Duet Maestro Freefall

dc42

I'm glad that worked. I am preparing a new release that switches from stealthchop to spread cycle at higher speeds, which may be part of a better solution.

dc42

PS - one way of reducing idle motor noise on the Maestro in spread cycle mode may be to use the M917 command to reduce the motor standstill current from its default of 75%, e.g. M917 Z40.

deckingman

@sigxcpu said in Duet Maestro Freefall:

Or you can wait to get old. My printer is almost quiet, I just need few more years

Like it (and I know the feeling).

dc42

Thise who have experienced this issue, please see https://forum.duet3d.com/topic/6614/new-duet-maestro-firmware.

LeonMF

@dc42 Closing the loop on this. I haven't had time to do any real significant testing but here's what I've got.

Disabling stealthchop does solve the problem.
This makes way worse stepper noise than the older drivers on Duet 2 ( I think? Maybe It's a system difference between this and my Delta)
Tuning stepper noise with M569 P2 F# changes noise but not necessarily for the better.
The new 2.0RC1 does not inherently solve the bed crash problem because the motor is in stealthchop when doing the types of move (slowly approaching a nozzle for calibration) most likely to result in exceeding the motor torque causing the freefall bed crash.
Given the above the video of the fix demonstrated by M3D on their discord channel doesn't make sense unless they are running the Z axis into the nozzle past the stealthchop/spreadcycle transition and thus demonstrating a use case not normally seen in operation.

Would substituting a Duet 2 Ethernet in place solve this problem?

dc42

Here are some options:

Switch the Z motor only to spreadCycle.
Increase the M569 V parameter sufficiently so that the corresponding mm/sec value displayed by M569 P2 is below the Z movement speed that triggers the problem.
Does the machine have two Z motors? If so, connecting them in parallel (instead of in series, which is what you get if you plug them into Za and Zb) and using just one of the Z motor connectors may solve the original problem and/or make spreadCycle quieter. Don't forget to double the motor current setting.

The Duet Ethernet doesn't do stealthchop mode, so it would be in spreadCycle mode all the time, with the same noise issue if you are unable to find settings that silence it.

LeonMF

@dc42 There is only one Z motor and it's on a novel belt system. This has low friction and allows the motor to freewheel fast enough that the motor doesn't catch itself when it drops.

Option 1 does solve the problem. Option 2 would do so if we were moving faster than the transition to spreadcycle. However, the most likely issues occur when the motion is very slow as you are inching up to the nozzle. That means that you'll always be in stealthchop in this situation.

For now, I've added code into the calibration scripts to enable spreadcycle during the risky moves and re-enable stealthchop during normal operation. It's better than nothing since what I'm interpreting here is that this can't be fixed and have stealthchop enabled.

jwmueller

@dc42 any thoughts or updates on this?

dc42

@jwmueller said in Duet Maestro Freefall:

@dc42 any thoughts or updates on this?

You may still be able to use stealthchop when the motors are idle, if you set a very low speed threshold/high M569 V parameter for the Z motor. But that may make the Z motor too noisy when moving.

napierdmark

@leonmf Can you add a friction break? A solenoid that retracts a brake when the table is free to move but automatically re-engages when power is lost (or the driver is not enabled).