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

Hi guys,

Well, it's still there.... After changing from Gecko dive to AM882 drivers, I am still getting random X and Z this time, faults, as in the tool has missed 5 to 6 cm of travel in an instant.

Ran for 55 minutes doing a constant Z level face, with long straight lines, no errors.

Changed to a parallel operation, with constant Z changing, constant Y position every 3 seconds then move 1mm, and constant X reversing left to right while Z up and down.

The error that occurred within 5 minutes of the new operation starting, ended up shifting my X position by 70mm, AND shifting my Z up by 3mm - no fault detected by AM882 because speed too low for stall detection.... or because fault was not a stall.

Thought that the X/Z error was caused by a jam and motor loosing torque after overheating ( which still may have something to do with it ) So I started Parallel again while motor was boiling hot and pushed with about 40kgs force and was utterly unable to cause any kind of jam, let alone 70mm of jam.

Re-homed position, reduced speed by 50% and ran operation for 25 minutes , and through the previous anomily without error ( not a G-Code error ).

So now I'm thinking I have only ever had this error when running parallel operations with changing Z values..... Could there be something happening momentarily with the motion controller getting overloaded from all the commands and it be responsible?

Assuming this is not a lack of stepper power, or a machine jam related error then what else could cause it to jam at higher speed, but seemingly not at 50% speed?

Nick

Quote:

---

Originally Posted by dachopper

motion controller...

---

Good question, what motion controller are you using? Do you have a USB breakout board in play?

Quote:

---

Originally Posted by dachopper

what else could cause it to jam...

---

Good question, are you 100% positive everything is mechanically sound with the machine?

UC400 ETH controller, UCCNC software

I don't know, I can turn the ball screws by hand from end to end..... what more could cause the X axis to seemingly bind, not get noticed by the 822, and then immediately unbind in reverse direction?

Next step is GOPRO an entire cut in the air, and see if it jams unloaded under inertia or no jam and program telling it to move?

Have you checked the pulse width from the UC400 ETH?, I don't know how but I've noticed its an option in the UCCNC software. The AM882 needs >2.5us for reliable operation. Worth a look?

Does the power supply have enough output!

Peter

Quote:

---

Originally Posted by Doddy

Have you checked the pulse width from the UC400 ETH?, I don't know how but I've noticed its an option in the UCCNC software. The AM882 needs >2.5us for reliable operation. Worth a look?

---

Pulse width is not an option. It's a function of the kernel frequency. I believe that you need to use 200Khz or less to get 2.5us pulse width.

Using 50 khz , .1 sec com buffer, and 300 line look ahead

Driver PSU voltage: ???
Micro stepping: ???
Stepper current: ???
Speed and acceleration: ???

The AM882 seems to be pretty complex compared with the traditional analogue drivers, are you sure it is been configured right, according to the manual?
Have you checked the signal levels at the AM882 end? How about wiring?

Quote:

---

Originally Posted by dachopper

Using 50 khz , .1 sec com buffer, and 300 line look ahead

---

No benefit in using 50kHz and it is too low if you have a fast machine. Raise it to 200kHz.

What's the benefit of 200Khz?
50Khz gives you a longer pulse width, which can be a benefit with some drives.

The fact that he's had this issue with both Geckos and Leadshines, means it's probably not the drives.

That would point to maybe the power supply, an issue with accel/velocity settings, or a mechanical issue.

You said the motors are very hot? What are the motor specs, and what do you have the drive current set at?