Linear artifacts

[[ro...]]

I have a lab in GE where I do their calibration, recently I have also been running their Accura Zeiss CMM(new facility, new machine). They found a document that says I need to do Linear artifact checks and volumetric checks every quarter. It appears I have to cover the whole measurement area. This is a big machine...I think it would have to be at least 4 feet long and 4 feet wide measuring area. They talked about having Zeiss coming in and do it every quarter....but its very expensive. They looked at a couple bars<---?? and artifacts and they were extremely expensive! I told them that I thought they could make something and have it measured at an outside lab and that would suffice. They aren't sure if this is proper. Anyone else have a big machine that they have to do this same check with? They made it clear that I cant use a small part, I have to use one as big as the biggest part I measure, which is 48" Diameter. Any help would sure be appreciated.

[[Me...]]

I used to program for a 6 post gantry and back then you could do multiple ball bars that covered the machines volume. The ball bar was only 36" ball to ball, but buy moving it through out the volume, +/-Y, +/-X, and +/-Z you could calculate the volumetric accuracy of the machine as a whole. I can't see any reason why you couldn't do the same for linears. Look for B89 volumetric performance on the internet and you should find all the data that you need.

Bill

[[De...]]

According to ISO 10360, you need to cover a minimum of 66% of the length of each axis as well as 66% of the length of the diagonals (spatials).

To accomplish this when Zeiss does a calibration they use a shorter test length and run multiple runs. For instance if the Y axis is 1800 mm they might run a 700 mm step gauge in two positions along the Y axis, thus getting results along 1400 mm of the 1800 mm axis of the machine and thus satisfying the requirement to measure a minimum of 66% of the axis.

This is a compromise based on the extreme cost of purchasing and maintaining calibrations on longer length standards.

I have been working on implementing a system to re-verify our machines throughout the year between the yearly Calibrations we have Zeiss coming in to do, and I can attest that it is not possible to do correctly without laying out a pretty penny.

We will use a 700mm step gage, in our case we manufactured it, however they can be purchased but expect to spend at least 10k dollars and likely more for one. You then have calibration of the step gauge which is also going to be quite expensive since you need a very good calibration (which should be done yearly). You will also need some kind of stand to hold the step gauge in multiple orientations which must be completely stable in order for the results to be worth anything.

In order to get setup I would say you need to budget $25-30K or get really creative in manufacturing your own step gages / standards and associated stands / fixtures to hold them. If it is only for one machine, you might be able to get away easier since you don't need to maintain flexibility so much, in our case we need to be able to handle multiple different machine sizes.

Assuming you get all the equipment you need and have everything calibrated, you then will need to create something to plot results based on whatever level of accuracy you plan on checking the machine to. In our case we are simply going to verify the machine to the limits stated in the manufacturers specs for the machines.

I think the initial setup cost can be recovered within a few years and the added peace of mind of knowing the machine can be checked at any time throughout the year with a bit of work is nice to have. After all, no one wants to deal with figuring out what to do when they find out their machine has been out of spec for some unknown amount of time since its last calibration.

All in all though this is a pretty in depth topic that is not easy nor cheap to manage which is why in most cases people simply have their yearly calibrations done by Zeiss and call it a day.

[[Ow...]]

Yeah, them certified length gauges are "new car" expensive and hard to justify if you don't drive it everyday.

Quality standards/auditors differ but, maybe you could get away with keeping one of your largest parts in a controlled environment and also had all of the characteristics verified/certified externally on an annual basis.

[[Da...]]

Do you have adequate temp control and a temp probe on the machine?

[[ro...]]

Yes, the CMM lab is controlled

[[Da...]]

In my opinion, I would push for a certified master. ( ball bar or whatever ) The cost of having Zeiss in every Quarter, I would think the artifacts would pay for themselves in a few years. Plan B would be to relax the requirement to twice a year...

[[Ja...]]

For this level of accuracy I would not suggest using a manufactured part or making your own standard. Metals will move over time – by that I mean they will attempt to stress relieve themselves and change shape/size, especially the larger you go. In most cases we’re only talking millionths or maybe a ½ tenth, but when your CMM volumetric accuracy is guaranteed to millionths or a tenth or two, this is a big deal. That’s why artifacts that are purchased cost so much – They are usually thermally or mechanically stress relieved a few times, and then allowed to sit and stabilize for very long periods of time – usually several months or even years. Making your own artifacts or using an already existing part is ok if you’re aware and willing to take some risk, but I would make sure management is aware of what can happen and get their blessing first.

[[De...]]

I don't in any way dispute that this is an important consideration, but I am fairly certain that using proper planning and manufacturing methods it is certainly possible to manufacture a high quality artifact / gauge so long as you go into the process with your eyes wide open.

In our case it was management that decided they would rather take the risk of making something than purchase one, they felt the cost savings was worth the risk given we have the equipment and the experience in this type of thing to justify making the attempt.

That said a lot of research went into the material and heat treatment process prior to making the step gauge.

Also, in order to verify that the gauge is dimensionally stable, we are having it calibrated at NIST, and for the first two years it will be on a more aggressive calibration cycle to monitor for dimension stability over time.

Due to scheduling and other non related issues, when the gauge goes to for calibration it will have been in its completed state for about 18 months, although to be fair that was not so much planned as it was a result of the amount of free time available to put into this project.

Either way, it will either be stable or it won't, only time will tell. I personally am confident, but I concede that there will be no 'proof' our process worked until we reach our second or third calibration.

From my perspective, whether this works or not, this has been a very interesting project to work on in all aspects.

[[ro...]]

It appears they now are focused on getting a Quik Chek gage by Glastonbury. Not sure if this meets all their needs but it seems to be on todays hot email topic. 😃

[[Ow...]]

I used a 16" quikcheck at the last place I worked.
It did what we wanted it to do, which mostly was just to check the basics after a crash.
The accura 1 cmm it was used on mostly had actually been knocked out of square with a crash into a probe bay, believe it or not, and the standard probe qualifications wouldn't detect it.
Running 3 shifts 24/7, you have no idea what can happen. After acquiring the quickcheck, the CMM checked it good for 3 years that I was there.

Keep in mind that the gauge doesn't come in perfectly square or perfect to 16" between holes, it comes in to what they certify it to be. Had a CNC machine technician come in and borrow it to square up a machine and it didn't work out the way he wanted it to
_
.

[[Cl...]]

Owen, how exactly do you use position this? Do you lay it
square to each axis and measure distance from hole to hole?

If so, in the middle of the surface plate, or in each corner?
Management has been talking about doing something similar

[[To...]]

Please sign in to view this quote.

Not speaking for Owen, I would say "All of the above", but you forgot mounting it at various angles, sorta like when Zeiss mounts the step block in a 3 axis direction to measure spatials. 😃 . If I were to guess, I would write a single program, using 1 probe and then change the probe to match the orientation that I was going to mount the block. You would then execute a manual alignment and you're off an running.

[[Ow...]]

It's been a while but, we didn't get real technical with it.

We checked it laying down referencing one plane as spatial to the other, linear lengths between holes and so on and then stand it up to check Z axis specifically for squareness to Y and X.

Protocol was to check it after any crash or anytime results from the CMM were being questioned.

Not by any means meant to be conclusive but, enough to have confidence the CMM was checking it the same as it did prior to being suspected as erroneous .

We were not held to any kind of strict ISO quality standard and didn't send it out for annual calibration but, we would check it directly after having the CMM calibrated by Zeiss on an annual basis.

[[Cl...]]

Thanks for the info.

[[Ke...]]

A little late to the conversation:

I worked in an ISO-17025 accredited lab (for what thats worth), and they used a "home built" artifact that was 16"x16", with a tooling ball in each corner, and a centrally located tooling ball that was elevated about 12" form the other four. It was never calibrated by another facility, as it was used on a bi-monthly basis and we only tracked the trend of each sphere relative to another...

My questions (as we are looking at doung the same design(ish) where I a am currently employed:
- for a 16"x16" base plate, is there a recommended minimum thickness of the material (sag/twist?)
- to minimize weight, we are considering using brass vs steel (not sure I can justify aluminum regardless of our labs temperature stability.

Thoughts?

[[De...]]

I don't think your going to be doing yourself any favors using brass, it has a coefficient of thermal expansion of 18-19 µm/m/°C which is really not all that much less than aluminum.

If you want to use this for tracking the machine over time you need to know the artifact itself is stable, I am not sure how stable brass is.

Due to weight, I think you would be better served to avoid a solid slab of any material and mill something with a T shaped web around the outside and perhaps a supporting lattice of some kind in the center if necessary. The T shape will help stiffen it while allowing you to remove a lot of weight in material.

My suggestion for a material would be 440c stainless hardened to AMS 2579/5D. It should give you something that is very durable at about 60 Rockwell C and given it is a cryogenic heat treat process you should end up with something with good dimensional stablity over time.

Unless the artifact is pretty big then you should be able to mill it within 0.03 on a side prior to heat treat and then do final machining post heat treat prior to finish grinding. Just make sure the machinist is aware of the hardness as its not a picnic to machine.