What's new
What's new

Best methods for measuring Diameter/radius of a semi-circle

MotoX

Cast Iron
Joined
Nov 14, 2011
Location
Enid, Oklahoma
Just as the title states, I would like to know the various methods of measuring the diameter/radius of fixtures we have.

Some of these fixtures have been here longer than i've been alive. The fixtures are supposed to represent one-half of a bored hole split on C/L with a tolerance of +.001" , -.000" on diameter.

These are usually in the 2.5" to 8" diameter range.

These fixtures are used in a single axis production horizontal boring machine. Heald bore-matics to be exact.

I need to make new ones and also trying to minimize waste. It would also help to be able to determine wear on the fixtures down the road.

In the past, i have just made an entire fixture then turned one half into chips.

My predecessor used to use a gage pin and a depth mic, but I've come to realize this is prone to error.

Half-height of the fixture is not as critical as the actual bore. I need the bore to be as close as possible for the best possible outcome(within reason) of the parts.

The final cut on the new fixtures is bored in situ in the machine with a short gage length boring bar.

Some of the parts have a wall thickness tolerance of +.0000" -.0005". Normal tol is +0.000"-.001".

I would like to hear any solutions you may have. I have mostly just basic hard gaging, but i am not against buying whatever i need to accurately measure these.

Thanks
 
We don't know what equipment you have available but assuming you have some kind of metal working machinery you could mount the part on a machine, measure 3 points with say an edge finder or probe and use math to figure out the radius (put "solve for radius of a circle using 3 points" in google and find something that works for you).
 
To add a tool for following the above response, this is the best circle calculator I know of. If you can accurately measure any two things on your partial circle, you can solve for the rest. Might be worth trying a couple different ways to see if they come out the same.

 
you could probe it, like mentioned above. If you have access to a Haas with probing, you can just use a three point bore, the radius will pop out on, I believe, variable #188
Or, you could sweep it with an indicator, then measure the radius of your indicator against a vise jaw.
 
You can sweep a known inside diameter, maybe 1/2" but can be anything, with your dti and zero the indicator. With the dti set sweep the fixture and move whichever axis until dti is zero again. Repeat 3 times as far apart on the fixture as you can get. Calculate the radius from the 3 points and add half of the initial known diameter. If you can't probe or cmm it's a very easy way to get about as close as your machine is.
 
How many do you have? If it's more than just a few, i'd take them to a shop with a modern CMM. In a few minutes they can tell you how cylindrical they are, how close to true the radius is and what the average size is to tenths with much more certainty than manual means.
 
Thanks for the replies.

I do have a fairly modern(2018) doosan mill with probing(no scales), but i feel like i always get a little bit of error (±.001") when using it.

indicating and calibrating the probe often seems to reduce that, but its still always there as compared to hard gaging.

Last year i bought an old manual CMM machine that had a bad scale. I recently got a replacement scale for it, but have not had a chance to get it up and running fully and calibrated. I'm thinking i should prioritize that. However, i don't know how an old 3 axis manual cmm will probe an arc. I was kinda under the impression that this old of a machine will only probe in one direction at a time? As in probing only parallel to an axis, but I don't really know. I don't have any experience with one yet.

The CMM is a ESM manufacturing 24x20 with a Renishaw TP1s probe, and a heidenhain ND1403 quadra check readout/computer.

Can anyone elaborate on this CMM's limitations?
 
IME CMMs are much better at roundness, ie form, than they are at diameter once you start getting under ±0.0004". Most of this is due to forms of probe wear that don't calibrate out IME. If you stay on top of your probes and master to a good ring gage you can keep that minimized. This assumes the machine is otherwise capable/repeatable, which is a separate question.
 
Most of this is due to forms of probe wear that don't calibrate out IME. If you stay on top of your probes and master to a good ring gage you can keep that minimized. This assumes the machine is otherwise capable/repeatable, which is a separate question.
Sounds like i at least need to get the machine up and running, then have my calibration guy come in and see if the machine is worth anything.

If it can measure form decently, i can average out a bunch of points. Surely that will be my best bet, no?

I like the idea of a DTI on a known diameter as guythatbrews said, but this method assumes i can postion the known diameter concentric and parallel to the spindle axis, i think?
 
Somewhere in a reference I saw a method described to use three rolls and two gauge blocks to measure an unknown radius. I can't seem to find it at the moment, but if I have three more cups of coffee I might be able to work out the math.
 
I like the idea of a DTI on a known diameter as guythatbrews said, but this method assumes i can postion the known diameter concentric and parallel to the spindle axis, i think?
You can just superglue your "known diameter setting ring" on anything that is perpendicular to your spindle. Sweep the ID, zero dti, and it's set to known diameter. Doesn't have to be a ring, can be gage block with 2 outside blocks standing proud so you can indicate known inside dimension. Hold gageblocks with a magnet.

Really then all you need to do is sweep the bottom of fixture radius on previously found centerline, add the dti radius to distance from center. Et voila, this sum is the radius. You can also use the dti to check if split line is on center or not.
 
Standard chord calculation math but nice suggestion for measuring with gauge blocks on the surface plate.

TG, Thanks for those suggestions. I think the last one will be the easiest to implement.

If you are ever in Enid, Stop by. I'll treat you to lunch
Really then all you need to do is sweep the bottom of fixture radius on previously found centerline, add the dti radius to distance from center. Et voila, this sum is the radius. You can also use the dti to check if split line is on center or not.

I'll see if i can rig up something along those lines. I want to try everything and see what works best.

Thanks everyone. This is what PM is about
 
So many of these processes seem to eat up a bunch of tolerance range before you get into actual errors that I can see why probing is the default.
Would it be a bizarre thought to turn a diameter you wish to check to a very high level of confidence in true size and then probe that part that you know is correct, see what your repeatability is in probing that part?
IOW if you turned a part that was 3.0000" and when you probed it in the machine it always told you it was 2.9997, when you measured fixtures in that range you could be pretty certain that it is going to tell you a dimension 3 tenths under
My thinking being that round parts are easy to turn accurately, easy to measure accurately.
IF the probing was repeatable within tolerances that work, then accuracy becomes less important.
One could make a set of standards in a few hours out of soft material that ought to work for a number of diameters, maybe
If made of a similar material [IE 'steel or 'aluminum'] as the fixtures and kept in the same environment, then when you want to measure your 4.3457 diameter fixture you use your 4 inch standard in the machine and see that 'today; it is showing your 4 inch standard to be 4.0002 where in December it would show it as 4.0000
 
So many of these processes seem to eat up a bunch of tolerance range before you get into actual errors that I can see why probing is the default.

This is my feeling as well. Although there are some ways to measure shown here, I feel like there are just too many chances for error to get a reliable reading.

Maybe I'm asking too much of my measuring equipment?
 
I'm always an advocate for learning and understanding how to take accurate, repeatable measurements. There are always some cases where gauge blocks and micrometers just can't get a reading for practical, geometric reasons, but I've seen people fall back on CMM's and think they've got an accurate number just because the readout goes to tenths and beyond.

Some cases in point. I had a complex molding (large circuit breaker case) that first piece QC was failing for holes mispositioned. I couldn't find a design error and the toolmakers couldn't find a construction error. Everything should have been right, but QC said it wasn't so what could we do? That was go back to QC and ask to have them show how they were measuring. That turned out to be a tapered probe to drop in a molded hole. Only problem was that this was the shut-off end of the hole and the molding department uses a rat tail file to remove the flash. The probe (this was back in the old days) could not actually establish hole center no matter how many digits the readout had.

And on the other end, and I'm sure I've told this story before. We used telescoping gauges to measure a jig ground hole to tenths. One had to understand the correct procedure, to follow it rigorously, and be sensitive to micrometer feel to pull it off.

Even using gauge blocks to size a ground groove. There's a difference between "it will fit" and "it will only fit with effort", and "it's not going to fit at all."
 








 
Back
Top