PCDMIS Software/Process Validation

[c_bearden]

I have fought it tooth and nail but the Quality Manager tells me I have to do Software/Process Validation for PCDMIS. I was under the impression that I could take PCDMIS out of the box and start using it. Also, the annual Calibrations from B&S should tell us that the CMM is doing what its suppose to. Anyway, he not only wants PCDMIS validated but he also wants all of my individual CMM programs validated......kinda like I'm writing a software program instead of a CMM program......bottom line, its just another knee jerk reaction to getting certified to ISO:13485 for medical devices
In my mind, I just want to know that PCDMIS is giving me the true results from my measurements........Plus, I want to keep it Simple (K.I.S.S)
So, here is what I want to do.......Take several known standards (Class 1 GageBlocks) and measure them. If the Software tells me that a 1" gageblock checks 1.000005", then I know its giving me correct results.......Software VALIDATED!!!!!!...........
Do you think this would be acceptable to an auditor?

Thanks,
ZeissUser

[cmmguy]

Ummmm nope. Medical device companies require software validation. This will require a certification from Wilcox/Hexagon stating that the software is certified. There is a standardized data set that is used to validate the software. Known numbers in, expected results out. Had to do it with other equipment that I sold in the past.

As far as your individual programs go, you may get away with comparing the results with another method and then "locking" that version of the program. Each time a change to the program is made, re-validation is required. BTW, writing a part program is exactly like writing any other kind of program. Anytime you have control over the outcome - it qualifies as a program to be certified/validated.

This will be difficult to impossible with PCDMIS because as I understand, everytime you run a PCDMIS program it is loaded and changed as you measure a part. This technically should not pass the certified program test.

Talk to your Sale Rep. Ask for a letter from Hexagon stating that the software is validated. I dont know what the solution is regarding the part programs - maybe someone else has gone through this process with PCDMIS.

PTB Certification of PCDMIS 3.7

[c_bearden]

Very good info.....Thanks for that......

As for validating our individual CMM programs, the QA manager said something about doing a correlation study between the (5) CMMs and using that as our validation.......would you consider this acceptable?..........that just don't sound like validation to me......

Doberman

[cmmguy]

If you are using the same cmm program, then I would say no. If you used a customer program, then Yes. I would correllate the results against traditional methods such as a surface plate check

Here is FDA site

[Andrew Gardner]

At a previous shop, we had to validate for Boeing. What we did, was took the B & S demo block, make a few changes, and had it inspected by a "calibration" company. Then, we took those nominals and programmed the block to them using at least one kind of every feature. Every time we uprgraded versions (LOL) or at least once a year, we ran the program over the block to make sure everything was still doing what it was supposed to. IMO, this is BS - I've never had PC-DMIS not translate or check something wrong. Aside from the software bugs, and operator errors (not me of course), it has been pretty reliable.

[Mark Farren]

We virtually had to do the same thing as Andrew for the same company. We created a artifact soild model, wrote a cnc program, machined the part, brought the model in to PCDEMON, created an inspection program, checked the artifact and got our results. Then we sent everything to a calibration house. They loaded our program and inspected the artifact on their machine, compared the results to ours and said we were good to go. This was pretty much all under the direction of a Boeing representative who was working directly with the cal house. I agree, seemed like a bunch of BS to me, but if ya want the work you gotta jump through the hoops! We get a lot of work in that all we get is a model. The drawing is just an iso view with a tolerance block identifying tolerances for different types of features and they(Boeing) wanted to assure we weren't losing anything in the translation process. So that was our cluster we had to go through.

HTH

[cmmguy]

I dont believe those methods will fly with Medical with regards to the validation of PCDMIS. It is a different beast. I think the origins of this requirement go to the custom software written for machines that make medical products, like automated medicine mixers, etc... It has carried over to other devices controlled with software. There is a cert for the Application software(PCDMIS) that you can get, but the validation of your part program should run along the same lines as described above.

[kbotta]

Whats an operator error?

[cmmguy]

a phone call to the wrong person.

[SuperCrash]

I have fought it tooth and nail but the Quality Manager tells me I have to do Software/Process Validation for PCDMIS. I was under the impression that I could take PCDMIS out of the box and start using it. Also, the annual Calibrations from B&S should tell us that the CMM is doing what its suppose to. Anyway, he not only wants PCDMIS validated but he also wants all of my individual CMM programs validated......kinda like I'm writing a software program instead of a CMM program......bottom line, its just another knee jerk reaction to getting certified to ISO:13485 for medical devices
In my mind, I just want to know that PCDMIS is giving me the true results from my measurements........Plus, I want to keep it Simple (K.I.S.S)
So, here is what I want to do.......Take several known standards (Class 1 GageBlocks) and measure them. If the Software tells me that a 1" gageblock checks 1.000005", then I know its giving me correct results.......Software VALIDATED!!!!!!...........
Do you think this would be acceptable to an auditor?

Thanks,
ZeissUser

We had to do something similar for our Statistical Process Control software; verifying that the new releases of the software met repeatability standards for given, measured data sets. It's a blast alright! I couldn't imagine having that apply to individual programs.... Good luck with that!

SC

Relevant stats:
CMM: B&S ONE
PC-DMIS 3.7 MR3
Probe Head: MH8
TP20 probe system
Probe Rack: MCR20
Extension(s): 50mm carbon fiber (PECF1)
Datum ball: straight fixed, 0,0,1
OS: MS Windows XP Pro Version 5.1.2600 SP2 Build 2600
Total Physical Memory: 512.50 MB
Graphics: Mobo integrated

[kingsld1]

I'm in the medical device field also.

What we did was a Gage R&R for the validation. We also included the B&S cert for PC-DMIS.

For each program we do a very short repeatability test and a peer review of the program.



DRK

[c_bearden]

Explain your peer review process.....Is this where someone verifys your CMM findings using an alternate method such as height gage/idicator/surfaceplate?

Thanks for everyones input
Doberman