Quik-Ways, (one uses Lacey Williams collets), and a Black & Decker ... guess which one is the most concentric ?
The Black & Decker; every time.

IIve used a B and D and it was good. There was a blue one bought brand new in about 1999 and that thing was spot on. Dead nuts on. I can't think of the brand. I loved that thing

[SupStk]

Concerning testing a valve grinders accuracy and repeatability, grind a valve, indicate, unchuck and give the valve 1/2 turn and rechuck. Reindicate and see what the runout is. Closer it comes to zero better the machine is.

Vacuum testing with springs on will cover alot of error. The springs will load the valve on the seat using guide clearance and deflection of the valve to indicate a perfect seal.

[Carnut1]

1103171708.jpg

I like my old Sioux. Fixed it up when I got it in the early 90's. To get an idea how low the hours are on it that is the stone I installed in the 90's! How accurate? If the valve seat measures concentric a valve that is not bent and ground on it will lap in in a second.

[Zmechanic]

David,

What brand collets and style are you using? If they are 5C, i'd highly recommend Hardinge collets if you aren't already using them. You might want to sharpie or dykem all the mating surfaces (collet to closer, closer to spindle, etc) and check the rub pattern to see if you accidentally raised a burr somewhere.

[HDBD]

I am not sure I understand the issue exactly.
I am very fussy with valves too. I cut the od on a lathe then move them to a sunnen vr6500. The Crawford multi bore system is very accurate. I keep the face to tip lengths plus or minus 1/2 thou which is easy using the same goodson checker after I grind them. But really a back type travel indicator such as the baker is a pretty crude gauge to expect anything in tenths. Put a test indicator on there for accuracy. An indicol, federal, or starrett. Or are you indicating the stem when it is chucked and that is the issue? Have you checked the collet system back to the lathe bore for runout? I think something may be wrong in the stack before the collet itself. What type of collets are they?

[dwilliams]

Then the next question would be, how does a cold head on a bench compare to a tq-ed head on an engine at operating temperature?

I wondered about that, tried to figure out a way to tell, and finally decided not to worry about it. If they hold vacuum, I don't care how round they are.

I've noticed that the first cleanup cut on a used head often shows the seat is considerably out of round, or perhaps offset from the valve guide bore. More than once I've re-cleaned the guide and turned the pilot 180 degrees, just to make sure I hadn't somehow bent the pilot.