How do I calibrate Checking Fixtures?

[Allie]

We supply bent stainless steel tubing to the automotive industry. After we bend the tube to the required shape, it is placed in a checking fixure that verifies the correct length of the tube, the correct angle of the bends and the diameter of the tube. If it fits, good part, if it doesn't, bad part is basically how it works.
The checking fixture came with a certificate of measurement of some sort. Obviously it has to be on the calibration schedule, I just don't know how I should calibrate it. Do I send it back to the manufacturer to have the measurements reverified or is it possible to have it done locally by CMM? I can set up a daily visual check of an burrs, signs of damage, etc. But I am unsure what to do about a calibration of the actual dimensions of it.
If anyone out there has an opinion on this, I would love to hear it.

Allie

 

[Al Dyer]

As this sounds like a go/nogo gauge, calibtation is probably impractical.

Do you have studies on the raw material going into the bender to ensure that the correct length of tube is being introduced into the process?

Are you relying on an attribute gage to validate final product?

What kind of tolerances are we talking about?

Al...

 

[Allie]

The tubes are cut to length on a laser cutter which is accurate down to tenths of millimeters, I believe. The length is then verified before going into the tube bender. On the longer tubes (about three feet long) they are only verified in the checking fixture after bending . If a problem were encountered, the tubes could be validated in the Trace Master (a primative CMM)but I find the Trace Master is not a gauge with a realiable Gauge R&R. We don't use the Trace Master for any official measurements. The tolerences are usually something like 180mm +/- 2mm.

Al

 

[tomvehoski]

It sounds like you could calibrate it with a CMM if you feel it is necessary. If it came with a certificate of calibration, you may still be able to rely on that. You have to look at the potential for it to change and the severity of the issue if it does change. Any calibration lab with a CMM should be able to help you.

You might also be able to verify it by taking known good and bad parts (checked with some other measurement technique) to see if it will pass good ones and fail bad ones. It may be difficult to get parts that are marginal to use for a test. This method may not be as accurate, but may be good enough and more cost effective.

Tom

 

[Allie]

That sounds like a plan Tom. I can use the good part/bad part as in in house verification and send it out every so often (yearly?)to get checked by a CMM.

Allie

 

[tomvehoski]

How often is the fixture used? You may be able to go to 2-5 years if there is little chance of change or wear & tear. Don't waste money calibrating things that won't change.

Tom

 

[Allie]

It is used to check every 20th part in a process with runs roughly 2000 parts a day

Allie

 

[tomvehoski]

You will have to set up the frequency based on your judgement. Perhaps build a set of good and bad "master" parts. Monthly or quarterly put them in the fixture and make sure they are still being passed or failed. If everthing is fine, continue checking this way and increase the time between checks accordingly. If you start to see good ones being failed or bad ones passing, send it out to the CMM. If everything remains stable, save the cost and use your internal method.

 

[Allie]

I think that might be the best way to cover it off. I am not anticipating too many problems with these fixtures. They are built very well and solid.

Thanks for your time and thought

Allie

 

[Daryl Montie]

Just as a confirmation... I've been involved with a similar situation where a test fixture was used to check the positioning of several features. The method of verifying the fixture was done just as Tom presented -- a five year calibration (by CMM) with a quarterly check using a "master" go/nogo. It worked great and was inexpensive. Love the Avatar, Allie.... ACKK!!