Collision Detection is not activated on V2015.1

On this topic, do you have issues with it not functioning correctly? When I run mine, (2024), it gives me a long list of collisions that don't exist. For example, I get it on a bolt hole pattern that clearly has functioning clearplanes activated before and after each hole. In my experience, running it and trying to decipher the collisions being reported, is a waste of time. It's quicker and more effective to just prove the program out. I haven't had an opportunity to try it in 2025 yet.

On this topic, do you have issues with it not functioning correctly? When I run mine, (2024), it gives me a long list of collisions that don't exist. For example, I get it on a bolt hole pattern that clearly has functioning clearplanes activated before and after each hole. In my experience, running it and trying to decipher the collisions being reported, is a waste of time. It's quicker and more effective to just prove the program out. I haven't had an opportunity to try it in 2025 yet.

I am still new with this function. But, I guess it is all about tolerance( Clearance Plane?), for example: 2mm apart, 3mm apart. If the potential distance is closer than the setting, it gives out warnings.

I wonder whether we can setup this distance ?

I found this from YouTube about Collision Detection:

https://www.youtube.com/watch?v=2Dp7bqm85M8

And try to get same appearance he had   

Todd Noname  I think this is a yes AND no situation. Granted I am only running 2020 here but what you say is still true, sometimes it does have issue with holes (only the Demon knows why) but that does not stop it from from revealing boneheaded mistakes or things that clearly wont work which could really trip you up during a first run. Like missing a minus on a move point, or forgetting to retract in Z before a probe rotation.
We program exclusively offline here so its a very handy tool for us. 
Also it seems if you F9 the holes, move the rotation a teeny bit then it will stop identifying them as a crash. 

When you say to F9 the holes and move the rotation a tiny bit, do you hit OK after doing this or Cancel? I was wondering if it was PC-DMIS just not understanding what was going on, and somehow doing this trick makes it act properly, whether hitting OK or Cancel

I hit OK, but I am willing to give Cancel a shot and see if that also fixes it! 

Also, here at least since we run OLD programs that are just updated over and over, sometimes its due to some joker keying in dimensions which make the hits go sub-surface to the model by a small amount. 

Have you experienced the dreaded CAD vectors being flipped? I've only experienced this maybe once or twice ever, but it has happened causing this kind of issue on our parts with collision detection being funky

Not really since I learned that IGES automatically fixes all the surface vectors on import. I have noticed occasionally that some surfaces let you "click through them" not that the point vector is wrong, but that they seem to have their cloaking fields on until you rotate your view slightly. 

I have experienced a few models that were broken in some way no matter what you did or how you imported them. Twice from the same customer with weird "tears" in the surfaces (that was super not fun) 

Thank you for your sharing.

I tested Collision Detection (CD) this morning and found that most of the warnings in our current programs are related to circles and single-hit features( Edge type). I will use F9 on those circles to fix them.

Another issue I noticed involves probe movement between CIR12 and CIR6 (see Pic 1), specifically in the area circled in Pic 2. Visually, there seems to be plenty of buffer room, and the movement looks safe.

However, I wonder whether there be any potential issues when the probe moves in and out from the surface over a certain distance, similar to  clearance tube box?

Pic 1

Pic2

Look down the holes and then click the relevant feature in the collision list, it should highlight the actual move that is having an issue. Sometimes it has an issue with the fist or last hit before it retracts and moves to another location. 

Also click the button that stops the animation on collision and it will freeze the probe where its hitting. 

There are certain instances where CD kinda acts wonky and threaded features is one of them. Micro differences in the CAD also don't work in our favor. One of the biggest frustrations with CD is that sometimes I select a bore with some SAMPLE HITS and don't realize the software hasn't compensated for a chamfer so the SAMPLE HITS are slightly below the surface and CD picks up on a collision. 

Coll Detection isn't a flawless tool, but it would certainly help programmers getting a visual reference on how their program is going to execute. 

Thanks. I did have a warning  with a nut, even I adjusted and rotated hits, it was still red...