What's new
What's new

Measuring roundness/rotary table repeatability

jccaclimber

Stainless
Joined
Nov 22, 2015
Location
San Francisco
I have a part the size and shape of a margarine tub. I need to measure the roundness at a few different heights on both the ID and OD. Some of the parts are transparent and others are very shiny so my usual laser ranging device isn't getting a good reading. I do have a Tesa gauge head that is consistent around 0.1 to 0.2 microns, assuming very good fixturing. I suspect I'm looking for form error in the range of 1 to 3 microns, so I started by taking a practice measurement on a class X ring gauge. What I found is that my rotary table (small, normally used for rotating optics) has >20 microns of X/Y position error.
I mounted the ring gauge, rotate the table a full revolution, and my measurement differs by 20 microns. Rotate again and it's anywhere in between. The plan was to measure the gauge in a few different orientations to back out error in the table, master ring, etc., but I'm not there yet.

This is something I'm going to be doing many times for a month or two, then infrequently after that. A manual method is ok as I do want to do it in house vs. send to an inspection house with a CMM if at all possible. If I end up needing to send it out I will, but I've had issues with external CMM work not being accurate to this level in the past.

Rotation isn't critical, even 1/4 degree error would probably be ok, but staying centered is important. Would a standard rotary table work for this? If not, does anyone have recommendations on how this would normally be done? I have a good X/Y stage, so getting a part centered to 1 micron for measuring is easy, but spinning it seems to be causing heartburn.

Every place I've ever worked has had a rotary table for mounting on a Bridgeport.....except this one, so I can't just clamp my ring gauge to that and try that out.
 
If you are looking for single-digit micron measuring accuracy, you are in air-bearing rotary table territory. I don't think a "standard" rotary table made for machining will give you this level of sensitivity. Of course, if you manage to get your hands on a Moore or SIP rotary table, you may be closer to your target. The air-bearing tables made by (for instance) Pneumo Precision (or whoever they are now...) are accurate to about 0.2 micron, and are calibrated with spheres that are 0.05 micron accuracy.

Talyrond is a Taylor-Hobson brand name for a roundness measuring device employing an air-bearing. There are others as well.
 
If you are looking for single-digit micron measuring accuracy, you are in air-bearing rotary table territory. I don't think a "standard" rotary table made for machining will give you this level of sensitivity. Of course, if you manage to get your hands on a Moore or SIP rotary table, you may be closer to your target. The air-bearing tables made by (for instance) Pneumo Precision (or whoever they are now...) are accurate to about 0.2 micron, and are calibrated with spheres that are 0.05 micron accuracy.

Talyrond is a Taylor-Hobson brand name for a roundness measuring device employing an air-bearing. There are others as well.

I was wondering about a SIP. They seem to be plentiful and cheap online, with the minor inconvenience of often being 500 to 1000 pound objects 2k miles away (a solvable problem if needed). I'd forgotten about Moore, so I'll look at that as well, thank you.

Do you think that characterizing error against a ring gauge would work, or do you think the repeatability will also be insufficient?
 
Take a look at this chart, and you'll need to consider all the sources of error involved, including the chart tolerances shown for the ring gages:

Ring gage and Master Disc Tolerances per B89.1.5 | Engineers Edge | www.engineersedge.com

If you are in the 4" diameter neighborhood, the diameter tolerance shown is already at .6-.8 micron, depending on which side of 4.5" you are on. I believe this tolerance includes roundness error within the diametral specification. This is for the XXX grade (the most accurate), and will likely cost you some big bucks.

Regarding repeatability, this will really be (also) an amalgam of all the errors present. If the expectation is to have R&R down in sub-micron range, to assure an unimpeachable result in low-single-digit micron range, you have a serious task in front of you.
 
Last edited:
Do you need in-process inspection or can samples be sent to gauge house with a roundness gauge?
A bearing alone down under 1 micron is very expensive let alone the rest of the assembly and if you want to check 1 micron you need a 1/10 micron axis.
Mits, Taylor-Hobson, Adcole and others are sources for gauges in this range but you had better be sitting down when you open the quote.
Bob
 
I've had good experiences with Adcole and Taylor Hobson in the past, but this task won't have the volume to justify either. On that note, I did discover that old analog Talyrond machines are cheap on Ebay. For what I want to do with it I'd rather pay for tabular results than deal with transcribing from a paper wheel, but if I only needed to audit roundness it would be a good solution.

It's a check that I think I'll do a couple times a month right now, and less as time goes o. I'm open to sending out samples. My concern is that none of the external inspection companies I've worked with in the past have particularly impressed me. Typically either they think their CMM is perfect, or refuse to guarantee anything to the degree needed (was asking for 2-3 micron accuracy on a 100 mm part at a past employer). As soon as I add an MMB callout they want me to buy them a hard gauge, pretend the callouts don't exist, or calculate them wrong. In the end it takes more time and money than just doing it myself. If I can find a good one I'll go that route.

Rant over, we were able to find a place within driving distance that has a newer Talyrond. They use it mainly for internal work, but they're open to metrology as a service so long as I don't mind waiting for gaps in production, which is ok.

I still think this can be solved by mapping spindle error in a controlled environment, but if the guys down the road can solve our problem we'll hire it done to those with the expertise.
 
Long term follow up. The answer came down to finding someone else with a Talyrond and sending them parts. Of the places we've tried, direct to TH has turned out best.
 
question on talyrond: how can a gage have a resolution of 0.3 nano meters? is that somehow electron microscopy based?
 
question on talyrond: how can a gage have a resolution of 0.3 nano meters? is that somehow electron microscopy based?

No idea, but I've seen it before. The bolt on DRO on the diamond lathe at a nearby shop reads out in angstroms, so 0.1 nm. It's not 100% stable as you might guess, but you can read the last digit when the air platform is on and everything is just right. I have no idea what the overall accuracy is, but it's better than I can make, even in a temperature/humidity/phase of the moon controlled environment.
 
I have a talyrond dating from 1978.(Bought for 50 notes.)

The original purchaser, who I tracked down, said it cost around £18000 then and was calibrated to millionths of an inch.

The gauge head seems to be a type of LVDT but with a different type of geometry.

I connected a DVM reading in 0.1 mV to the output of the all analogue computer that makes up the system and was able to resolve to a 1/200 of a millionth, say 0.1nm. You could measure flex in a milling machine when pressed with a feather.
 
I will second CMMs are not accurate enough to under 5 micron. One of the Zeiss is, but it has to be in a controlled atmosphere. Its also about 1 mil. Also, for form testers Mahr are excellent. Accuracy in the Nanos.
I do measure parts with 1-3 microns on one of the CMMs. More to see the form of a diameter. I do get some good results, but not to the extent to use the results for nothing more than reference. I notice a bit of a bias in the measurements. Also, alignment is very critical. If you are measureing with the axis basically tilted, then you'd be measuring as an ellipse.
 
I believe the Mahr MFU100s are rated to .350 and .250 microns for accuracy. Some of the parts I deal with have a 1 micron total tolerance on some features. The Ziess Zenos CMM is rated less than a Micron.
 








 
Back
Top