Z way too deep intermittently

[donchapman]

Twice in the past week my Z has gone nuts and suddenly plunged a half inch or more too deep after otherwise working perfectly hours and days on end.
According to the SB screen it's at -.25" but in reality it's -.50" or more. I Stop and Quit the program, reset the bit to zero Z and everything works fine again for days, then it happens again. My best guess is that it's a static electricity problem, but I have everything well grounded and this is the same machine I've owned and used for 7 years. Any thoughts? Maybe sometimes it's just eccentric, old, and cranky like me. Are ShopBot owners and their machines like people and their spouses/pets who become more alike the longer they're together.

[Brady Watson]

Don,
Was this in the middle of a file with no human input or did you at ANY TIME use the S key, spacebar or insert command functions, and then resume?

I've never experienced it dropping during cutting, only after I paused and changed something & resumed. What version of the software are you using?

-B

[sawkerf]

Don, are you using a router or spindle? I'm using a Milwaukee router and last year I had the same problem. Turns out it was the bit turning inside the collet, moving up and down as it spun. I was very suprised(and relieved)when I finally figured it out. I have broken many,many new .25 compession bits due to movement in the collet. I suggest you check or replace your collet and see if that helps. Kip

[donchapman]

Brady: It was in the middle of a very large VCarve file and it did it all by itself. I'm sure I did not touch the controls during the running of the file.
I'm using the current DOS 2.38 SB control software, which is all I can use on my old PR96.

Kip: I'm using my original PC 3.25HP router and at the time of the Z drop was using a 60degree V carbide sign carving bit and it was doing very detailed but shallow routing, so the bit wasn't under any undue strain, and when I re-zeroed the Z it completed the file perfectly. Nevertheless, your point is well taken and I'll start marking and taking note of the position of the shaft of the bit in the collet and check to see if it has moved if and when this happens again. A year ago I did have a V-bit come out of a too loose collet and bury itsel in the material, so I've been very careful since then to make sure the collets are very tight.

Logically, movement of the bit in the collet is a more likely cause than my static electricity theory, since each time it has been only Z that has been affected, not X nor Y.

[donchapman]

All the sudden my Z went too deep twice again today. I checked the 60 degree V-bit I was using and it had NOT slipped down in the collet.
I was watching the router the second time it happened today and at about line 64,000+ in a complex V-carving job it just all the sudden starts routing about .09" deeper. I stopped and quit the file, rezeroed the Z, and used the File Part Go To command to resume the file just before where it went too deep. This time it stayed at the correct Z depth.
I'm back to thinking it's a static electricity problem, although I don't know why it would show up now, 7 years after I started using my SB, and only on the Z, not the X or Y.
Has anybody else experienced this problem?

[doglaw]

Don, you are not alone. I experienced the same problems several months ago. In my case I had two things going on. First, I was using version 3.4.26 which had a bug concerning preview. Prior to going into preview I had set the 'z' to .5". Coming out of preview, the controller thought the 'z' was set to .788". That was my first problem and Shopbot corrected it in version 3.4.27. The second problem was the bit slipping in the collet. I now ensure that I really tighten up the collet after bit changes.

[donchapman]

Thanks for the info, Douglas, but (1) I marked the position of the bit shaft in the collet and it had not changed and (2) my Z works perfectly through tens of thousands of lines of commands and then all the sudden starts going too deep right in the middle of running a long file. It then does just fine as soon as I rezero the Z.
It happens very late in long files with so much Z movement that the Z stepper motor is quite warm to the touch. Could overheating of the Z stepper possibly cause it to send the bit too deep? I don't see the logic in that. The random nature of the glitch still makes me think it's an intermittent interference and all I can come up with is a static electricity problem, but why now after so many years of no problems? I'm going to replace my ground wires with heavier guage wires, and see if that helps.

[davidallen]

Don,

I'm having a similar problem.

I was drilling some shelf support holes and the z kept dropping after a few holes were drilled. I thought it might the the z speed, so I programmed a file to move up and down continuously (I had the z plate below it just in case). After several hours, it still held position.

The next theory was that the plunge bit was dull and wouldn't cut fast enough. I slowed the z speed from 1 to 0.25 and tried to change the bit.

It was stuck to the point that I had to remove the chuck. I'm using the RC Eliminator and the screw wedge had worn to the point that it cut a notch in the bit collet (used a 1/4 to 1/2 adapter to hold the bit). It seems after 6 years, those things tend to wear out.

After replacing the chuck and bit, it still moved.

Plan B, change the design.

da

[donchapman]

Thanks for your input David.
Some months ago I had problems with the Z going way up instead of a little bit down whenever it was supposed to be routing quarter inch holes, and I solved that problem by reinstalling my SB control software.
I also once had a problem with a bit coming loose in the collet when I didn't tighten it enough.
My current problem is neither of those.
Yesterday a 60 degree V-bit once again suddenly started routing (about .07" this time) too deep toward the end of a large and complex V-Carve job. I heard it happening, stopped and quit the part file, rezeroed the Z, reset the file to where the problem happened, and the file ran perfectly to the end. The bit shaft had not slipped down in the collet, the command lines were correct, the bit was not under strain as it was performing very shallow details at the time, and no one touched the controls. My dust collector was off and ground wires were in place, so static electricity wasn't a likely factor. I'm running my SB 8 hours or more per day on a major job and it works perfectly through tens of thousands of lines of command on a number of different part files and then about once or twice a day it does this crazy too-deep-Z thing on different part files, the command lines of which are ok. I have 3 different 60 degree solid carbide V-bits that I change, clean, and inspect after each file run, which takes 1-3 hrs./file, and they are in good condition and cutting clean. I'm at a loss to figure out what's wrong and more tired at the end of the day because I'm having to be hypervigilant to keep the sudden Z plunges from destroying hours of work.

[richards]

Don,
There's a reason for everything - and usually the reason is to teach patience.

It looks like you've ruled out mechanical problems, such as a chipped tooth on the Z-axis rack, built up grease and dirt on the Z-axis and worn or chipped teeth on the Z-axis spur gear.

Assuming that the problem is NOT repeatable, we can safely rule out the cut file.

That really leaves static or faulty electronics. Something has to command a motor to move. Usually it's the program that does the commanding and the electronics that do the bidding of the program; however, here in Utah, with temperatures ranging between 0 and 25 degrees F, and humidity too low to measure, there is static. Depending on your gearing, it would only take a very few extra pulses to move an axis 0.07" OR a few missed pulses to keep an axis from moving 0.07".

On my machine, I use 6-guage cable as my ground line. It connects everything, including the computer case to a common earth ground. It does NOT connect to the controller's logic ground. It is meant only to drain off any electricity from any touchable surface.

Solving intermittent problems is probably the hardest thing to do. (I basically had that kind of problem that was caused by a wireless network card, which kept disconnecting and reconnecting at random, which interrupted the control computer, which caused all kinds of grief. The solution, once the problem was identified, was to remove the wireless card, install a normal wired card, and recommence cutting.)

You mentioned that you've had your machine for seven years and that you use it eight hours a day. As a process-control computer designer, I've seen non-typical, non-repeatable problems start to occur when by-pass capacitors crack or dry out. Those are the little capacitors that are usually mounted near the ICs on the controller card. I haven't looked at the controller card on my Alpha lately, but common practice is to have as many as one by-pass cap per IC and as few as one by-pass cap per four ICs. Sometimes getting out a magnifying glass and giving the controller board a quick look might show something. In this world of throw-away electronics, most circuit boards never live long enough to dry out and become flaky, but 14,000 hours on a machine would put it in the "it's time for an inspection" class.

Good luck.

[donchapman]

Thanks, Mike.
I agree with your reasoning and when faced with such a persistent problem I usually solve it through the differential diagnosis of rulling out all other possibilities until I'm left with the cause.

My PR96 is 7 years old but most of that time I've probably averaged 40hrs/month rather than the 40-60hrs/wk that I've run it the past couple of weeks.

I do think I've ruled out mechanical problems in the Z axis, which on this old machine is a screw drive powered by a stepper motor with a direct shaft connection, so there are no teeth to break off.

That leaves something to do with electronics/static as the likely culprit so I'll install larger gauge ground wires and take a look at the by-pass capacitors under a magnifying glass.

I cut just two files yesterday running a total of about 4 hours and had no problem, which may or may not be significant since it can sometimes run perfectly for hours before the problem recurs.

[richards]

Don,
I'm taking a wild stab here, since my total experience with ball screws is information that I've gleaned from web sites, but do you have thrust bearigs on the ball screw? A standard ball bearing allows axial play (up/down in this case), but a thrust bearing should eliminate all axial play. Is there any possibility that the ball screw can move 0.07" due to a worn/faulty thrust bearing? (I'll admit that this is a long shot, but anything is better than trying to find a leaky capacitor.)

EDITED: After reading your first post, where you had 0.25" movement, a faulty thrust bearing can probably be ruled out.

[workingdog]

Hi Don,
I have a PR with a ball screw. You have to watch out
for dust. It gets into the threads and builds up, until the Z looses step.

[donchapman]

Thanks Mike and Mike for your feedback.
The ball screw is clean and works fine with no vertical play.
I knock on wood but have gone two days totalling about 10 hours without Z problems and don't know why.
The only change I made about the same time was to a new CMT 60 degree V-bit rather than the Woodcraft bits I've used just fine in the past and which still look good and were cutting fairly clean.
I have a hard time imagining the bits were the culprit since they were making shallow detail cuts under very little strain.
Who knows? It may be that I inadvertently moved a sometimes-shorting wire or a spider web fell off an electical connection somewhere.
I'd like to know what caused the Z-drop problem, but I'll be happy if it just doesn't happen again.
I do appreciate your help and will let you know what happens via a post here.

[bob_dodd]

Don Check coulping between Z stepper & ball screw , make sure set screws are on flat part of shaft & tight Bob

[davidallen]

Don,

Can you correlate performance and humidity?

da

[donchapman]

Bob: The set screw is on flat of shaft and tight.
David: I haven't paid attention to a possible correlation of performance and humidity, but I think it's happened under both humid and dry conditions. I'll take note from here on.