Scanning results

[robehor4]

I read on a previous thread that the way PCDMIS calculates a profile is incorrect. I could not find anything confirming or denying this. Does any know of a thread I can review or have comments for this?

[BKulpa]

I use profile all the time and the results seem to be correct.

The only thing to watch is plus and minus deviations are: Material on / material off, not dimensionally posative and negative.

[AJHVW93]

I read on a previous thread that the way PCDMIS calculates a profile is incorrect. I could not find anything confirming or denying this. Does any know of a thread I can review or have comments for this?

You may be thinking of probe comp ? One of the aces can help you on this one.

[Matthew D. Hoedeman]

I read on a previous thread that the way PCDMIS calculates a profile is incorrect. I could not find anything confirming or denying this. Does any know of a thread I can review or have comments for this?

It depends on exactly what you mean:

IF you are scanning WITH cad data, then Pcdmis uses the cad nominals and cad vectors for all touches. The results will be reasonably close (on the order of microns) unless your form is really whacked (on the order of 15 degrees off or more)

IF you are scanning anything with a touch-trigger-probe (the only kind I have so the only kind I know for sure), then unless you are using SURFACE POINTS with a minimum of 3 sample hits per hit, then the comp for the probe radius will NOT be correct, it will only be along 2 axis, not 3. This is fine if you are scanning a 2-D profile. However, if you are scanning a 3-D surface, then the probe comp will not be correct. If you are scanning 'up/down' in the 'x-axis', you will get no probe comp in the 'y-axis' direction and you could easily be off a BUNCH. For this reason, when doing a NON-CAD scan of a 3-D surface for reverse engineering, COMP=OFF is your best bet.


IF you are talking about the PROFILE dimension itself, it will depnd on what version you have. It WILL be correct, but you may need help in understanding what it is telling you. In some versions it takes teh highest deviation and doubles that value (treating it like true-pos) and in some versions it takes the highest to lowest and reports that value. If you report each point as a point and not part of a scan or set and use the T axis, that will always tell you the correct value (use 1/2 of the profile tolerance for the +/- value ie profile tol of 1.0mm = +/-0.5mm).

[Underspec]

What I usually do is enable text analysis to prove it. This will list the point deviations.

[robehor4]

Thanks for your info. I wish I could find that thread again, since I use CAD for everything here and have compared results to a profrile tracer Im confident with my results. Since I didnt here any negative responses from you guys I will continue with my current process. Thanks again.

[Jan d.]

See this post:

http://www.pcdmisforum.com/showthrea...hlight=profile

Yes, I remain strongly opinianated against the way PC-DMIS reports profile using the 4.x XactMeasure profile dimension. It is dead WRONG. But Wilcox disagrees and that is OK because by using the MIN/MAX you can get to the correct answer. So as long as you know what the problem is, it is easy to get to the correct answer.


Jan.

[robehor4]

See this post:

http://www.pcdmisforum.com/showthrea...hlight=profile

Yes, I remain strongly opinianated against the way PC-DMIS reports profile using the 4.x XactMeasure profile dimension. It is dead WRONG. But Wilcox disagrees and that is OK because by using the MIN/MAX you can get to the correct answer. So as long as you know what the problem is, it is easy to get to the correct answer.



Jan.



Jan.

Thanks Jan, that is exactly what I was looking for. Too bad I missed all that.