Machine Simulation missing collisions?

[machineimpossible]

I got a model from our machine tool builder, and have it all setup and seemingly functioning properly in Machine Simulator. I went to test out the collision checking, so setup a drill to plunge 12" thru the part. When I go into the Simulator, the tool goes thru, then when the holder hits the stock, it detects a collision, but then when I continue, the drill gets deeper and the Z axis head goes right into the A axis tilt without showing any collision. 

When I run the exact same setup with one of the Generic MachSim 5 axis (trunnion style) it stops at the tool holder collision and machine component collision

I checked that my parts of the machine are "geometry" and not "machine housing" I'm stumped as to what is wrong?

[pro grammer]

We are also having MANY 5-axis problems with 2020. Swarf and curve ripping right through the part. Simulations shows a perfect path. It seems if we try to make a full loop everything goes haywire. So, I have to break up everything into sections. It is a massive headache. No support from our vendor. To make matters worse, when we try to use a new model 2020 refuses to let go of the old model. Paths have to be completely regenerated. STILL no support from our vendor. 

[Guest]

Just now, pro grammer said:

We are also having MANY 5-axis problems with 2020. Swarf and curve ripping right through the part. Simulations shows a perfect path. It seems if we try to make a full loop everything goes haywire. So, I have to break up everything into sections. It is a massive headache. No support from our vendor. To make matters worse, when we try to use a new model 2020 refuses to let go of the old model. Paths have to be completely regenerated. STILL no support from our vendor. 

If your vendor isn't replying you need to escalate to CNC software. Also I would call them if they are ignoring, go ring their doorbell and peek through the window!!!

[machineimpossible]

I'm going to try to call customer service directly in the next few days. 

[pro grammer]

Just now, byte me said:

If your vendor isn't replying you need to escalate to CNC software. Also I would call them if they are ignoring, go ring their doorbell and peek through the window!!!

WE DID! Have not heard anything after almost 2 weeks. Sent them a zip2fkinggo with one file that had a corrupt path. Never heard back from them. It was a file with about 40 processes (at that time) that would crash the simulator. When we tried posting it out it locked up with errors and a shtload of $$$$$$$$$$$$$$$$$$$$$$$'s. I was able to figure out the problem area after about an hour after we sent the file to MC and get that specific process working. I don't have time to babysit the fckedup children at MC. I told my boss before this 5x project started that MC could not handle it. They are probably too busy making new bugs and designing prettier icons. Will the new icons be pastel or earth tones? Can't wait for 2021 to see!!!!!!!!!

[JParis]

5 axis machsim is going to be mostly useless unless it is tied to the post....every move has 2 possible solutions and the output will be determined in the nc file....in the nci file you can and will at times see improper output...

2020 Machsim was broken for me.....it became utterly useless....

Jury is out on 2021 but I have begun the 3rd party CAV verification drumbeats....to me, as implemented in Mastercam, machsim is just a dog & pony show....if you want it done right, you have to go to a company that does just that.

JM2C YMMV

[machineimpossible]

well, in my case I'm only using it for 3+2, and my current issue really is just that it is not detecting collisions between machine components, which is all I'm really wanting it out of it right now.

[JParis]

You have to make sure your collision checking is set up properly

and even though it's "only 3+2" there are still multiple solutions.....components may not in machsim be in the actual positions that the machine will be in...

[pro grammer]

16 minutes ago, JParis said:

5 axis machsim is going to be mostly useless unless it is tied to the post....every move has 2 possible solutions and the output will be determined in the nc file....in the nci file you can and will at times see improper output...

2020 Machsim was broken for me.....it became utterly useless....

Jury is out on 2021 but I have begun the 3rd party CAV verification drumbeats....to me, as implemented in Mastercam, machsim is just a dog & pony show....if you want it done right, you have to go to a company that does just that.

JM2C YMMV

Let me know when the jury returns a verdict. Until then I will avoid 2021. The NCI file is where I discovered the $$$$$$$$$$$$$$$$'s. That is how I tracked down the faulty process. The funny thing is I recreated the entire process. When doing so, the new process auto loaded most the same settings from the problem process. That caused the new process to do the exact same thing. It was not until started unchecking depth cut boxes, regenerating and rechecking and regenerating them that the issue was resolved. 2020 has a very very bad issue of not letting go of things. Just like my wife. I spend a lot of time recreating ops because of corruption. I believe the real corruption is the lack of help we get after paying maintenance.

[Guest]

Just now, pro grammer said:

The NCI file

can you duplicate this, in a discloasable file?

[machineimpossible]

26 minutes ago, JParis said:

You have to make sure your collision checking is set up properly

 

and even though it's "only 3+2" there are still multiple solutions.....components may not in machsim be in the actual positions that the machine will be in...

Thank you, that was it!

So does it matter which components are in each box? I put the z axis head and tool in 1st box, and all the trunion components and fixture in the 2nd box?

[pro grammer]

4 minutes ago, byte me said:

can you duplicate this, in a discloasable file?

This is a 5ax aero prototype part. Unfortunately, no.

[JParis]

Just now, machineimpossible said:

Thank you, that was it!

So does it matter which components are in each box? I put the z axis head and tool in 1st box, and all the trunion components and fixture in the 2nd box?

Yes....you're choosing which components to check against for collision...

As mine is set up in the pic, my tool and spindle are checked for collisions against the rest of the XYCB and tool probe

[Guest]

Just now, pro grammer said:

I spend a lot of time recreating ops because of corruption.

pro grammer,

 if you copy the chains/drill points to a new operations instance, does it fix the problem?

If it does, then you may have external geometry being wrongly associated with your operation,

It happened to me before..

 

[machineimpossible]

well it seems to be working now on my little test. It registers a collision when the Z head hits the Rotary table. 

[Guest]

I have proof, foreign entities can be randomly bound to your operation for unknown reasons, this happens when the operation is created...

[machineimpossible]

Now my next question, as others have added, your machine simulation must be "tied to your post". How do you do this?

[JParis]

1 minute ago, machineimpossible said:

Now my next question, as others have added, your machine simulation must be "tied to your post". How do you do this?

$$$ to reseller

[machineimpossible]

 

3 minutes ago, JParis said:

$$$ to reseller

Got it. Now would that be included under the umbrella of Post support maintenance? because we are up-to-date on that for this year. 

[JParis]

1 minute ago, machineimpossible said:

 

Got it. Now would that be included under the umbrella of Post support maintenance? because we are up-to-date on that for this year. 

Nope, it is a seperate add-on purchase. Pricing can vary depending upon machine and requirements, so a direct ? to them is best to ascertain the actual cost

[pro grammer]

Just now, byte me said:

pro grammer,

 if you copy the chains/drill points to a new operations instance, does it fix the problem?

If it does, then you may have external geometry being wrongly associated with your operation,

It happened to me before..

 

That was my thought. Just select the new chains/surfaces. It would not let me. The 2 models were placed in the same exact positions on different levels. Then I turned off the original model layer, deleted all of the entities and selected the features of the new model and regenerated. Then to test I turned on the old model and translated it 5 inches. When I did that It showed that I had to regenerate the ops again. NOT COOL! The only way it would work was to completely recreate new ops and copy the parameters over. That had to be done for 17 out of 52 processes. The original model was a catia import. Not one straight surface on this thing. We were told the model had been translated from metric to inch. That is usually a bad idea. After that we had some many gouge issues (just really stupid stuff) we tried converting to STP and reassigning geometry. That reassociation did not go well at all. But some of the gouge issues were fixed, not all.

[Guest]

Just now, pro grammer said:

The only way it would work was to completely recreate new ops and copy the parameters over.

You can automate this process, that's what I did.

Which operation types do you use? i.e.

It sounds like the same bug.

Just now, pro grammer said:

The original model was a catia import. Not one straight surface on this thing. We were told the model had been translated from metric to inch.

Are you using moldplus catia converter ?

did you try using the model prep optimize/simply functions?

these aren't sheet body solids are they? 

[pro grammer]

9 minutes ago, byte me said:

You can automate this process, that's what I did.

I would like to know how.

The issues were primarily 5ax curve and 5ax swarf that were giving me the headaches.

[Guest]

Just now, pro grammer said:

I would like to know how.

First create a chook worker.dll . Then you can use prebuilt Mastercam Sdk functions to copy the parameters and toolpaths from one operation structure to a new one you create, then delete the original, It works everytime for me.

Just now, pro grammer said:

The issues were primarily 5ax curve and 5ax swarf that were giving me the headaches.

Thanks.

The code is mostly trivial it's just long to type in all the parameters from the operation, because mastercam operations structures are really really big.

[Guest]

Just now, pro grammer said:

I would like to know how.

If you want to build the patch yourself, you need to download visual studio community edition (free). 

Follow the first and second video in this topic to access Mastercam's api.

I'll hook you up with the code to get the info you need.