HMC programming training?

[Sticky]

CJep, not really. What I want to do it is program one part, then populate it vertically in the Y axis, and then on 2-6 sides of tombstone while maintaining individual offsets for each part.

 

The above picture is just a quick dummy model to test transform ops. What I want to do is use G54 for parts mounted on the B0 face of the tombstone, and add G55-59 based on how many sides of the tombstone I have. So a two sided tombstone would use only G54 and G55, while a six sided tombstone would use G54-59. And any side work on the parts would use the extended offsets for that side. IE G54 is for parts on B0, so side work on a part mounted on B0 would use a G54 P12 etc

 

My controller uses J1-27 for extended offsets. So that is what I will use for the rest of my example.

 

 

What I would like to be able to do is program the bottom part that is facing out (B0), working on the face, right and left sides. I want the face of the first part to be G54 J1, then transform the tool paths Y+ to get a total of 9 parts, G54 J1-9. Then I want to work on the right side, starting at the bottom I would use offsets G54 J10, and transform up to G54 J18. Then on the left side start at G54 J19 and transform up to G54 J27.

 

Now that all the parts on B0 are complete and ready to go I would like to just transform to the other 4 sides. So parts on the B90 side of the tombstone would start at G55 J1 for work on the B90 index, and go to G55 J9, right side of those part G55 J10-18, and left side G55 J19-27.

 

On the contrary. You can use G54 for all of your faces. Our machine repeats within .001". Anything that needs to be closer can be tweaked with your system variables and macro variables with the WC shift for each face and then cleared after that face's process is complete. Yes, it DOES work for multiple parts.

 

Yes you could use G52 or something and work shift, or use the same offset for each face, and write over it each index using G10 or something, but both of those methods introduce more headache and problems then just having dedicated offsets for each face.

 

And no transforms for multiple parts on multiple panes is not so great, go back and read the link I posted earlier. Sure its fine if you only have one part on each index, or if you are just translating parts on one plane, but anything beyond that it takes a lot of post modifications to get it to work.

[guitar]

Actually. this is not a headache. We do it all the time, and have my posts setup appropriately. I'm sure if you were new to this, it may seem a little confusing. But, once the process is tweaked, it is extremely beneficial. Using the macro variables allows changes on the fly. Thank you for the interest. Multiple parts are not a problem.

[Sticky]

Actually. this is not a headache. We do it all the time, and have my posts setup appropriately. I'm sure if you were new to this, it may seem a little confusing. But, once the process is tweaked, it is extremely beneficial. Using the macro variables allows changes on the fly. Thank you for the interest. Multiple parts are not a problem.

 

What is the advantage to doing it that way? I'm not seeing any advantage over having offsets for each part/face.

 

Can you provide an example of the transform tool path doing what I am asking for, both in this thread and the above link?

[guitar]

Without writing you a program, try this:

Indicate (or probe) your part. I assume you indicate since you said this was a tight tolerance issue.

In your control, hopefully a Haas (since fanuc has not yet learned to store MDI values), if not write a program that loads the current machine position or the amount it is off position info into a macro variable location. Yeah, you can change the variable location with another variable on the fly. After indicating, you hit cycle start and your work shift is automatically loaded where you want. Piece of cake.

[Sticky]

Not a Haas. I already do in process probing, for setting offsets and inspecting parts.

 

I think you are missing what I am doing, read post 51 again.

[CJep]

Sticky can you associate the B rotation with the work offset in the registry of the machine control?

Then you would have the post setup to output B0. and the rotation would be picked up with the offset,..

 

The J output would require some custom logic in the post,..

[Sticky]

Sticky can you associate the B rotation with the work offset in the registry of the machine control?

Then you would have the post setup to output B0. and the rotation would be picked up with the offset,..

 

The J output would require some custom logic in the post,..

 

Yes I can, and I can write values there via G10. The post is already setup to output J, so that is covered.

 

The problem is transforming a translate/transform operation.

[CJep]

Yes I can, and I can write values there via G10. The post is already setup to output J, so that is covered.

 

The problem is transforming a translate/transform operation.

 

Well you wouldn't do a trans rotate using the method I'm thinking of,..

You would still use translate to setup the work offset and number of parts.

[Sticky]

Well you wouldn't do a trans rotate using the method I'm thinking of,..

You would still use translate to setup the work offset and number of parts.

 

So you are thinking of doing everything inside the post?

[CJep]

Yes, I'll PM yah,..

[Bob W.]

I ran into the transforming crash issue tonight. It is not related to drilling alone. My issue was with a contour toolpath where a contour was mirrored, then transformed (rotated 180 degrees). The tool wanted to go directly through the tombstone. Thankfully I caught it in Vericut and I remembered the issue from reading this post. Armed with the info in here I was able to quickly modify the program, re-simulate, and get to the machine before the current cycle was complete. Many thanks to the people chiming in on this issue, it undoubtedly saved me several hours of frustration. I am optimizing an existing program by transforming toolpaths instead of the entire program at the machine so losing 30 minutes because of an issue like this is a pretty big setback when the cycle gain is measured in a few minutes.