Cnc brainfart

Whitbread · Feb 19, 2020

Was running my 96 BP VMC1000
for 8 hours straight today doing 3d profiling on an investment cast mold part. At 530, I was a few feet away at the workbench waiting for the nearly complete program to finish. Suddenly I hear endmill bits go flying and a rapid table move. Run over, mill is error'ed out, and table is 14" X negative of part zero and endmill clearly lost all 9 lives in one go. I can save the part with a touch of weld, but dammit it was a brand new maritool long reach profiling ball endmill haha!

The program simulation definitely did not show any kind of X-12"+ command, and machine readout was at X4.xxx" when the table took off to the negative. After a power cycle, tried homing, and it's now stuck on home position and errors out as soon as drives are powered on. Been running perfectly for last 9 months. Oh the joys of cnc hah. Off to troubleshoot after dinner!

YouTube

guest · Feb 19, 2020

Wow, looks like you dodged a bullet there.

plastikdreams · Feb 19, 2020

Might want to clean the machine a bit lol

G00 Proto · Feb 19, 2020

Almost sounds like what happens to me when the CAM software randomly throws in a G92... as far as I can tell, that resets the entire machine coordinates, crashes the machine and forces me to do a complete re-setup. For me, it's a Mastercam glitch that occurs every couple hundred programs... preferably at the least opportune moment.

dksoba · Feb 19, 2020

Are you using DNC or is the program stored on the machine?

Whitbread · Feb 19, 2020

guest said:
Wow, looks like you dodged a bullet there.

No kidding! Had it hit the side of the fixture table, it would've surely cost more than a dab of weld and a $30 endmill!

plastikdreams said:
Might want to clean the machine a bit lol

Yeahhhh, the plan was now that all hogging was done, run the 3d profiles, let machine sit for 2 days to drain, then empty it all out.

G00 Proto said:
Almost sounds like what happens to me when the CAM software randomly throws in a G92... as far as I can tell, that resets the entire machine coordinates, crashes the machine and forces me to do a complete re-setup. For me, it's a Mastercam glitch that occurs every couple hundred programs... preferably at the least opportune moment.

It's definitely a possibility! I'm going to search out the line of code it puked on and see if there's any clues there. But most likely the code will be just dandy because murphy is a total asshole haha.

dksoba said:
Are you using DNC or is the program stored on the machine?

It was running in DNC mode from a file on the hard drive generated by fusion 360 with the DX32 specific post. I was doing a few surfaces at a time to keep programs in the 700-1000kb range and 45-60min run time.

litlerob1 · Feb 19, 2020

G00 Proto said:
Almost sounds like what happens to me when the CAM software randomly throws in a G92... as far as I can tell, that resets the entire machine coordinates, crashes the machine and forces me to do a complete re-setup. For me, it's a Mastercam glitch that occurs every couple hundred programs... preferably at the least opportune moment.

Waaaaa??¡!!?? Your post isn't perfect? Like all the other dudes on this site claim!! (Note, just breaking balls. Some of them are my friends).

R

Whitbread · Feb 19, 2020

Took a look at the code, it looks good to my newbie eyes. The line in red was the current line when the machine barfed out.

X4.4624 Y0.3252 Z0.8941
X4.4579 Y0.3411 Z0.894
X4.451 Y0.3648 Z0.8941
X4.4486 Y0.3728
X4.4409 Y0.3966
X4.4355 Y0.4124
X4.4352 Z0.8945
X4.4358 Y0.3966 Z0.9009
X4.436 Y0.3886 Z0.9042
X4.4361 Y0.3807 Z0.9075

Powered up the machine, enabled drives and it didn't immediately error out like before. Hit home axes and of course it homed just fine like nothing ever happened :angry:

. Must be the ghosts of previous operators decided to mess with me. Guess it's time to clean it out tomorrow, check all cannon plugs, home switches, cables & cards for seating, and hope it never happens again!

mmurray70 · Feb 19, 2020

dksoba said:
Are you using DNC or is the program stored on the machine?

At my last job we had a Fanuc mill run by DNC and it would randomly make completely messed up moves like that every now and then. Maybe once every couple months.

I dont miss working on that machine!

G00 Proto · Feb 19, 2020

litlerob1 said:
Waaaaa??¡!!?? Your post isn't perfect? Like all the other dudes on this site claim!! (Note, just breaking balls. Some of them are my friends).

R

Post is solid, however, there is an occasional glitch when I propagate old legacy files that were created in older versions of Mastercam. My old machine definitions and posts had started life in 2002ish in version V8, and then migrated and modified for the last 18 years... somewhere along the line something new was introduced that very occasionally pukes up a G92. No problems on new programs, just legacy files... it doesn’t happen frequently enough to remember to check.

Sent from my iPad using Tapatalk

Perry Harrington · Feb 19, 2020

Hardinge CHNC lathes with the Siemens control would do that when the motor had issues. IIRC the wires to the encoder would break and it would go off yonder in a flash.

Shawnrs · Feb 20, 2020

Whitbread said:
Took a look at the code, it looks good to my newbie eyes. The line in red was the current line when the machine barfed out.

X4.4624 Y0.3252 Z0.8941
X4.4579 Y0.3411 Z0.894
X4.451 Y0.3648 Z0.8941
X4.4486 Y0.3728
X4.4409 Y0.3966
X4.4355 Y0.4124
X4.4352 Z0.8945
X4.4358 Y0.3966 Z0.9009
X4.436 Y0.3886 Z0.9042
X4.4361 Y0.3807 Z0.9075

Powered up the machine, enabled drives and it didn't immediately error out like before. Hit home axes and of course it homed just fine like nothing ever happened . Must be the ghosts of previous operators decided to mess with me. Guess it's time to clean it out tomorrow, check all cannon plugs, home switches, cables & cards for seating, and hope it never happens again!

Look ahead in your program for a large x move. It shows it erroring on one line but the error may be further along and your screen did not keep up with the code. Most errors in our programs are usually a few line past the spot where the machine errors out.

gustafson · Feb 20, 2020

If that pic is right after it errored, which it appears to be, the machine has no idea what has happened. I have not run a DX, so I don't know what it does when it loses encoder signal, but the same machine with a Heidenhain would have EMOed in ~50 thou. I see no M functions in the code, and it appears to be telling you it is trying to move to the red highlighted line, and the stated[wrong] positions are correct except for Y.

Wild guess: X encoder loses connection, keeps travelling to try to make X, hits part, throws Y out of spec, which is the error displayed. It has no idea the X is out 20 inches.

Ianagos · Feb 20, 2020

I have not had good luck with dnc. Possibly got a code sent incorrectly through dnc crashing the machine. The serial connections can act strangely. At least that’s my experience.

gmc1724 · Feb 21, 2020

Do you have any power cables running along side your data cable? Had a similar problem many years ago where data was getting corrupted when being tranferred, turned out to be interference from a nearby 415v cable.

Guest · Feb 21, 2020

I had a PC driven CNC Mill that would go bonkers from time to time. It would be profiling a plate and then just all of a sudden take a wrong turn at high speed and try to cut the thing in half. It would over travel and show insane "distance to go" numbers. After a half dozen or so accidents I found the culprit. The fan in the CPU cabinet had died and it was like an oven in there. A new fan and that WTF incident never happened again. I suppose since it was PC driven the PC itself would not trigger an overheat alarm.

Griffin · Feb 21, 2020

We have a vmc1000 with dx32 myself. We also use fusion360. We have had it occasionally loose it's mind and just cut through the middle of our part. Go post it again and everything worked perfect with no changes in CAM. We also had an x axis encoder go bad and cause the issue you experienced. I'd offset it high enough to clear everything and run your program again to see what it does. You can also choose no z moves if you want to run it that way. I'd also check the cables going to the x axis encoder. They run down the left side of the bed and under the left end of the table to the x axis drive motor. If the chips get real deep, they'll be dragging on it. The connectors aren't awesome and deep chips wiggling them around might cause some communication errors.

charlie gary · Feb 21, 2020

There was a time when a shop I was in had trouble like this. We checked all kinds of things, and then the owner decided to check the electrical connections in the circuit breaker box. The terminal for the ground wire was loose. He swore, tightened the screw and the machine never went rogue again.

Whitbread · Feb 22, 2020

gustafson said:
If that pic is right after it errored, which it appears to be, the machine has no idea what has happened. I have not run a DX, so I don't know what it does when it loses encoder signal, but the same machine with a Heidenhain would have EMOed in ~50 thou. I see no M functions in the code, and it appears to be telling you it is trying to move to the red highlighted line, and the stated[wrong] positions are correct except for Y.

Wild guess: X encoder loses connection, keeps travelling to try to make X, hits part, throws Y out of spec, which is the error displayed. It has no idea the X is out 20 inches.

Yes, that Pic was as I walked to up the machine after I heard the noises. I definitely agree with your wild guess as mine was similar haha.

Ianagos said:
I have not had good luck with dnc. Possibly got a code sent incorrectly through dnc crashing the machine. The serial connections can act strangely. At least that’s my experience.

The dnc code came right off the machines hard drive. It was not transferred over a drip feed cable.

gmc1724 said:
Do you have any power cables running along side your data cable? Had a similar problem many years ago where data was getting corrupted when being tranferred, turned out to be interference from a nearby 415v cable.

No, I don't have a data cable hooked to a computer. I use a floppy to transfer to machine hard drive, then run dnc off the hard drive.

Griffin said:
We have a vmc1000 with dx32 myself. We also use fusion360. We have had it occasionally loose it's mind and just cut through the middle of our part. Go post it again and everything worked perfect with no changes in CAM. We also had an x axis encoder go bad and cause the issue you experienced. I'd offset it high enough to clear everything and run your program again to see what it does. You can also choose no z moves if you want to run it that way. I'd also check the cables going to the x axis encoder. They run down the left side of the bed and under the left end of the table to the x axis drive motor. If the chips get real deep, they'll be dragging on it. The connectors aren't awesome and deep chips wiggling them around might cause some communication errors.

I've had that happen just you describe as well! Was making a flat plate with 2 posts sticking up 1/2" as a grinding jig for wood chipper blades. Machine was 2d adaptive clearing about halfway through operation, then just went Y- right out the front of the part cutting off half of the 2 locating the posts in the process. Reposted job with no cam changes, ran program again and part came out perfect.

I cleaned chips out and checked the X encoder cable. Turns out it was replaced within the last few years with a new cable from EMI. Opened encoder connector up, no signs of anything foul going on so I blew it out and reassembled. Checked other end of cable connection on axsbob card, all looked good there too.

charlie gary said:
There was a time when a shop I was in had trouble like this. We checked all kinds of things, and then the owner decided to check the electrical connections in the circuit breaker box. The terminal for the ground wire was loose. He swore, tightened the screw and the machine never went rogue again.

I'm going to hit all the grounds this afternoon just to be safe!

Cnc brainfart

Plastic

guest

Guest

Diamond

Hot Rolled

Hot Rolled

Plastic

Diamond

Plastic

Stainless

Hot Rolled

Titanium

Stainless

Diamond

Stainless

Aluminum

Guest

Guest

Aluminum

Stainless

Plastic

Similar threads