What's new
What's new

Grinding Shims

HMT8

Plastic
Joined
Jan 15, 2021
Location
Southeastern Indiana
Looking for some advice. I have a project that requires me to grind a bunch of 3/16 thick shims. The material is 4140 PHT and the sizes range from 4x4 to 1x1 ish. Each shim is ground at assembly and I need to be able to hold plus or minus .0002 inches on thickness. Since it is done at assembly, each shim requires a unique thickness. Each shim usually only needs 1 to 2 thou taken off total. This operation takes forever on a surface grinder. Any ideas how to do it faster? Vertical head surface grinder maybe? I have about a thousand shims to do.

Thanks in advance.
 
Nice to have a bigger machine so one can walk away and do other things. if the price is OK farm it out to a shop with a bigger grinder.

RE: Since it is done at assembly, each shim requires a unique thickness. *Cant send out..
 
Your thickness tolerance makes me wonder if the shims (especially the larger ones) might also need to be tapered.

In nearly 40 years of shimming "alignment interfaces" of aerospace structures I've seldom encountered as-found interfaces that are parallel within one part per thousand, and forcing one part per thousand tapered gaps closed often creates enough distortion to impair functioning.
 
Each shim usually only needs 1 to 2 thou taken off total. This operation takes forever on a surface grinder.

Define "forever"?

One side or both?
(IOW, are the shims as received flat?)

What is your process?
wheel spec and dia.?
coolant?
method of grinding?
auto 2 or 3 axis?

Were you practiced at grinding other tooling flat/parallel before being assigned the shims?

smt
 
Your thickness tolerance makes me wonder if the shims (especially the larger ones) might also need to be tapered.

In nearly 40 years of shimming "alignment interfaces" of aerospace structures I've seldom encountered as-found interfaces that are parallel within one part per thousand, and forcing one part per thousand tapered gaps closed often creates enough distortion to impair functioning.

That's a good question. The answer will probably depend on what he's shimming. If the shims really need to be within .0002" on size, you may well be correct. (As usual!)
 
If you are going to live on the front lines of fussiness, you are going to have to pay to be there. I have to wonder what type of assembly requires a shim flat to .0002 and made of hardened steel? Could you sandwich various thicknesses to achieve the correct stack? Blued tempered 1095 shim stock is available in a wide range of thicknesses and could easily be laser cut to your shape. You would have to check hardness, but should be comparable to your pht.
 
The process may speed up the job. Rather than running one-ups you might fill the SG chuck and with setting your stops to grind one part at a time, grind the highest part and then go to the next size, using individual spotters so having no pull and test/measure parts. Yes, you may need set parts 1" apart(?). Make a gauge jig/fixture so measuring is as fast as can be.
Don't pre-measure parts, but use your SG dial as the measuring device.
Wet grind to save swelling.
Down-grind to +.0005, and use one incremental crossing.
Choose the open wheel that just makes surface finish needs.
Have a spotter on the chuck that is just for setting zero
+- .0002 is a wide range for a good SG hand.
Have a size-set dressing diamond off the chuck set to +.001 the highest part.
Sometimes going to a harder but more open wheel can be good for pre-hard.
Determining the dress schedule, perhaps dress at every x number of parts.
Have your gauge set-up on a push truck at your side.
Use a vertical set big dial, dial indicator with a set of well-marked gauges.
Have a set board so missing the mark set that part on the board in its place perhaps .187 .186 .185.. but you should not miss the mark.
You might price having the first side Blanchard ground.

I think reading the down dial is faster than using a DRO, the reaching up to set is a big time consumer. A very big asset is a set zero hand wheel down feed dial.. with not having that I would have a slip of masking tape with a marked line to set on the down dial for zero(where I dressed) so going off the spot x number of thousandth made my size.
Q: what wheel are you using
Q: about how much time per part
Q: what is your Surfaace Finish spec
*looks like you haven't listed your location.
Buck
 
Last edited:
The process may speed up the job. Rather than running one-ups you might fill the SG chuck and with setting your stops to grind one part at a time, grind the highest part and then go to the next size, using individual spotters so having no pull and test/measure parts. Yes, you may need set parts 1" apart(?). Make a gauge jig/fixture so measuring is as fast as can be.
Don't pre-measure parts, but use your SG dial as the measuring device.
Wet grind to save swelling.
Down-grind to +.0005, and use one incremental crossing.
Choose the open wheel that just makes surface finish needs.
Have a spotter on the chuck that is just for setting zero
+- .0002 is a wide range for a good SG hand.
Have a size-set dressing diamond off the chuck set to +.001 the highest part.
Sometimes going to a harder but more open wheel can be good for pre-hard.
Determining the dress schedule, perhaps dress at every x number of parts.
Have your gauge set-up on a push truck at your side.

looks like you haven't listed your location
Your thickness tolerance makes me wonder if the shims (especially the larger ones) might also need to be tapered.

In nearly 40 years of shimming "alignment interfaces" of aerospace structures I've seldom encountered as-found interfaces that are parallel within one part per thousand, and forcing one part per thousand tapered gaps closed often creates enough distortion to impair functioning.

No
Your thickness tolerance makes me wonder if the shims (especially the larger ones) might also need to be tapered.

In nearly 40 years of shimming "alignment interfaces" of aerospace structures I've seldom encountered as-found interfaces that are parallel within one part per thousand, and forcing one part per thousand tapered gaps closed often creates enough distortion to impair functioning.
The shims are to position parts on hydraulic clamping fixtures for some 4 spindle machining centers. The part datums (where the shimmed bits actually touch the part) are all less than .25 inches square so we don't have to worry about taper (within reason).
 
If you are going to live on the front lines of fussiness, you are going to have to pay to be there. I have to wonder what type of assembly requires a shim flat to .0002 and made of hardened steel? Could you sandwich various thicknesses to achieve the correct stack? Blued tempered 1095 shim stock is available in a wide range of thicknesses and could easily be laser cut to your shape. You would have to check hardness, but should be comparable to your pht.
The hardness isn't really an issue. I think stacking up shims would be more difficult than grinding to thickness. Thanks for the idea though
 
Michiganbuck, thanks for the process details. There are some good ideas here that would help if we end up surface grinding them. We have a decent process for grinding them, holding the tolerance isn't difficult. it just takes too long to get across a shim with a surface grinder.

How much stepover would you use for the final grind?
 
Define "forever"?

One side or both?
(IOW, are the shims as received flat?)

What is your process?
wheel spec and dia.?
coolant?
method of grinding?
auto 2 or 3 axis?

Were you practiced at grinding other tooling flat/parallel before being assigned the shims?

smt
Current process is as follows:
Grind shims flat (oversize 3 or 4 thou)
Assemble fixture and measure
Disassemble and grind as close as we feel comfortable.
Assemble and measure - then usually do the same thing for a final grind.

Measurements are made on a CMM - no way to easily measure all of the positions of the datum pads any other way.

Mitsui auto 3 axis surface grinder - not sure of the wheel (I don't personally grind these). Wheel choice seems good though.

It's a long process anyway you look at it, but If I can grind a shim in considerably less time it adds up pretty quick.
 
Thinking about his a bit, we end up grinding each shim 4 times.

(2) both sides for initial cleanup
(1) semi finish
(1) finish

The initial cleanup grinding is done with a full chuck, but after that it's each one individually - for about 1000 shims. Yikes.

We've done it before, and it took longer than I expected. Now I realize why!

Could maybe eliminate the semi finish.
 
OP Q: (How much stepover would you use for the final grind?)
Stepover is very slow, with the right wheel you should be able to down-grind to need only one step-over pass of .0003 to .0005

QT:
Thinking about this a bit, we end up grinding each shim 4 times.
(2) both sides for initial cleanup
(1) semi finish
(1) finish

Re:
(1) Seems that you should grind one side for flat & surface finish.
(2) and one grind to make size

*unless the part is prone to release stress warping.
 
Last edited:
This might be a stupid suggestion, but do you have enough data from your previous work to have a distribution for the shim sizes that you've needed?

Grinding each shim four times sounds like and awful lot of work. Your tolerance is +-.0002, so .0004 total, and the total span is only a few thou if I read you right. So really only about ten sizes. It sounds like this is something in series production.

Why not just make all the sizes once, rather than faffing about with each shim four times? Lay out as many blanks as fit on your table. Grind to the max size you might need. Take a set off. Grind off another .0003. Take a set off, etc.

Then mic the shims, write the size on each, hand them to assembly and they can mix and match once. Every now and again they'll run out of the sizes in the middle of your distribution, so make another set of those.

That has to be cheaper and faster than grinding them one at a time, one up, multiple times per shim and multiple shims per assembly.
 








 
Back
Top