What's new
What's new

Size is jumping during cylindrical grind

I do stick the wheel after I dress with the steel plate. I don't stick the wheel after grinding carbide parts. Wheel looks open to me (correct me if I am wrong).
Try a stick at each part. See what happens. But normally this solves a part growth problem, not shrink.
Can't really see the diamonds and bond tails at this magnification.
And then there are sticks. Type, grit, hardness, how applied. We have 6 different ones.
Normal wheel stick is plunge and removes about .100/.200 from the end of the stick.
Check with the DTI in grind cycles. If wheel load is up or down the deflection of all involved varies.
Yes your non-loaded static test should tell about encoder and such if it mimics the actual grind cycle moves.
Bold print here. Have had stuff that would rapid in, touch rapid out and check fine. But the actual cycle was crawl in and then rapid out with different results.
This was motor coupler.

One can buy a tip touch temp probe that plugs into a any voltmeter for low cost.
 
Is the size shrink repeatable?
Sometimes we just chart this and comp at each part knowing it will move this way or that into a run.
Ever tried to run the same part twice or three times without removing it? Where did it's size go? Able to measure in the machine without removal?
Parts are expensive here so I'd be doing the testing at oversize so that they can be saved later.
My guess is that this is a 1.2 lb piece of custom carbide. That has to have a big price tag.

Flexing/movement test.
Bring your wheehead in against the DTI. Push back on the wheel with about 20-40 lbs of force. Change?
Grind part. Put the DTI on the ground section push-pull on the faceplate the same. Change?

Another thing to watch is spindle power. This can sometimes tell you the force of the wheel against the part.
Sometimes more sparkout revs at depth help this settle down.

It sounds like the vise has solved your part movement problems. If you can live with locating this way it is for sure the best.
Your parts are out of the furnace. I'd maybe suggest sand blast or rubbing on a piece of sandpaper on a plate before loading for a better surface.
Also you may see black specs in the skin or powder on the parts if not cleaned.
This is graphite that is sprinkled on the plate so that they will not stick during sintering. Best to get rid of it.

Frustrating... so far I am of no help :bawling:. You did throw me a couple of curve balls here.
 
Last edited:
Is the size shrink repeatable?
Sometimes we just chart this and comp at each part knowing it will move this way or that into a run.
Ever tried to run the same part twice or three times without removing it? Where did it's size go? Able to measure in the machine without removal?
Parts are expensive here so I'd be doing the testing at oversize so that they can be saved later.
My guess is that this is a 1.2 lb piece of custom carbide. That has to have a big price tag.

Flexing/movement test.
Bring your wheehead in against the DTI. Push back on the wheel with about 20-40 lbs of force. Change?
Grind part. Put the DTI on the ground section push-pull on the faceplate the same. Change?

Another thing to watch is spindle power. This can sometimes tell you the force of the wheel against the part.
Sometimes more sparkout revs at depth help this settle down.

It sounds like the vise has solved your part movement problems. If you can live with locating this way it is for sure the best.
Your parts are out of the furnace. I'd maybe suggest sand blast or rubbing on a piece of sandpaper on a plate before loading for a better surface.
Also you may see black specs in the skin or powder on the parts if not cleaned.
This is graphite that is sprinkled on the plate so that they will not stick during sintering. Best to get rid of it.

Frustrating... so far I am of no help :bawling:. You did throw me a couple of curve balls here.
Here is a quick update I just did... I do not have a way at the moment (waiting on gage) to measure the part without removing it from the machine. So the next best thing -- I ran the same part 4 times without moving anything, just pressing the start button. I marked the part with sharpie every time I ran it, and it cleaned up the sharpie every time, completely cleaned it up. Does this mean anything? Technically the sharpie is adding thickness, but the way it cleans up and when it touches in the cycle makes me think it is "growing" somehow.

Very custom part.

I pushed against the wheel with all I got, and it moved the DTI maybe 5 tenths, very hard to measure though without rotating the wheel which causes the needle to jump around.

I pushed against the part/vise/faceplate as well, 1-3 tenths with all my force. Again alittle hard to measure.

Surprisingly enough this machine does not show spindle load.

I am sparking out 12 revs during the rough grind and another 12 revs for the finish grind.

The Vise is exponentially better at the moment, it has eliminated the runout issues completely. Setup is slightly longer, but still under 1 min.
 
Crazy money but Etalon or the ilk indicating mics.
Sensitive to holding in such a part but gives you some idea where you are.
Not sure as the boss so short if these will reach in so maybe a stupid option.
One can make new tips but that not easy.
How long in seconds is 12 revs?
 
Last edited:
Here is a quick update I just did... I do not have a way at the moment (waiting on gage) to measure the part without removing it from the machine. So the next best thing -- I ran the same part 4 times without moving anything, just pressing the start button. I marked the part with sharpie every time I ran it, and it cleaned up the sharpie every time, completely cleaned it up. Does this mean anything? Technically the sharpie is adding thickness, but the way it cleans up and when it touches in the cycle makes me think it is "growing" somehow.
I kept running the same part, backing off the wheel 0.0001" every cycle. At cycle 6 (+0.0002") there were small spots of sharpie left, but still mostly cleared up. Cycle 7 (0.0003")
 
I would get that temp probe. Check your wheel. If it's rubbing due to a poor dress/sticking, perhaps it may be getting hot and growing. Check part too.

About how much time is represented along the bottom of the graph?
 
Last edited:
I would get that temp probe. Check your wheel. If it's rubbing due to a poor dress/sticking, perhaps it may be getting hot and growing.

About how much time is represented along the bottom of the graph?
Hmm. About 5 mins per part, so 90ish minutes including setup.
 
You could probably eliminate this as a possibility (or not) by just touching the wheel and seeing if it feels hot. The average trend is toward smaller, but that's a good bit of bouncing around up and down, too.
 
Ground Dia. is 1.5000 at post #1.
What is this 0.9985? I am missing something.
Just trying to use scrap. They’re pieces that are undersized (1.30”) from previous tests and I target a nominal diameter taking off .030 grind stock to simulate the real part. In this case I’m putting in X coordinate (1.0”) and running 14 parts just pressing cycle start.
 
I am going to go way out on a limb here and into silly land.
The trend makes some sense. The variation off the trend line does not.
So the hard question is actual measuring.
If you take the same parts and mix them up (keeping track of which is which) and send them to be measured over different days is there a different number on a second or third check?
I know this sounds crazy but one needs to know how much to trust the measuring process and allow for it's float.
I may be all wrong here. Just throwing it out for comments.
Bob
 
If you have a carbide part against a hard steel plate(face plate) with a center point applying pressuer at perhaps a 1/16 (or what) diameter the part is most likely moving during grinding, so you are not getting any closer to success..

If size is wondering about it is most likely because your part is wondering about.

it is doubtabll that your wheel head or work head is changing. I can run a half thow on my least grinder with the only changing being a constant change due to wheel wear.

and measuring within .0003 should be pretty easy with simple gear.
 
Last edited:
Can you pause for a few seconds and then take .002 in slow feed?

I used to give a tap with a small taping hammer when I had odd happenings when grinding, to be sure my part could not easily move..
 
Last edited:
Yes I was surprised too. Most of mine take about 1/2 to 1 hour from cold start.
Sometimes it does get up into close to two hours. Some are happy in 15 minutes while doing sub 1 minute cycles.
Seems It is a decreasing curve before it goes the other way I am told. (This the same with mine)
He is working in hard steel now but spent his early years in a carbide shop.
His stuff is tighter tolerances than here and I trust him.
At the age of 18 he learned gauge R&R, SPC control, machine response to runs and what happens if you go to lunch.
He knows how it acts and they just keep comping bigger and bigger before the next part as things warm up. Then it goes the other way.
When I asked he spit out all sorts of Gxxx codes for different Kellenberger machine grind cycles but I was "do not know these".

Trend or best fit line on the chart above is nice but the flyers are not.
Wonder if more sparkout helps? At 120 grit and this type grind cycle is the surface finish great?
 
Last edited:
Good morning All,

Sorry for the absence -- I had a technician come in and do a PM on the machine. I was not aware of this fact, but the Motor is actually water cooled. There is an outlet nozzle for a visual check on whether you are getting coolant flow into the motor.

Well, we found out that I was getting water into the wheel head but not out of the wheel head, so the technician deemed that it was clogged with rust and carbide (We could not get the motor out without removing the wheel head... Took 24,000lbs on a press to get the motor out). 2 Weeks later and the machine is back apart, and I am cooling the motor properly. Other than the cooling the machine was in great condition, ways, ball screws, etc.
CloggedMotor.jpeg

The technician says that the lack of cooling could've be part of the problem I was having with the size inconsistencies due to thermal expansion of the wheel head from the motor heating up... What do you guys think? Is that plausible this was the cause of all my issues? I guess there is only one way to answer that properly. Time to run parts!
 








 
Back
Top