The Dreaded X axis stepper error using UC400ETH and UCCNC !!!

[dachopper]

So guys,

I am fairly confident that I have diagnosed the problem correctly.

Background - I seem to experience random jamming in the X / Z axis, where by the stepper would loose any huge amount of steps - up to 5 or 10CM of error per event.
It only happened when running parallel type operations with multiple quick Z or X axis toolpaths, or both at the same time. I can run a parallel operation in the Y direction with Fixed Z height, and have no issues whatsoever for hours, but the other direction - within 10 to 15 minutes guaranteed motorStall in Z or X

I changed the power supply from 600W switching, to 700W torroidal, to 1000W torroidal.
I tried 2 x different Gecko drive motherboards
I bought new Ethernet cabling from the laptop to ETH400
I replaced the Geckodrive with a PDMX126 + 5 x Leadshine AM882 digital stepper drivers
I ran the motors under both 3.5 Amp peak setting, and 3.5 Amp RMS setting.
I observed stalls occurring randomly 10-20 minutes in, in every single one of those tried changes
I currently have the Max speed set at 80% of motors peak torque rating, and acceleration set at 10% of the slowest acceleration speed I was able to generate an inertial induced motor stall.
I cannot physically make these motors stall no matter how hard I push / pull / sit on the gantry while it's moving, yet it still stalls mid air unloaded.

After finally videotaping 3 stall events, none of which were cause by excessive speed, loading or acceleration, and also noting the AM882 failed to detect the stall.. I then unplugged the spindle and on the first attempt got to the end of the toolpath without error ( It had failed 100% of previous attempts, within 13:20, 12:00, 9:00 and 17:00 minutes on the recorded occasions - including 2 of those where the spindle was spinning in the air, with and without a tool in, and not cutting anything during the test run but with stalls both occasions. )

Based on the above evidence, that the stalls are not occurring in the same place, or even at the same time, and I can't find any mechanical issue with the machine, It has to be an electrical problem.
It is not effecting the Y axis, and the Y axis cabling does not run anywhere near the spindle cabling, but the X and Z axis runs right next to it the entire wire length.

So, I have ordered some heavy duty VFD cabling ( siemens Protoflex EMV ) and I am also going to re-wire the Z and X stepper motor cables, with some different Foil + wire shielded, and move them to either sides of my cable tracks, with the water lines in the center.

After re-doing the cabling, that is almost the entire electronics replaced, except the ETH400, and stepper motors themselves.

My toolpath is being tested by the UC400ETH retailer, to confirm they do not see any issues from it during the parallel operation, to try rule that out also.

Should be able to test in a few weeks when new cabling arrives, Let you know how it goes.


M_C, the stalls are loud high pitched, and without any movement at all from the jammed axis while the high pitched noise is active, which is normally for the 3 - 4 seconds, or the duration of the attempted move for that axis, until commanded to hold or reverse and it unstalls at that point. I tried straight after it stalled to force it, and it was futile, I would have needed a serious tyre lever to have any chance of making a mechanically forced stall happen with these low speed / accel settings.

[A_Camera]

So guys,

I am fairly confident that I have diagnosed the problem correctly.

Background - I seem to experience random jamming in the X / Z axis, where by the stepper would loose any huge amount of steps - up to 5 or 10CM of error per event.
It only happened when running parallel type operations with multiple quick Z or X axis toolpaths, or both at the same time. I can run a parallel operation in the Y direction with Fixed Z height, and have no issues whatsoever for hours, but the other direction - within 10 to 15 minutes guaranteed motorStall in Z or X

I changed the power supply from 600W switching, to 700W torroidal, to 1000W torroidal.
I tried 2 x different Gecko drive motherboards
I bought new Ethernet cabling from the laptop to ETH400
I replaced the Geckodrive with a PDMX126 + 5 x Leadshine AM882 digital stepper drivers
I ran the motors under both 3.5 Amp peak setting, and 3.5 Amp RMS setting.
I observed stalls occurring randomly 10-20 minutes in, in every single one of those tried changes
I currently have the Max speed set at 80% of motors peak torque rating, and acceleration set at 10% of the slowest acceleration speed I was able to generate an inertial induced motor stall.
I cannot physically make these motors stall no matter how hard I push / pull / sit on the gantry while it's moving, yet it still stalls mid air unloaded.

After finally videotaping 3 stall events, none of which were cause by excessive speed, loading or acceleration, and also noting the AM882 failed to detect the stall.. I then unplugged the spindle and on the first attempt got to the end of the toolpath without error ( It had failed 100% of previous attempts, within 13:20, 12:00, 9:00 and 17:00 minutes on the recorded occasions - including 2 of those where the spindle was spinning in the air, with and without a tool in, and not cutting anything during the test run but with stalls both occasions. )

Based on the above evidence, that the stalls are not occurring in the same place, or even at the same time, and I can't find any mechanical issue with the machine, It has to be an electrical problem.
It is not effecting the Y axis, and the Y axis cabling does not run anywhere near the spindle cabling, but the X and Z axis runs right next to it the entire wire length.

So, I have ordered some heavy duty VFD cabling ( siemens Protoflex EMV ) and I am also going to re-wire the Z and X stepper motor cables, with some different Foil + wire shielded, and move them to either sides of my cable tracks, with the water lines in the center.

After re-doing the cabling, that is almost the entire electronics replaced, except the ETH400, and stepper motors themselves.

My toolpath is being tested by the UC400ETH retailer, to confirm they do not see any issues from it during the parallel operation, to try rule that out also.

Should be able to test in a few weeks when new cabling arrives, Let you know how it goes.


M_C, the stalls are loud high pitched, and without any movement at all from the jammed axis while the high pitched noise is active, which is normally for the 3 - 4 seconds, or the duration of the attempted move for that axis, until commanded to hold or reverse and it unstalls at that point. I tried straight after it stalled to force it, and it was futile, I would have needed a serious tyre lever to have any chance of making a mechanically forced stall happen with these low speed / accel settings.

OK, so you continue to change/rebuild without knowing what actually is wrong. Maybe you will eventually fix it but without knowing what was wrong. It was waste of my time here since you are not answering any questions. Good luck.

[m_c]

The fact the motors are still trying to move, tells us the drivers are still trying to move the motors, which leaves only two possibilities. Either there is something mechanical jamming, or there is some kind of glitch in the step pulse stream causing the motors to stall.

Where did you source the ballscrews?
Something like a bit swarf in the nut, can cause these kinds of symptoms. It just occasionally gets dragged between a ball and the screw, causing things to jam. Stop motion, swarf gets moved out the way, and things start running perfectly again. Only way to find out if this is the case, is to dismantle the nuts and clean them.

Noise could cause a glitch resulting in a sudden deceleration/acceleration, which could be enough to cause the motor to stall. However I'd still expect that amount of noise to cause other problems. Key things I'd be checking/double checking, is any shielding goes back to a star point, and you've not introduced any ground loops anywhere.

[dachopper]

With the shielding, is that all the shielding, including spindle braid and stepper shielding together at one end on mains power earth, or just together?

[dachopper]

Hi guys,

My suspicions were correct, the culprit is EMI coming from the spindle, but now that I have replaced all of the cables...... I need some electronically minded advice.

The spindle is leaking current to the entire machine when it is powered somehow.

Inside the 2.2KW inverter, I have the house power connected to R /T on the left for active / neutral, and Ground on the far right.

Then I have the spindle power on U/V/W in the middle.

The shielding on the spindle cable is not connected to anything at the Inverter end, but it overlaps onto the metal plug at the Spindle end - could this be the cause?

The other possible issue is that this spindle does not have an earth - aside from it's power return cable.... If I did run an earth wire attached to the spindle or aluminium Z axis block, can I just plug that into the house power earth inside the inverter box, or could that cause some kind of fault/loop?

The question is, is the charge coming from the shielding getting charged, and because it's touching the outside of the metal plug, it's then charging the machine, or the lack of a spindle frame Earth is now somehow causing the issue

[Clive S]

Ok First make sure the machine frame is earthed to the star point in the control box, make sure the screen on the spindle cable is earthed at one end only (solder a wire to the screen and ground it to the spindle body)

Are all your other cable screens grounded to the star point (In the control box you should have only one point where everything is grounded to)

[dachopper]

Ok.... I disconnected the Spindle power cable braided sleeve so it is now fully insulated at both ends, and so is not physically touching the metal spindle plug either.

With the spindle turned on at 18,000 rpm, I have a 40 Volt ac signal generated, between the Spindle/Machine, and the spindle power cords' Braided Sleeve, at 9,000 hz. The machine ran without issue.

So - should I run an earth from the braided sleeve to house . power inverter earth at one end.... I don't see how this would dissipate electricity though as it's an AC signal effectively being generated by EMI.

2. I will probably run another earth wire back from the spindle to the House earth now - to stop electric shocks in the case of other failures.....

If I have the AC leaking braided sleeve earthed with the inverter / house earth, and also conect an earth from the spindle to the inverter / house earth, I could be back at square 1 as the spindle could get some of the AC?

[dachopper]

Ok First make sure the machine frame is earthed to the star point in the control box, make sure the screen on the spindle cable is earthed at one end only (solder a wire to the screen and ground it to the spindle body)

Are all your other cable screens grounded to the star point (In the control box you should have only one point where everything is grounded to)

In the control box, I have main 220Volt house power earthed to the corner of the mounting plate, The Stepper Drivers are bolted to the plate, and the Torroidal Transformer is also bolted to the plate.
Then I have a single earth wire coming off that same corner to a small ACDC transformer, and another single wire that goes off then splits to 4 and is the AC earth for my woodchip / 2 x pond pumps and Spindle power inverter.

So I should earth the spindle + Spindle sleeve to earth in the inverter? or back to star?


From DC transformer - all that suff plugs directly into the stepper drivers etc....

[phill05]

[QUOTE=dachopper;93928]



The spindle is leaking current to the entire machine when it is powered somehow.

Inside the 2.2KW inverter, I have the house power connected to R /T on the left for active / neutral, and Ground on the far right.



Why is the blue Neutral wire going to a frame screw and jumping back to the R/T side? could this be the culprit.

Phill

[dachopper]

Hi Phill, it's illusion, the neutral 240 Volt connection for this VFD is active / neutral on far left, and ground far right, then 3x spindle wires in the middle ( there is no spindle ground from manufacturer)