Z axis seems to lose it's position midway through cut

[brandon]

I just finally finished installing a 2.2kw spindle and getting the VFD set up with GRBL (what. a. pain.) and was doing some test cuts in aluminum. The first part went great apart from a slight amount of chatter in one specific area. (And it was QUIET!!!). However the next part is giving me trouble. I am trying to cut the same exact thing and essentially mid cut the Z loses its positioning and is higher than it's supposed to be. It seems to be a consistent 2.5mm above it's actual 0 point. It doesn't sound like the stepper is stalling out and I don't necessarily know how to tell if it loses steps. Admittedly the new spindle is heavy and I tested the motor stalling myself just to be sure that wasn't the case. Basically it is sort of sensitive when moving the spindle up but won't stall unless I put force on it. I couldn't get it to stall at all going down (presumably gravity + weight of spindle). Very weird that it would work fine then not work. I also DID happen to change the gcode between the two parts, but the only difference was I manually added an M5 command at the end to shut of the spindle since UGS disables M30 commands for whatever reason. I can't imagine this would have anything to do with it.

[Derek the Admin]

I would very it a couple times by air cutting to make sure it is losing steps again before you dig too much into things.

Did it cut all the way through the part or leave 2.5mm there?

I would check to make sure the coupler is good and tight to both the shaft and the screw also.

[brandon]

It didn't cut all the way through, no.

It seemed to do fine air cutting. Like I said before it also somehow managed to cut my first AL part fine. For the air cutting I set the bit on top of my material (with the paper method) and went up 5mm and re-zeroed. After doing the air run it was still 5mm (or extremely close at least) above the material. The coupler was pretty snug on both the screw and stepper shaft, but it does flex apart under weight. I assume that is normal given the design of the coupler and vertical weight of the rest of the axis? But hang on let me go test it again, but with the spindle on. I wonder if there's some sort of interference going on. Doubtful, but who knows.

Edit: Double checked with the spindle on. No change, still air cut fine.

[Derek the Admin]

The ability of the coupler to separate is inherent to the design of the coupler and OK. Seating the Z screw against that UHMW piece in the bottom plate limits unwanted axial movement and coupler separation. I would diagnose your issue as a matter of just plunging or ramping (please ramp instead of plunge) too hard in certain areas in the aluminum OR not hard enough allowing the aluminum to gum instead of shear. The Z stage is trying to lower but it isn’t clearing material well enough given the material, end mill geometry, and cut settings which results in a brief stall or series of brief stalls .

[brandon]

Alright so I just retested again with only ramps instead of any plunges. It still lost its position, but this time it went to deep and went through, well into my spoilboard. Then proceeded to break my bit. Woo. Ran the identical program on an air run, worked perfectly fine. Also double checked stepper voltage and that was actually very slightly higher than it's supposed to be so if anything it should have more power.

This wasn't an issue before installed the new spindle. I want to blame the spindle weight but I know I'm not the only CK owner with one of these spindles and I haven't been able to find any other posts about it.

[Derek the Admin]

When Alex installed his spindle like yours he had to build a faraday cage for the vfd. It was a giant electrical noise generator. When you air cut it’s not under load and I suspect it could be behaving differently in terms of the noise generated with load vs no load.

You could also try to lower your Z acceleration setting in Grbl just for good measure. That spindle is probably what, 10 or 12 lbs? That is certainly on the heavy side but it should be able to lift that all day long. Lowering the acceleration setting will reduce the risk of lost steps with a heavy spindle though.

[Derek the Admin]

Side note... the over voltage on the driver won’t help. It may be the issue actually. When the voltage is higher, the current is higher, and the temperature is higher. These drivers have thermal overload protection and will cut off very momentarily to try to protect themselves from cooking.

[brandon]

I'll be sure to lower the voltage back down. It was at something like 690 so it wasn't incredibly far off. I used to run it slightly higher on my MPCNC I built without (seemingly without) any issues. Also the first thing I did was halve the Z max speed and acceleration but I'll halve it again. And yeah the spindle is probably 12lbs or so with the mount. Honestly the included mount is a little overkill in itself, it's basically a huge cast aluminum block.

You may be correct about the electrical noise being the issue too. I had issues getting GRBL to output the RPM correctly because the frequency was bouncing off the walls until I added a capacitor (that was a fun week of troubleshooting).

[Bruce]

Where did you buy the VFD/spindle motor set and what model/brand is it?

[brandon]

This is the link for the exact set I bought.

[Bruce]

I think Alex also had to add shielded control cable or add a shield to the existing stepper motor cables.
(Speaking of Alex where's he been? Seems he fell off the earth!)
Here is his build thread:
millrightcnc.proboards.com/thread/598/aforww-ck-build-thread

[brandon]

Well I shielded off all stepper motor cables and the spindle cables (with the same stuff Aforww used) with no luck. Air cuts just fine, I put a piece of acrylic up just to test in and it first hole it cut seemed normal, second hole it went too deep on the first pass. Cut all the way through a 2mm (or 3mm can't remember, it was just for testing) in the second slot on the first pass. Stopped it, sent it back to zero and my new zero was about 1mm too low.

[Derek the Admin]

Do you have a router and mount? I wonder what would happen if you just remounted it and checked.

It's probably worth verifying your actual rpm. If something was set wrong on the VFD then it is possible that you’re getting a lower RPM than you are specifying. If that happened it might not be able to cut and clear material, thereby causing a temporary stall. It’s also worth while to make sure that the spindle is up to speed before it context of material. The VFD will be able to control how quickly the spindle comes up to speed. If it was still winding up when it first met the material you would also possibly stall.

[brandon]

Ok just tested it with the dewalt router. Figures, works perfect. Made a cut just in wood since the spindle even had problems with acrylic. I may weigh the spindle and mount and try to mimic that amount of weight with the dewalt on to see if that could be the issue.

I'm not saying it's impossible but I'd have a really hard time believing that, even in acrylic, the spindle with its torque would be causing a stall on material entry.

Edit: Ended up trying it again with the dewalt + a 10lb weight on top of it. This got both setups within 1/4lb of each other. Worked like a charm. I also had the spindle off to the side and let it run so maybe if there was any RF or electrical interference it would show. But it seems to work fine as long as the spindle doesn't touch any material while it's spinning. I would wonder if the spindle being under load would cause more messy electrical noise that could throw it off, but I think I'd notice issues in more than one motor.

[Bruce]

This seems like it is an electrical noise problem with the VFD to me. Are all the wire shields tied to ground or negative? Are any of the VFD wires in close proximity to other stepper control wires? I think the Z axis would be the hardest to isolate since the spindle is mounted there. If you move all the spindle wiring away from the stepper control wires...maybe route them to the front of the machine, does that help?
Also The VFD may need to be in it's own grounded metal enclosure. Also signals may need to be isolated with a 4-20mA signal converter. I think Alex used one for the PWM spindle speed signal from the CNC shield to the VFD.