Couldn’t find any other threads about this issue.
As I was cutting some pockets, the Z was slowly moving down in a way that seemed programmatic. Using an offset cut, each pass Z would fall, I’d guess ~6/1000ths of an inch per ‘lap’. Here’s the result. This was supposed to be a flat pocket.
Bit was 1/4" downcut, 18000 RPM, 180IPM, 40% stepover.
Couple of things…did the wood move as you removed material (surface tension) In Jim Neeb’s laser documents he suggests changing one of the settings the fixes Z “drift” . I don’t remember the settings, but might be worth a try.
People have reported that loose stepper motor connections make this occur.
I have personal experience this with the brake in the Z axis slowly going bad which resulted in a slowly falling Z.
Before you mess with software parameters, reliably prove it is happening in a repeatable manor.
Stick a sharpie in the spindle and draw a series of lines back and forth across a poster board. Start with the height so that the line is very thin so you can gauge it losing height.
I do not think this was a case of wood movement. It created a fairly consistent sort of mountain across the pocket. I would expect wood movement to be more random. I’ll see what I can dig up from Jim’s videos.
Thanks, I’ll give this a try.
For additional details, I had 4 pockets that I was cutting all in the same file. The 4th of the 4 pockets was where the most pronounced drift was. After this happened I restarted Mach4 and cut my second piece, but with the 4 pockets in separate files run one at a time. The drift did not occur on the second piece.
Ah! That is a clue. You aren’t crazy, this gets reported a lot.
The dirty connector, bad wires are easy to check. While cutting in the air, wiggle your wires around. Just be quick with the estop!
For me it was a bad brake and thus every time it would do a rapid to Z 0 it would lose one or two steps. Over time it got progressively worse until it started diving 2 or 3 inches and burrowing into the spoilboard or worse, crashing into the tool carousel.
Best of luck. Never fun when the “precision machine” has troubles being precise 
https://drive.google.com/drive/u/0/mobile/folders/1xB6J0RZIU1j8g951NkP-8HXC1swC2C-B?usp=sharing
See if this works
It’s probably not the z drift issue that is fixed with changing the z step signal. In all cases I know of the symptoms of that are a drift up and it takes a lot of Z movements to accrue enough error to notice. It’s usually noticed on a 3d carve.
Thanks as usual Jim !
This is consistent with the bit working its way downward out of the collet from not being tightened enough or otherwise not being installed correctly (e.g., not snapping the collet into the collet nut prior to installation, or installing with debris in the collet). It happened to me, and I swore it couldn’t be the issue, it was too consistent, but sure enough, tightening to the correct torque fixed it.
Nothing else has shown up in my tests, and I want to believe this is what it was since that’s easier to solve.
I should also note that I re-tightened the collet between the first and second piece, which lends credence to this theory.
Well that is great news! Nothing more irritating than futzing with bad wires or flaky brakes.
So, I was doing a 3D carve today and I’m running into trouble with the z-axis. I did a roughing pass with a 3/8" EM, and then started my first clearing pass with a 1/4" tapered ball nose. The TBN clearing pass was supposed to take about 1 hour. About midway through it had turned into an air pass - it was never making contact with the wood at all. I stopped the cut, and commanded the machine back to work x0, y0. It was in the correct place lower left corner. But when I jogged the bit down to almost kiss the board it was showing as about z -.75".
So, removed the bit and collet - blew out any dust that might be present. Re zeroed the bit, and re-started the gcode at the right spot in the file. Seemed to be going ok, so I stepped out for the hour it should take - came back, and again I had very little cutting into the wood.
Commanded it back to x0 y0 - looked good. Jogged down to z0, and again it’s showing a negative z value at the top of the board. I am positive that the collet was clean, and the bit was appropriately torqued down.
Based on this year old thread - is the troubleshooting process to unmate and blow out/clean connectors, and then “wiggle” wires while it’s doing an air cut ??
TIA,
-scott
Scott, that is consistent with a problem a number of us have had with these machines on long 3D carves. If you don’t find any mechanical problems, then try setting your Z step signal’s active level to low and see if it fixes it.
Just to close the loop on this - I tried making the change as shown in the image posted. For me this did not alleviate the problem. After further consultation with Avid - I ended up changing from (Avid) Mach4 version 2.5.5 to version 2.5.10. After first test - this seemed to have fixed the problem.
Appreciation everyone’s feedback on this !!
-scott
