Features incorrect in CAD after angle rotation

[Mike Ruff]

Hi all,

I have a strange one that happens intermittently. This issue is not program specific and usually does not happen, but it's super irritating when it shows up. If you take a look at my picture, you can see features way off of where they are supposed to be. This screenshot was taken after the program finshed running. Those features that are way off on the CAD were measured on the part in the correct location. Every once in a while, the demon decides to measure features correctly and then put them in the wrong spot in the CAD (this is a 4-sided pyramid and the other 3 sides were fine). This also translates to the report and this time it reported my position at 11 inches (yes, 11 inches) out of tolerance. I can rerun those features with Ctrl+E, redraw the report and everything is good. And it probably will run correctly the next time I run the same part. This happens randomly on pretty much any program and I'm not sure what's causing it. I have a lot running on my computer all the time (Solidworks, Outlook, PC-DMIS, Chrome, Adobe, Word, Excel, etc.), but my computer has 16GB of RAM and another 4GB in the video card for graphics (nVidia Quadro K2200). Has anyone seen this before or have any idea what's causing it?

One interesting note is that this only happens in programs with angle changes. However, if it happens in the same program when running a lot of parts back to back, it is never the same features, it is random features every time. Sometimes it will be 3/20 features in A90B90, other times it will be every feature in A90B-90. It's totally random where it happens.

I'm stumped, and finally irritated enough to ask around

Running CAD++ 2015.1 (Release) x64 build #791
features in wrong location.jpg

[BKulpa]

Are you by chance using moveall command for those wrist angles.
I've had a similar problem where is seems to get confused on the tip angle.

[Mike Ruff]

It happens with or without MOVE/ALL

[louisd]

Are you executing your routines from local drives? If you are executing from a network drive, this could be a direct cause.

[louisd]

test it out. put your routine that's wigging out on a local drive, see if it resolves it. if so, you have your answer, as crappy as it is.

[louisd]

We saw pc-dmis just straight-up SKIP lines of code, specifically file/writelines for report outputs.
I presume the local drive enables the machine to output & function without any delays or interference.
--When you are reading/writing to the network, you are throttled by the network bandwidth. Not to mention fluctuations caused by EMI/noise induced along the cables, a poor cable or connector conductivity, a junky network switch... There's so many more variables that can potentially slow down or interrupt the data transfer causing it to wig out when accessing the routine from a network drive.

[Mike Ruff]

Hmmm.... That makes sense. I can see it being caused by a slight communication lag with the network. As far as I can tell, it can happen on any routine, and I'd guess it only happens maybe 1 out 40-50 parts, so transferring programs to the local drive and running from there will be tedious to see if it goes away- this is the best idea I've heard so far to test it out, but it could potentially take a few weeks of running off the C drive before I'd be confident that that's the issue. I'll talk with our IT guy too and see if he has any input on the situation. Thanks for the input!