Do computers/controllers make mistakes?

[powerglider]

I've always been under the impression that computers don't make mistakes only the programmers do. But what happenned a few minutes ago just makes me wonder if it only happens to me.

I'm new to cnc machining, bought a used SuperMax milling machine with Anilam Crusader 2 controller. I was able to make my computer download programs through the rsc 232 port on my own. Everything looked rosy.

I've been successful at cutting so far in maybe 20-25 different toolpaths.

So today I load a program and run it, no problems for the first 8 pieces I cycle through. But on the 9th the Z axis starts heading south progressively on its' own! Cutting through my material (lucky for me the vise wasn't in the way or it would have ate it too!). Well there is no change in Z axis, only one line of code to lower it (third line) and then one line of code at the end to bring it back up before homing it to the start x=0, y=0 position.

I run the program again 5 times and can't reproduce the same effect (with no material).

SO have others had this happen to them? There have been no power variations I could account for, no other buttons were touched. The only thing I can think of is that I do have a few small magnets that I use to hold pieces of paper close to the servo motor and also close to the shielded control lines to the Z axis. COuld they possibly have set it off? They've always been there (except now I removed them).

Or is it just something else I must have done and the controller or servo couldn't have just acted on it's own?

Well, I don't know what to do now, I just ruined a pretty expensive piece of material.

Better go eat some jelly beans to relax...

Francesco

[Al_The_Man]

The method used by most controllers is that in the case of an axis running away, a parameter in the controller is set to alarm if any axis following error exceeds the number of encoder counts in the parameter, it should then turn off the drive enables or issue a E-stop which should achieve the same thing.
Either this control does not have it or the wrong # is in the parameter.
Al.

[powerglider]

Thanks for the reply.

I should explain that the z axis didn't take a dive at a fast feed rate.

The program called for the tool to move from a starting position down to a Z coordinate, then to move in a straight line in Y then in X then in Y again to create a slot, but on the second pass it started also slowly and progressively moving to a new Z position with no code to account for it and no buttons pushed.

I haven't been able to understand what might have happenned.

Francesco

[HuFlungDung]

Did you, prior to the crash, abort your program for any reason and start over on the 9th time?

Did you EStop it, or how far did it go when it took a stroll? Did it stop at any significant level? What Z level did it return to after the errant move? If the controller loses position, then it is unlikely that any movement after that would recover the error.

Maybe post your sample code here, in full, in case we can spot anything unusual about it.

[powerglider]

OK, I found out what I did wrong, it has nothing to do with cnc or the computer it has to do with collets and the right way to secure a tool.

I was using a 3/4" DA180 collet to secure a 3/4" end mill and it must have "climbed" out of that arangement after hogging 8 pieces. The giveaway for my friend that figured it out was the blackish residue left between the collet and the end mill (I had never had that happen to me in the 3 years I had used my manual knee mill).

So now I secured it in a proper tool holder with the set screw nice and tight (my friend showed me how to do that as I had never done that before being new to Cat-40 tooling).

Well thanks for the ears and help, I'm just glad it didn't slam into my nice Kurt vises!

Francesco

[Geof]

Originally Posted by powerglider ....I run the program again 5 times and can't reproduce the same effect (with no material)....

Yes, very difficult to replicate what happened without cutting material. I have had a similar occurence but identified it almost immediately by the slowly descending gouge in a vise jaw.

[powerglider]

Exactly why it drove me crazy for a few hours this afternoon, because I had not taken the tool out to see the blackish residue until my friend asked me to do it so he could inspect it. Gee I wonder now what he milled out the first time it happened to him, I'll have to ask him.

Also because I have not indexed any of my end mills as I change them too often and just reset the Z zero with the jog and a feeler gage I never even noticed the tool had moved downward, and also when it started hogging material I didn't want I immediately stopped the motor and then sent the Z axis up with a manual jog. Once I started problem solving I had no more material.

Sorry to hear you milled out your vise, I would not have been happy about that!

But overall boy am I happy that I know what caused it instead of doubting the cnc machine or controller which can really mess your head up (I was thinking about the quotes I had just put out this morning and wondered if I wanted to call them to put them on hold).

[Geof]

Originally Posted by powerglider ...Sorry to hear you milled out your vise, I would not have been happy about that!...

You get philosophical or you stress out too much. Major problem occurs when someone else screws up and you are all ready to ream them out when they let their eyes stray over to the vise in question.

[Willbird]

This is exactly why I don't put endmills in collets on cnc mills :-). Some folks do it all the time, but one deal like that will convince a man :-).

I have seen a few odd ones that made me tear my hair out, one was holding a part in 4th axis, in 6 jaw chuck, and milling (4) windows in the od of the part, two of the ports had a funky deal in one corner up near the chuck, the window wasnt square there. After about an hour of messing around we figuired out the machine was bumping into 2 jaws of the 6 jaw and rotating the A axis..that was why it happend in 2 out of the 4 windows, and the.

I have seen a fadal 3 axis wander off while cutting, usually happened during an axis rotation (G68) the huge following error was one clue, also if you defragged the machine meory it would gain 5% to 10% more free space for programs, my guess was some kind of memory leak in the code the machine runs on.

But for the most part garbage in garbage out :-). machines seem to KNOW they have a new guy to play with, they settle down after you get a few more hours at the switches :-).

Bill