G-Code Toolpath Verification As A Service?

[Matthew Hajicek - Singularity]

Are there are companies out there that do this? I imagine there's a fair number of shops that are getting by without g-code verification software which would occasionally pay to have a more complicated job verified.

[JAMMAN]

+1, I work at one of those shops without verification. First runs on new programs we call "white knuckling" because of the tension on your finger at the red button. I would pay for it.

[Sticky]

+1, I work at one of those shops without verification. First runs on new programs we call "white knuckling" because of the tension on your finger at the red button. I would pay for it.

 

Serious question, how much safer would you feel running it through G code verification? Would you actually just let it go at full rapid? Or are you still going to turn the rapids down and single block your way though things with caution?

[Watcher]

I don't think is realistic yet. The number of scenarios and machine configurations to support is beyond countable. Nobody can guarantee the customer success.

 

(Sent from my Moto X via Tapatalk)

 

 

[Mick]

I believe it is doable. It just purely depends on the machine, and the job.

[Helmut]

It's absolutely doable. It would just be limited to very specific machine/controller combinations. I'm sure IF one started a business doing that they would come up with a good way of incorporating new machine/controller combos as a way of expanding business. I think the biggest variable would be the post processor. Even with the same machine/controller the ability to have infinite different posts would make that the biggest issue.

 

IMO of course.

[robk]

I belive CCAT offers that service.

http://www.ccat.us/mantech/section/Our_Resources

[Newbeeee™]

Here's a scenario (why the bloody hell is it always me who'se different...)

 

We just bought a new VMC (with 60m/min rapids).

Toolpath cutting down inside a 'box' with internal lugged sidewall. Tool rapid ups 1/8 and rapids to next position and back down 1/8 and keeps cutting.

Machine is set for straight move (parameter check and later proven in mdi).

Machine ran ok at 25% rapid - machine clipped internal boss at 50% rapid, machine fooked internal boss at 100%.

AE from MTB says it's because the acc/dec of the rapid is soooo fast, the axis launch full bore in G0 state and before they are back under control, the lug has been clipped (the programmed rapid move was a straight move - but it was shallow in X and long in Y).

Fanuc will be visiting to see if this is correct or BS...

 

BUT, I wouldn't have the 'nads to hit the green button and go and put the kettle on.

1st prove out here is hand on rapid override running between 0+25% and hand on feed override, because with feed override at 0 machine axis stop (aka feed hold).

Now we are self taught so if we're doing it wrong please tell me...

 

We also run everything through NCPlot 1st.

 

And hats off to the mould (that's MOULD ) makers like Rick who run overnight and can actually sleep, without thinking that they're gunna come into work in the morning to a train wreck!

[Mick]

We did it as a favour to a customer of ours twice. Purely just a favour, as we owed them. I actually built a machine/control model so I could do it. It worked a treat, and they were very grateful

 

I've Vericutted a huge NC file, cutting several dies all set up on the machines table. I pushed the green button mid Friday afternoon,and left it running all weekend. It finished Monday morning at about 5am. I walked in the door, looked in the machine, and......

 

 

 

 

 

 

 

 

Everything was perfect. Great product, that Vericut

[?Mark]

Serious question, how much safer would you feel running it through G code verification? Would you actually just let it go at full rapid? Or are you still going to turn the rapids down and single block your way though things with caution?

I feel a lot saver after running thru vericut. Not even one "bump" so far, but as watcher mentioned, there are many variables.

Single blocking IS impossible though, not with 250K to a milion lines of code...I'd break a cycle start button going thru one program

[Sticky]

I feel a lot saver after running thru vericut. Not even one "bump" so far, but as watcher mentioned, there are many variables.

Single blocking IS impossible though, not with 250K to a milion lines of code...I'd break a cycle start button going thru one program

 

I only single block through rapid moves, plane changes etc I mostly just machine nasty stainless so single blocking through a feed movement isn't going to work so well for me

 

When I do mold style part with long surfacing parts I let them run un attended through the night. I just go through the code, check all my retracts and rapid planes, triple check tool length offsets are correct and let it roll.

[Guest]

Or are you still going to turn the rapids down and single block your way though ...

 

Single block? People do that?

 

 

I mean if your programmer is bad, but jeez... I just can't imagine single blocking through an entire program. I'd quit my job before doing that. Seriously. I'd quit, without even thinking twice.

 

When I run a program for the first time, I still run 10% or 25% rapids. Just because I'm normally running other people's machines, to me that's a reasonable precaution the first time through but single block... I just can't even fathom that.

[Sticky]

 

Single block? People do that?

 

 

I mean if your programmer is bad, but jeez... I just can't imagine single blocking through an entire program. I'd quit my job before doing that. Seriously. I'd quit, without even thinking twice.

 

When I run a program for the first time, I still run 10% or 25% rapids. Just because I'm normally running other people's machines, to me that's a reasonable precaution the first time through but single block... I just can't even fathom that.

 

 

Lol, why don't you just crank it up to 100% rapids, turn out the lights and go home on an unproven post James? You could maybe ask questions before saying I'm a bad programmer.

 

 

If I hadn't been single blocking and checking clearances last weekend I'd be out 1 spindle, 1 tombstone, 1 tool holder, 2 parts, and about 50k. All because of a Mastercam BUG. NOT because I didn't program the correct values, but because Mastecam didn't DO what it said it would.

I am ONLY working on hmc's, and most of it is multiple sided work on tombstones, so taking a little bit of extra time to make sure all clearances, tool offsets, and work offsets are correct is a good way to stay crash free, particularly when I am constantly trying to find/deal with bugs and post modifications. When I have a spindle inside a window fixture, in between corners of a tombstone, or on the left side of a tombstone I use caution.

[Guest]

Lol, why don't you just crank it up to 100% rapids, turn out the lights and go home on an unproven post James? You could maybe ask questions before saying I'm a bad programmer.

 

Maybe you could read the entire post before you put words in my mouth.

 

When I run a program for the first time, I still run 10% or 25% rapids.

[Sticky]

I didn't put words in your mouth, I was making a sarcastic comment about why it evens matters if there aren't any problems in your programs?

 

I'm not sure if you actually thought I was single blocking though feed movements through out the program, which you would have known is not the case if you read post #11.

 

Then I had to give you an example of how single blocking through planes and clearances saved my spindle and more due to a BUG in Mastercam, which running at 25% would NOT have saved anything, The spindle was about 1.5" away from the side of the tombstone, the machine would have crashed before you even knew that it was going in the wrong direction.

[Newbeeee™]

I only single block through rapid moves, plane changes etc I mostly just machine nasty stainless so single blocking through a feed movement isn't going to work so well for me

 

When I do mold style part with long surfacing parts I let them run un attended through the night. I just go through the code, check all my retracts and rapid planes, triple check tool length offsets are correct and let it roll.

Sticky - just a thought.

Have you looked at NCPlot? It's a code backplotter but you can configure it so rapid is red (for example) and you can quickly run through your code and look for the red moves.

We use it and wouldn't be without it, but we're only running vericals with 4th's mounted.

Also, it maybe worth having a word with Colin G now he's working for himself. If you want some bullet proof post mods I'm sure he'd help you out.

[Allan]

Hey Newbee; I've seen the issue you describe, many times.

It has everything to do with "S" curve acceleration/deceleration.

And Vericut or any other Simulation software is not going to catch the issue period.

 

Typically we set-up posts to make rapids as high feed moves (G01), this forces the control to not introduce acceleration/deceleration motion that is unexpected.

The other option is to use absolute retract values, you'll end up with more Z rapid motion but at least you wont clip features on your part based on the rapid setting.

 

Allan

[Newbeeee™]

Thanks Alan,

We went the retract route and all was well, but I'm *sure* the machine can be tuned to sort this?

Any linear motor machine (Matsuura/Dmg???) doing 10000000000 miles an hour in G0 surely doesn't have this problem???

[YoDoug®]

Single block is for sissy's. They put a green button for go. If they made it for girly men to run they would have put a chartreuse button or some other color that real men can't identify.

[Mick]

Single block is for sissy's. They put a green button for go. If they made it for girly men to run they would have put a chartreuse button or some other color that real men can't identify.

 

Okuma breeds'em tough

[Sticky]

You guys are right, I'm just gonna start running these crappy unproven posts, from my unreliable software, at full rapids and go home, and just buy some spare $30k/pc spindles and just change them out when things go boom. Who cares right?!

 

[Sticky]

Sticky - just a thought.

Have you looked at NCPlot? It's a code backplotter but you can configure it so rapid is red (for example) and you can quickly run through your code and look for the red moves.

We use it and wouldn't be without it, but we're only running vericals with 4th's mounted.

Also, it maybe worth having a word with Colin G now he's working for himself. If you want some bullet proof post mods I'm sure he'd help you out.

 

No I haven't looked at it. How do you like it for 4th work? Is a code backplotter only or can you have your part model in there and everything?

[Joe788]

Single block is for sissy's. They put a green button for go. If they made it for girly men to run they would have put a chartreuse button or some other color that real men can't identify.

 

We have a little Okuma lathe with the absolute insanity of not having a rapid over ride. We call it the Charlie Sheen lathe - One Speed, One Gear, GOOOOO.

 

[Newbeeee™]

No I haven't looked at it. How do you like it for 4th work? Is a code backplotter only or can you have your part model in there and everything?

It's good. It only plots toolpath and you can't get a model in or anything like that.

Email Scott @ NCPlot.com and he'll answer your queries and I believe you can download and run a trial.

[Newbeeee™]

We have a little Okuma lathe with the absolute insanity of not having a rapid over ride. We call it the Charlie Sheen lathe - One Speed, One Gear, GOOOOO.

The siemens lathes we have, have a combined rapid/feed switch.

So for proving out you could be at say 50 or 60% feed which then jumped straight to 50% rapid...

Luckily it's siemens and every click of the dial can be changed, so we now have:-

<70% feed 5% rapid

85% feed 10% rapid

90% feed 25% rapid

95% feed 50% rapid

100% feed 100% rapid

It maybe worth looking to see if you can do something similar?