Valve seat concentricity and dead pilots

[modok]

You have to check runout WITH a pilot.....so how can the pilot have "runout in the guide"? You spin the pilot and measure the runout? or are you just wiggling it and think you are seeing runout?
Or are you fighting a curved guide?
If you are noticing that a tapered pilot will be loose in the guide, well yes. A tapered guide in a straight hole is loose on the far size, because it is. If you are indicating as it goes in, seeing it go sideways as it is inserted, I think you are pushing too hard.
You can't FORCE it into the guide, as that will tend to bend it. Any bend in the guide will be magnified if the pilot is forced in. I rarely needed any tool to pull the pilot back out, besides my hand, if you know what I mean.

If you want to get the seats dead straight to prove a point to yourself, get or make a pilot that is THE size you want the guide to be, and size the guide to fit this pilot. No clearance- no wiggle-no problem
Hone until the pilot just barely fits through, or size with a diamond one pass hone, or solid carbide reamer. HSS steel reamer won't cut it. Always some bend and variance in size with HSS reamer although I will say the old straight flute k-line ones did very well, but a good HSS reamer must be a lost art because all the new ones seem to be worse all the time.

Sioux pilots are .0015 under the nominal size marked and are straight more than half way up. They taper larger only at the top, so they can act as straight AND tapered pilots. I personally think a straight pilot will work fine a lot of the time, if you have <.0005 clearance and put some heavy oil on it. Having it lock eliminates one variable and is necessary for some odd things, but I have ground seats and used a dead pilot machine with a pilot that was free and it seemed to work fine. Tapered plots shine for fitting tapered (used) guides, and that's great too. The more pilots the better! And yes, may need to level on every guide if they are all a little different, and than can be on some heads. The level is not necessarily better but it SURE faster way to setup most things, and it can be very accurate! The floating table has to be level to work anyway. A milling machine does not have to be perfectly leveled to work, but it is highly recommended.

[Dan Timberlake]

You have to check runout WITH a pilot.....so how can the pilot have "runout in the guide"? You spin the pilot and measure the runout? or are you just wiggling it and think you are seeing runout?
Or are you fighting a curved guide?
......................

===================

This post reminded me some boards have a nice feature to add a "star" to a particularly good post without adding another post just to say .

http://www.eng-tips.com/viewthread.cfm?qid=378451

[jcisworthy]

The guide is a new fresh honed bronze guide and is straight.

I installed the pilot in a fixture I use for checking runout on valves. I spun it in the fixture with a dial indicator on it and it had only .0001 runout.
I measured runout on the pilot in relation to the guide hole after it was installed using a coaxial dial indicator. I indicated the guide hole center, installed the pilot and indicate the outside of it. It had .004 runout compared to the zero I had on the guide hole.
I am gone for the weekend so I will figure it out next week and post what I come up with.

[Carnut1]

The guide is a new fresh honed bronze guide and is straight.

I installed the pilot in a fixture I use for checking runout on valves. I spun it in the fixture with a dial indicator on it and it had only .0001 runout.
I measured runout on the pilot in relation to the guide hole after it was installed using a coaxial dial indicator. I indicated the guide hole center, installed the pilot and indicate the outside of it. It had .004 runout compared to the zero I had on the guide hole.
I am gone for the weekend so I will figure it out next week and post what I come up with.

Possible the guide is not straight and twisting the pilot. Carbide pilots are tough and deflect very little but could be the cause. I am a honed guide fan. Properly honed guide is very strainght and your valve job will have good concentricity if done right and equipment is in good (not worn) condition. Thanks, Charlie

[Carnut1]

Has the guide been measured with a valve guide dial bore gauge?

[jcisworthy]

I use a diamond hone. The guides have .0016 clearance using dial bore gauge and they are straight.

Prior to this post I tried inserting the pilot straight in tight straight in tight with a twist, in snug and each time the pilot repeated .004 runout. It was consistent no matter how I inserted it but not straight in the guide according to the coaxial dial indicator and the deflection on the ball head tool holder when I brought the quill down.

My guess is thousands of dead pilots daily are inserted crooked but not noticed because they are not checked. As I said earlier, a floating seat and guide machine will float to the guide alignment so who would ever know in that case. I noticed it because I am on a mill and decided to start checking things out.

[cpmotors]

I will run the quill down with a valve stem in a collet after alignment and see if it engages the guide smoothly,

This will verify whether or not the pilot does not aling straight to the guide when I install it.

Your making this more complicated than you need too honestly. Cut the seat, check the runout/deflection direction, offset equal to runout, cut again and check. Put an indicator on the pilot and push it with your finger, it flexes all over the place. It's essentially just locating the bottom of the tool body.

I am going through all of this because when you get this retarded on set up and some seats still come in better than others I want to know why.

I am performing at least two checks to verify each particular operation so I can go through a methodical process of elimination to try to see something that will help consistency.

I was told to use a bubble level like Goodson VGL 100 to level the quill and head like you set up a Sunnen VGS machine and forget about it. I am starting to think he may be right.

Ill bet pilots are not going in the guide true everywhere people just do not know it because they nevel checked.

Set up on the pilot however it goes in the guide and cut a seat. That is how the floating seat and guide machines work.

I know a thing or two about being retarded on setups, I did everything your doing when I first started doing this.
Your pilot is straight, your guide is straight, your axis alignment is off.
A bubble level is pointless for doing this on a mill, even if the mill is perfectly level the guides won't be and you'll spend hours trying to jack around getting the guides adjusted to perfection.
Level the head along X axis by indicating it, indicate the pilot for valve angle, sweep the pilot for center with just a regular mag base and Last Word/Bestest, and then check axis alignment as I showed in the video. I cut seats to .001" or less all the time doing this, and so does every single customer that has bought my fixture for doing valve jobs in a vertical mill.
Well, all but the one guy with his pythagorean theorem arguement....

[Dave Koehler]

Forgive me if this has been previously mentioned.
The mill head level is the most important. It has to be on the same plane (trammed in) as the bed if leveling the mill via the bed.
If the head has been tilted to surface a head and not checked prior to seat work there is no hope of any of this working.
For the purpose of seat work, With the right level it is easier to level the mill head with a pilot in the spindle.

[modok]

I still don't get it.
If a pilot is or isn't straight in a guide has NOT something you check with a milling machine. That is one pilot verses another.

[jcisworthy]

I only have a few pilots because I use bronze guides in everything I do and hone to the same size each time. I only have three 11/32 pilots one of which is split and I don't use it.

out of the three the best register came from the split pilot at .002 runout.

The tram is good

I am going to using a level for quick close set up the fine tune it with the last word.

If the pilot does not install straight in the guide that is the problem in my view. What good is careful set up only to install a crooked pilot.

I am trying to reply the best I can on my phone I do not have access to my computer until tomorrow so I apologize if I am somewhat vague.

I appreciate all the input.

I think ultimately i will have to try a few different pilots and se if any register true to the guide

[Sir Yun]

On the risk of stating the bloody obvious/mansplaning

There are three axis. one through the guide, one through the pilot, one through the quill/tool axis.

the objective is to align them. (first zero out rotation, then translation)

If the axis of the guide does not coincide with the axis of the pilot, one side of the tool will start to cut before the other side introducing uneven cyclic loading and change the cutting geometry, ergo the pilot will bend (it will always bend with a single point tool, but hopefully reasonably symmetrical,) and as a consequence the hole/seat is not round.
Additionally ball head is like a universal joint not a CVJ so it will only tolerate a minor bit of misalignment before it starts to introduce all kinds of oddball behaviour.
When you use a ball head you are sort of using it like a reamer.
Just as you can't straighten a hole with a reamer, it will just follow the banana hole, it will follow the pilot axis ( of the part that sticks out).

So if you then stick an indicator on your quill and measure the seat it will be out (or not if the axis where aligned).

If you use an air float table (say because it is faster). It needs to be dead level (otherwise the weight slipping sideways will bend the pilot off axis) the axis through the guide needs to be perpendicular to the guide axis (the table can only translate not rotate, so a bubble level on a fitted dowel pin will help). Assuming that the guide axis is OK . When you then stick in a live pilot, it wiggles the head above the guide (like a Serdi ) and given a fairly good fit this works well enough. Then you lock the lot down and you are basically running a light mill.

Assuming your mill is properly trammed and level and all spiffy and slop free and you are able to line up the axis through the guide exactly with the axis trough the tool/quill why use a pilot and ball head to begin with ?
You have 2 tons of knee mill to keep things in place but you rather use a 6mm piece of carbide sort of supported on one end ??

There is no magic to it. Align the axis ( does not really matter how), then bore the hole/ cut the seat. If you don't, stuff WILL bend eratically and you will have a lot of unwelcome guests to your party

[jcisworthy]

Good point

[jcisworthy]

New bronze guides diamond honed to finish.

The tapered pilot only has a tenth of runout in it so the pilot is straight but when I insert it in the guide it has .004 runout where the ball head tool holder rides.

I am using a vertical mill with a rollover fixture. I indicated the guide with a last word then used a coaxial dial indicator to center and verify the indicated guide hole. I inserted the pilot with spring and ball head tool holder, rand the quill down and the tool holder deflected.

I ran the coaxial on the top of the pilot and it is .004 out. I tried taking the pilot in and out a few times and it repeats .004 out. I stuck an old split pilot I got with my valve grinder and it is better but .002 out.

Anyone make pilots that go in straight?

The pilot doesn't have runout, the guide isn't vertical/parallel axis to the quill axis. This is why I don't use a coax to center on pilots, you need to stroke the quill along the sides of the pilot in Y and X and see which way its leaning. Very few guides are actually perpendicular to the deck or parallel to each other. Trying to center off of the guide then inserting a pilot is just compounding the problem.

I can use a HSS Goodson tapered pilot, my own custom carbide pilots or Rottler pilots and get the same result. I do prefer the Rottler style with straight lower shank and the fast taper near the top vs the straight taper.

https://www.youtube.com/watch?v=FNckByP_TGw

I watched your video and that does not necessarily mean your seat is concentric to the guide. If your pilot does not go in the hole straight is surely is not. Seat run out in that case is not accurate.
I am indicating off the guide hole which is where I want my seat to be concentric to.

[Sir Yun]

Given a straight and round guide bore and a straight and round ground pilot with a tightisch slip fit the method shown will work fine to well within a 0.01mm If the pilot is tapered and has room you can stick it in lopsided and the method shown will only align the quill axis to the pilot axis. Try it. Stick a pilot in it and bump it gently sideways and stick the indicator back on and see what happens. A taper wil only fit a taper properly ( thats why they are on machines), so imho there is zero reason to use a tapered dead pilot in a good hole. Live pilot ok but only to translate :

If you indicate (or stick in a live pilot which has the same purpose) on the top of the inside of the guide but it is not vertical you are still not sure if you got the rotation out.

You need a dial test indicator with a long arm. You do exactly the same as is shown on the video but inside the guide.
Then stick in the pilot, repeat process. If it is still not concentric enough, that would be weird. then it is slop of the cutter on the pilot or/and cutters that are wrong causing it to wedge and sideload. Grinding the last bit is probably the easiest if you want real concentricity.

Ps the end of the guide needs to be clean and at 90 degree of the guides axis or it will cause pilot movement.

Pps


Given the fact that the whole lot will heat up 100C, enough is enough. Sure you could try to go for 3 micron total runout but i would be very surprised if that would still be the case once heated up and torqued down and assembled with a row of fat 300 lbs springs on it ) oh ,and the valve has to move as well.

[wyrmrider]

Exactly that's why the use of torque plates on some heads helps