Looking for new portable CMM

Hi all,
I'm looking for a portable CMM and found some Romers but not sure which one is best fit my need. I need a touch probe CMM, (PC-DMIS software preferred) to measure the mold/die feature mounted on the machine. Accuracy about 0.0002", arm is able to reach 40". I already submit a request to Romer but still want to ask your opinion on any pros and cons of the portable CMM? What machine type you have or suggest, etc...
Currently we have to disassemble the mold/die to put on the CMM to meaaure so I'd liek to avoid that.
Thank you in advance.
Nguyen

My experience with CMM arms is limited, but based on it and what I have heard from numerous people with much more experience, in the 40" range you will be lucky if can reliably report +/- .010".

When you start talking about the range of .0002" environment becomes extremely important even with the best stationary CMMs. There are machines that can do that, but they are in carefully constructed rooms that cost about as much as the CMM itself.

So, do you have a cool million $ to toss at this?

My experience with CMM arms is limited, but based on it and what I have heard from numerous people with much more experience, in the 40" range you will be lucky if can reliably report +/- .010".

When you start talking about the range of .0002" environment becomes extremely important even with the best stationary CMMs. There are machines that can do that, but they are in carefully constructed rooms that cost about as much as the CMM itself.

So, do you have a cool million $ to toss at this?

I'm not sure what equipment your PCMM experience was with but arms are certified to standards just like stationary CMMs. The volumetric accuracy of a 1.2 meter Infinite 2.0 is +- .0006”, so obviously not better than .0002”, but it is that accurate when evaluated by the ASME B89.4.22.

That may be, but I am not intersted in B89.4.22 results when in real world applications +/-.010 is what you get.

We had a salesperson bring an arm out for a demo. I presented a part that simplified was a block with a tab sticking out one side. There was a hole in the tab. The Datums were the pretty obvious 3 mutually perp planes. I wanted to know how well the arm could report the location of the hole and the thickness of the tab. Both had been measured several times by myself and 3 others using CMM, microhite, a micrometers (for the thickness.). I asked for thickness first, they could not get closer than .008 to the actual thickness and their repeatability was equally horrid. I did not even bother asking for position after that.

That is the extent of my personal arm experience, but I have conversed with many who use arms and CMMs for a variety of applications.

Arms are great for some things. They can do what they can do. Tight tolerances they can not do.

Oh, and I have not even gotten to the numerous reports of frequent encoder failure & costly replacement (In terms of both $ and downtime).

Lastly in the interest of full disclosure, do you work for an arm company or a division of a company that makes sells arms or arm software?

That may be, but I am not intersted in B89.4.22 results when in real world applications +/-.010 is what you get.

We had a salesperson bring an arm out for a demo. I presented a part that simplified was a block with a tab sticking out one side. There was a hole in the tab. The Datums were the pretty obvious 3 mutually perp planes. I wanted to know how well the arm could report the location of the hole and the thickness of the tab. Both had been measured several times by myself and 3 others using CMM, microhite, a micrometers (for the thickness.). I asked for thickness first, they could not get closer than .008 to the actual thickness and their repeatability was equally horrid. I did not even bother asking for position after that.

That is the extent of my personal arm experience, but I have conversed with many who use arms and CMMs for a variety of applications.

Arms are great for some things. They can do what they can do. Tight tolerances they can not do.

Oh, and I have not even gotten to the numerous reports of frequent encoder failure & costly replacement (In terms of both $ and downtime).

Lastly in the interest of full disclosure, do you work for an arm company or a division of a company that makes sells arms or arm software?

Look, I have been reading your posts on this forum for years and I know you generally know what you write about but this time I disagree.

Discussions like this are exactly why we have standards in this business. If the equipment is sold with a specification and no longer meets said specification, its not only “not accurate enough”, its out of spec. The standards are designed so that the give end users the ability to prove that their equipment works, not some arbitrary engineers guess at what is important.

You list 0 experience actually measuring anything yourself with an arm. I don’t care what some sales guy at a demo did or what other people did that you heard about. If I don’t measure it myself, with equipment I know meets its spec, I’m skeptical and I absolutely would not advise someone on a forum without that experience.

And yes, I do work for a large metrology company that makes equipment such as stationary automated CMMs, arms and trackers that are all certified to standards such as ASME B89.4.22. I have used them all in “real world applications” and I know what they CAN do, not just from measuring things but because of the standards.

lol, I love this guy.

Wes as someone who worked with DCC CMM's for about 8 years before messing around with PCMM's, here's my take on it. You are entirely operator dependant using a PCMM. Smaller probe shafts flex, and larger ones can't get into certain areas. The "cone" style calibration that is generally provided for tip calibration leaves much to be desired. In the world of PCMM's it is all about technique.

Granted you can't hit .0002" accuracy with an Infinite 2.0, but I do believe you can hit the +/-.0006" stated above, but precautions need to be taken. If you need tighter accuracy, you have to stick with the larger probes (8mm to 15mm range) and they can't stick out too far. Calibration needs to be done to a Calibration Sphere, not a single point cone check (this is 20 hits with the probe sitting inside a cone, rotating the wrist around in different directions. here the stylus nominal diameter is used for calculations, and if the styli isn't exactly the number used you have problems). After this, holes need good sample hits, single points need good sample hits, the operator needs to be careful. Technique and Setup are also important at this point. Whether it's on a tripod or using the magnetic base, it needs to properly secured, and the relationship to the item being measured needs to held. At one facility, we have a granite surface plate with a 1/2" plate of steel bolted down to it on one end. The Arm is attached via the magnetic base to that plate, and anything being measured is clamped to the granite surface so that the relationship between arm and part never changes. That same facility uses a 30mm Calibration Sphere. At that facility, probe calibrations are within .015mm(.0006") all the time (edit: this includes a 3mmx20mm probe, which is the one that gets close to the .015mm), and that arm is a 3000 series (old). Another facility uses a 5000 series arm (newer but not a 2.0). They have a steel table instead of the granite surface plate, and use the cone method for calibrations. Their probe calibrations range between .040 and .065mm. The steel table is about 7'x7', braced with tubing around the outside and a cross throught the middle. If you attached the Romer arm in the open areas between the bracing, then you are looking at the .010" accuracy that you speak of. At the facility where things are done the right way, I would comfortably say that +/-.0006" could be done with a smaller Infinite 2.0 Arm (if we had one), operator dependant of course.

I can only say from my own experience I've tested it out on a part that was roughly 25" long in the features I checked. I don't want to start the old Gage R&R on a CMM discussion, but I reran the same part 15 times without moving it and my total variation in any given measurement was .043mm. This was on the arm I spoke of that is not (IMHO) properly utilized. Obviously I avoid the springy part of the steel table, but my probe calibrated at .040mm, and I was using a 6mm probe (which as I stated above I don't like for the flex). I had to use the 6mm for the part I had, and I had to do it quickly because it was an unsanctioned test that I needed to perform so I knew what I was getting off that setup. I had known values on the part I checked from my Cordax RS-150, and those values were within the range of the .043mm (not necessarily the average of the range). Lastly, I tried to measure every hole with the probe coming in at the same angle each time and the same spots in the holes. That's the best I can tell you.
I am going back up to that facility in a couple days. I'm bringing a tooling ball with me that I've measured on my Cordax. I want to see if I can improve my results using that.

The accuracy of that 5000 series can definately be better, and I gotta think that with a better arm and a better calibration method I could bring the range of measurements to within a +/-.0006" window. Please noone berate me on my methods here, I was in a hurry and trying to work under the radar. I don't have a lot of support in my strive for excellence at this place.