During a recent audit, an inspector asked me how can do I "Calibrate" our CAD software. I told him that we didn't need to since it was commercial software. That it had already been verified by the manufacturer (Autodesk, Inc.). That seemed to please him and he moved on. When we got the final report, there was a recommendation that we develop a method to "Calibrate" our CAD software.
Rather than "calibrate", I think that we just need to verify that it is accurate. If we found some inaccuracies, then we could deal with that later.
One easy way that I could think was to draw a cube X x Y x Z and then verify that the volume is accurate. Of course it could be a 1in cube, 1ft cube or 100ft cube..... either way it should give the correct volume. I just don't know of any other way to prove to this dip-wit what I draw is an accurate length.
Has anyone else been asked to do this for Inventor?
How exactly would you calibrate a CAD package? That's completely rediculous. Did he ask you to calibrate your calculator? That's all Inventor is, a really high end graphing calculator.
Create a dimension that is equal to 2 inches + 2 inches and show him that the line comes out to be 4 inches long.
how silly IMO..
You aren't really measuring anything with Inventor. The measurements are done with phyical measurement devices based on dimensions provided by your software only. Frankly I would just stick a "calibration not required" sticker on the Inventor box and stuff it back in the closet.
Guess others have come across this.
http://asqasktheexperts.wordpress.com/2012/06/04/calibration-of-autocad-software/
@karthur1 wrote:During a recent audit, an inspector asked me how can do I "Calibrate" our CAD software.
I hold up my calipers to the screen.
Can you identify this person and organization?
take a look at this web page
http://asqasktheexperts.wordpress.com/2012/06/04/calibration-of-autocad-software/
I agree with everyone here that this request is TOTALLY rediculous. I think he is was just looking for something to put in his report actually. But, now that he has put this in there, we have to resolve it or we could lose our certification.
He didn't ask about the CAM software or Excel that we use. I didnt want to open that can and cause us more grief. I looked at the "Ask the Expert" page that a couple of you linked to, and that is basically what we told him.
JD, this is AISC auditor.
I understand what he was after (I think), how do you validate your processes.
But the terminology he used and if no other context of the question was presented I would liken this to a Dilbert cartoon ( http://dilbert.com/ ) - jargon with no meaning, trying to sound professional but instead sounding clueless. And to top it off it makes the responder seem to be the clueless one unless they instantly recognize the institutional jargon and respond appropriately.
In the follow-up when you do respond appropriately the reviewer will notate that as progress in response to his review. Nothing changed, just the perception.
Maybe the auditor was thinking about any analysis you do. You may have been on the right track with your initial mass properties idea. Do you perform any stress analysis on your designs? There are certifications for FEA codes. This is really CAE not CAD, but who knows. The guys over in the Simulation discussion group might be able to help. Here's a certification thread over there: http://forums.autodesk.com/t5/Autodesk-Simulation-Mechanical/Algor-Simulation-Accuracy-Benchmarks/m-...
I hope this helps,
FEA is NOT a measurement tool either and as such should NOT require calibration/validation.
No one in their right mind would rely 100% on the output of a FEA program. Its simply a "get in the ballpark" design tool that should NEVER replace real world performance measurments.
Export a model in a neutral format and verify that file in (something else)? But then, how do you verify / calibrate that (something else)? It's turtles all the way down.
Maybe ask Autodesk for some sort of certification, but I bet somewhere in the 1-point EULA text is something along the lines of 'don't bother asking, pal'.
Thankfully we have never had to deal with this "calibration" question.
Thinking about this I've realized we do have something that could be used to verify the accuracy of our software.
We make a variety of parts that are waterjet cut from steel plate. Typical we produce our part and drawing in Inventor and then save a copy of the Inventor drawing as an AutoCad drawing. The AutoCad drawing is imported into our vendors waterjetting software to be used as a template.
The fact that the finished parts can be physically measured and found to be accurate should verify that our software is accurate. Of course I guess you have to take into account the accuracy of the waterjet process, which is pretty hig but like any manufacturing process not 100%.
Finally, I agree with the others that have stated that this is a silly question.
Richard in Houston
CAD is all math. 2+2 will equal 4 no matter what means you use to calculate it. You can't measure math, you can't calibrate it, it just is what it is. Are you going to suggest that if you were to calculate 2+2 on a really mega-fancy pi calculating calculator that can go out to 2000 places, that 2+2 might actually come out slightly off from 4?
2+2 = 4 only for civilian work, when you work for the government, then it equals whatever the committee formed to evaluate it says it equals.
Can't find what you're looking for? Ask the community or share your knowledge.