Measuring distance between crests on keyway with angled walls.

[dandrummerman21]

I have a part that has keyways in it that have a 7 degree taper in the walls. I'm looking for ways to check it. We do have a CMM but I'm looking for a 2nd way to check it, out on the shop floor.

The part is 1-3/8 diameter with a keyway that is .210 nominal at the crests, and .140 deep at the crests (~.148 deep from full diameter). They gave a +.002 tolerance on the .210 dimension at the top of the crests.

What other options do I have to check it? I was trying to come up with a way to drop a ball/pin in it and measure the height but I think I would need 2 different diameter pins/balls and compare heights to calculate the wall angle and chord distance? That's more math heavy than I would prefer for my operators. I'd also have to make or modify pins, the keyway isn't long enough for gage pins.

I also have optical comparators with "face reflectors" that I could try.


My feeling is that the +.002 tolerance is too tight to rely on just using the end of gage pins or gage blocks to judge the width of the slot. Pins have small edge break, blocks even more so. I could grind the end of some pins down maybe? I haven't set up or made any parts yet, so maybe it might actually work but I don't have one yet to even try.

They will be coming from the grinder on friday for me to mill..

Ideas?

Thanks.

 

[Milland]

I have a part that has keyways in it that have a 7 degree taper in the walls. I'm looking for ways to check it. We do have a CMM but I'm looking for a 2nd way to check it, out on the shop floor.

The part is 1-3/8 diameter with a keyway that is .210 nominal at the crests, and .140 deep at the crests (~.148 deep from full diameter). They gave a +.002 tolerance on the .210 dimension at the top of the crests.

What other options do I have to check it? I was trying to come up with a way to drop a ball/pin in it and measure the height but I think I would need 2 different diameter pins/balls and compare heights to calculate the wall angle and chord distance? That's more math heavy than I would prefer for my operators. I'd also have to make or modify pins, the keyway isn't long enough for gage pins.

I also have optical comparators with "face reflectors" that I could try.


My feeling is that the +.002 tolerance is too tight to rely on just using the end of gage pins or gage blocks to judge the width of the slot. Pins have small edge break, blocks even more so. I could grind the end of some pins down maybe? I haven't set up or made any parts yet, so maybe it might actually work but I don't have one yet to even try.

They will be coming from the grinder on friday for me to mill..

Ideas?

Thanks.

As long as the angle of the cutter you're using is known good, shouldn't a specific ball bearing that fits the slot work? At the high end it's one dimension, at the low it's another bound value, so it becomes a measured "Go, No-Go". A micrometer measurement from shaft to ball top should be reliable with a skilled hand.

If a ball can't be found that doesn't hit the bottom of the groove, just grind a flat to remove it, then have a few marks on the ball near the "top" to ensure the flat does not intrude on the walls when inserted.

 

[jccaclimber]

As above, if your angle is consistent (or can be measured in a different way) then over ball/over pin with a single pin is the way to go.
You could verify this by figuring out how different the smaller pin is from the larger one at the extremes of acceptable angle.
If the conditions allow, I've also made a spreadsheet that does the math for you and left a (protected) laptop on the floor. Type in the numbers, spreadsheet goes green or red and tells you which direction to go.

 

[CarbideBob]

As long as the angle of the cutter you're using is known good, ...

What would be considered "good" on the angle in the cutter?
Would they all be the same?
Shallow angle a few microns makes the ball move a lot.
Bob

 

[Milland]

What would be considered "good" on the angle in the cutter?
Would they all be the same?
Shallow angle a few microns makes the ball move a lot.
Bob

Agreed, the cutter would have to be qualified and monitored for wear. As for the micron depth variations, a similar range of "good, not good" depths from ball top to groove bottom should show if there's been a change in side angle. While 7 degrees (per side, I presume) isn't steep, it's not 1 degree either...

I'm also wondering if I'm misunderstanding the question - is this a single groove in a part, or multiple (like gear teeth), and angular spacing from groove to groove must be checked also?

 

[dandrummerman21]

Thanks for the replies.

It is 7 degrees per side, with +/- .5 degree tolerance on the wall angle. There are 2 slots in the part, about 60 degrees apart. They resemble keyways except for the angled walls. So not gear teeth.

I'm feeling better about measuring over a ball/pin after qualifying a part with other methods, so the measurements agree.

When drawing up in CAD, a pin that's .200 diameter just barely misses the bottom of the slot (although I like the idea of grinding a flat so it doesn't hit, If i had to. thanks milland), a .210 pin nearly hits the points (drawn at .211 nominal center of the tolerance). So maybe I'll use a .208 or so for a quick check. We'll see how well it matches up with what the CMM says.

I just wasn't sure if there was some fancy hand tool made to measure this kind of thing accurately.

 

[guythatbrews]

Measuring over a pin the measurement change from high to low limit width by .0162 for 7*, .0176 for 6.5*, and .0152 for 7.5*. That is assuming 14* included angle, even more pin movement for 7* included. So that check is reasonable. I'd cut off a gage pin instead of a ball. Easier to handle and more measuring contact area and they are cheap. A ball will be affected more by small surface imperfections. Use the biggest pin you can so the tangent point is closer to the feature than the middle or bottom.

Wouldn't bother with two different size pins. Then you are also checking the angle. Or maybe DO use a matched set of two different size pins to check the angle, since the angle is hard to check since it's so short. One close to the keyseat top, one ground away to get close to the bottom. Haven't done the math though. The two pin difference may to too small to measure reliably.

Good idea to double check another way, cmm or comparator, to confirm the measurement matches.

 

[michiganbuck]

Having an optical comparator and any kind of grinder it would be pretty quick
work to make a go/ no go and dead-on gauge.
Perhaps an hour to make 3 flat stock gauges with a surface grinder.
+- .001 is a lot of size for a sight gauge.

 

[Crimi Giuseppe]

See if this system suits you.
One cylinder is enough, but with two the measure may be faster

 

[guythatbrews]

See if this system suits you.
One cylinder is enough, but with two the measure may be faster

Great method to find the corner as shown. Next time I need to find a corner I'll do this. Always a good idea to block open the other side of the vise with a piece the same thickness when you are using the edge of the vise. The vise holds better and less chance of tilting the part.

This method is harder to implement on the curved surface of a shaft and you can't check the part on the bench.

 

[dandrummerman21]

That is a neat idea for square corners. I'll have to remember that.

Won't work for my current application very well.

 

[Crimi Giuseppe]

The big shot:
a) you can build yourself a cylinder with the angle you have.
b) Mount two parallel cylinders, (motorcycle chain model.) Every time you have a different angle you have to rebuild only the two support plates of the cylinders

 

[Milling man]

@dandrummerman21 As far as I understand, you will do a lot of such details? Maybe it makes sense to modify the universal manual goniometer? Weld two narrow strips to it, then grind them and calibrate the goniometer scale once. Then just write down for the operators the range of values in which this angle should be.
For example, something like this https://www.aliexpress.com/item/400...00029291554919!sea&curPageLogUid=FHG2W5P6ha5i

 

[dandrummerman21]

2 things:

1: I've never seen that tool before in my life, and I'm about to leave so I haven't researched it any. I'll look up how to use it later when at home, probably. Fascinating looking tool, looks way more useful than the regular ol' protractor

2: From the 1 minute I stared at it, I don't see how that allows me to measure the crest width on this shaft. Maybe the angle only.

But I've got a "big" tolerance of +/- 30 minutes on the angle, and the tool (which I received today) looks really good. I'm not too concerned with the angle, more so with the distance at the top, and I cut a sample part today (didnt receive my parts from grind like I expected today). I feel really good about measuring a pin (currently using .210 diameter).

A .001 change in width of the slot equates to a .004" change in measurement over the pin to the other side of the part. And the comparator seems to agree with my mike. Didn't CMM it yet.

Note that I can only use the comparator on the "top" of the part. The slot is contained on both ends, about 1.5" long. So I can't view a section view of the slot unless I actually sectioned it.

 

[Milland]

You're probably already doing it, but I'd rough the slot with a stub 5/32" / 4mm EM, then just use the tapered cutter for finishing. I'd even "step" it with a higher rough pass too.

 

[dandrummerman21]

You're right, I am doing that.

Mostly because the material will be coming in 40-44hrc. And I want to baby the cutters just in case. I didn't mention it because it isn't important, but the bottom of the slot has a .030" radius, and I had to have the tool mfg add that radius.

I was surprised that I couldn't find anywhere that stocked angled endmills with radii on them (besides ball endmills. But I didn't want a 1/16" ball endmill with 7 degrees for deflection reasons.)

 

[75sv1]

Could you take a 'cast' of the slot? Then use an optical comparator? There are products out there for this.

 

[dandrummerman21]

Could you take a 'cast' of the slot? Then use an optical comparator? There are products out there for this.

Maybe.

Currently, going with mike over a pin method. Comparing it with the CMM and comparator (looking at the gap from the top), they all agree on the width within a few tenths.

I also did what michiganbuck suggested, made some flat gages, starting from .208-.214, incremented in .001. They work okay and I trust my feel with them. But I don't trust the operator with them. So I'm having the operator check over a pin and the flat gage. Told him that a flat gage that is .001 under the calculated size must fit, and it seems okay.

 

[75sv1]

We do ball over ball on cones. So, different diameter pins would be good. Probably line check, then CMM one every so often.