Need Help!- What causes plunging?

[Sinkoumn]

I'm just curious if anyone can tell me what causes the bit to plunge, more specifically in my case, I was cutting this board the other day. The total cutting time was 4 hours, everything has been running perfect; however, 3 hours 45 minutes into cutting the board the CNC decided to dive right in - luckily I was there and able to stop it before the bit snapped or the machine exlpoded in a fireball inferno .

But before I try to remedy this from happening again, does something like this happen because of a Z-axis binding, Mach3 missing a step, electrical spike, or what? To make it more broad, is this problem structural, or electrical? Or, does this just happen and there is nothing that can really be done about it?

I'm using Mach3, ArtCAM pro, Joes 06 CNC, and a HobbyCNC Pro board with 305oz motors.

I'm just curious and open to anything that you all have to offer.

Thanks much!
Mike

[ftkalcevic]

> Z-axis binding
Possibly.

> Mach3 missing a step
Unlikely. Mach3 is pretty well tested and is unlikely to have a serious bug like that.

> electrical spike,
This can be an issue if you are using servos. Electrical noise can corrupt the encoder data returning to the drivers. Did you measure the X and Y position? Were they off too?

> or what? To make it more broad, is this problem structural, or electrical?

I've had this happen once because the cutter was not secured in the collet and slowly pulled out.

If you are using steppers, it is usually because the acceleration is too much for the steppers and they miss steps. This is just configuration. The fact that the cutter went deeper would mean the stepper motor missed steps when going up, ie lifting your spindle against gravity. You can try and tweak the mach acceleration settings.

Strange though that it happened after 3 hours. Did you change the feedrate at that time? I'm not sure if heat will effect the motors. Were they warm?

> Or, does this just happen and there is nothing that can really be done about it?
Absolutely not. When set up right, the machine shouldn't miss any steps.

[casy_ch]

I had a case like that last year. The manufacturer told me I probably had resonances (mechanical). I reduced somehow the feedspeed and the problem was solved.

You do not say what size your board has and with what kind of cutting mill you have been working. It seems to me that 4 hours is very long unless you are using a mill cutter of Ø0.04". With finger mills of Ø 0.1 up to 1/4" I cut between 0.2 up to 0.4"/sec.

Jean-Claude, Switzerland

[troy.edwards]

Plunging into a job can be caused by many things, if you have post processed a programme then it is always worth checking it in your pc's softwares graphics first & do the same on your cnc (if it has a graphics mode). It is possible that the post process software has a bug if it has been recently updated. We had a similar problem with VISI 15. when we updated the post processor did'nt work properly & it took a call to the software company to rectify the problem (the end result if we did'nt check it in graphics would have junked a £70,000 form tool). Other problems that can occure are much more seriouse but are rare, we did have a machine once suddenly punch a milling cutter through a very expensive aviation component. The programme had been used on many occasions & had already produced a pass off & run a few others off. The end result of this was the TNC had failed & had to be replaced by a technician.

[oliver9866]

hi, dear friend. for any specification cnc routers information, pls contact me. thanks. hope we can share our ideas. It is my pleasure.

[Paulo E.]

I would go for the more likely culprit and them work my way back. 9 times out of 10 this happens because your tool was not secured properly and or over time it becames loose. More than likely that was your problem, not to say that anything else can't happen but 3:45 minutes later and it plunges it if was missing steps you would have probably been able to detect earlier. God knows it could be anything LOL. I think that if it was a spike in power and it missed it's location on the Z axis it would had happened for the other two axis. I guess the only thing you can really do is to try and replicate the problem. Secure the tool properly and let it run and it it happens again you can start to isolate the problem.

[High Seas]

First a question:
After the plunge wher did ZERO Z show itself? (Odd use of words but maybe you get my drift--I'll try this next one) WhHen you told MACH3 to zero out was the bit not at surface level -- but below? Could be missed steps -- OR Loose tool.

Thats what happened to me on a plunging incident. The bit was in the stock -- not at the surface when at Z Zero.
I used the fewer steps remedy - backed off the 1/8 steps on my Z axis to 1/4, changed and then the motor config -- DON'T forget to do that! While there I lowered speed and accel on the Z as well in the config panel.
I also reduced my feed rate on the Z axis (Plunge).
Can't say which fixed the problem -- to many variables changed - but I got the job done.
I can now go back and undo each.
BTW my board was Xylotex - but similar circumstances.
Jim

[Sinkoumn]

Thanks for all the great info guys! I want to add few more details on my particular situation to maybe narrow down the culprit.

I was cutting a longboard that was made of pressed ply 3/4" thick, and was cutting using a 60* V-bit going at a max depth of .15". The speed was at 70ipm on the X and Y, and 30ipm plunging for the Z.

The board is a symetrical design, so I had to cut each half individually (so cutting time was 2 hours total per end (thus the total 4hrs for cutting the whole board)).

The thing that gets me is that I cut the first half of the board flawlessly with no mess up as you can see in the picture I attached.

While I'm not ruling out the bit from coming loose, I don't think that it was the problem in this case because after it plunged I simply hit the estop, raised the Z axis back up, rezeroed the bit to the top of the material, and began to cut a few lines before the bit decided to dive in on the G-code - thus I didn't retighten the bit, and it finished the rest of the cut without issue (~15-20min of cutting time).

I wish I had taken note of what Mach3 said my Z position was, but I was about three seconds away from putting my foot through the concrete wall.

The part of the design that the bit dove in on was a simple "machine along vector," where it should have only cut .065" deep for the entire outline, so that is why I think it either missed a step or the Z bound up and didn't retract back to the safe Z height (which I had set at .25" - coincidentally this is about the same depth as the mess up).

I will try to monkey with a few aspects relating to all the info you all provided, but please if something else comes to mind from the additional info I added please share!

Thanks so much again!
Mike

[cncwhiz]

Did you look at the program to see if it had a programmed "Z" move to a deeper depth. With as small of cut you are doing, I don't see the tool sucking out of the collet? Those are some cool decks and with all that time in the work guessing at the issue is not the good way to go?

[Lindsay h]

One thing to check is that jogging is disabled when you run a program as a jog input could cause this result. I had it happen after a 2hr program and I thought of the jog causing it when pondering lying in bed that night.I always disable jog now and I haven`t had a problem since.

[jimmearns]

I have seen this, only bvery occasionally the Mach3 misses a line - don't know why but it always appears to be the Z up move. I have reduced the incidence of this by making sure the computer is not running ANY background programs. Also, on critical parts add another Z up to the g code so if it misses one it will do the movbe with the second command. Bit of a nuisance hand coding in another line, but better than remaking an important part.

[Sinkoumn]

I like the idea on the extra Z up, but I hope I can solve it by changing the accel/velocity, or the steps first. Because that would be a lot of adding code lines.

That longboard cutout was just under 150k lines of code!