Multiple issues with Duet Ethernet since 2 weeks

[V3DPrinting]

@v3dprinting said in Multiple issues with Duet Ethernet since 2 weeks:

Definitely it’s linked with temperature, but not consistent. It seems that the print done yesterday evening is good from the webcam view. Need to check when arriving in the office.

The part done yesterday evening is better, but not good : the first millimeters shows the same pattern with bigger layers, but less than yesterday.
So the part is only 10,75 mm instead of 10mm.

The problem is linked with some electronic component not delivering properly when cold or hot. Impacts the drivers.

[V3DPrinting]

@v3dprinting said in Multiple issues with Duet Ethernet since 2 weeks:

Forgot to say that the PSU is a 24V 400w or 450w, with dual rails. Bed is on a dedicated rail.
Load is only 50% max.
I’ve monitored the voltage : bet. 24.1 and 24.4V on the Duet sensor.

PSU is a TDK-Lambda GWS—500—24. Delivery 500w under 24V, max current 21A.
Definitely not overloaded.
As the voltage is pretty steady, no sign of issues on that side.

[dc42]

From your config.g:

M906 X1100 Y1100 Z850 E600:600 I30 ; Set motor currents (mA) and motor idle factor in per cent

• How did you choose the XYZ motor currents?
• Do you have the specifications of the motors?
• How may Z motors do you have? If more than one, how have you connected them?

M92 X160.47 Y160.11 Z1600 E505.7:505.7 ; Set steps per mm
M566 X360 Y360 Z24 E1200:1200 ; Set maximum instantaneous speed changes (mm/min)
M203 X18000 Y18000 Z360 E2400:2400 ; Set maximum speeds (mm/min)
M201 X1500 Y1500 Z300 E3000:3000 ; Set accelerations (mm/s^2)

For an axis with 1600 steps/mm, the maximum speed and jerk look OK, but the acceleration looks rather high. Try reducing Z acceleration to 100 in the M201 command.

[V3DPrinting]

Hi Dave,

XYZ motor current was chosen bet. 50 and 80 % of the max load of the steppers. XY are 1,8A and Z is 1,2A, 400 steps/rev

I only have one Z motor.

Remember that this machine had been running perfectly during nearly 1 year and half with these parameters.

I will try to reduce the Z acceleration, but I don't think it is the point, as the problem appears only in the first millimeters on Z axis.

And don't forget that I also have the wobbling pattern issue on X and Y too, even if I don't reproduce it at the moment.

[V3DPrinting]

I have done another test print today with the same gcode as yesterday evening.

Same issues on Z, except the part is now 11,8 mm height instead of 10 mm.

First 4,8mm of the part are with layers well over 300 microns instead of 200 microns.

As it is the same gcode, the issue is definitely on the printer.

[deckingman]

@v3dprinting I have an idea but it's a log shot. Before I say any more, do you use Z hop at all?

[V3DPrinting]

@deckingman

Very very scarcely. Only when the part is a bit tricky.
But most of the time, I need also dual extrusion, so I end to print them on my Ultimaker 3 and S5

[V3DPrinting]

@dc42

I have changed the Z acceleration to 100
I have printed a new test print, still the square. But with another end of spool.
Still the same issue.

This time the part was 10,8 mm height instead of 10mm

I haven't switched off the printer. Just a reboot in order to take the new config.g.

So it is not related to any skipped steps nor mechanical issue.

I am still unable to reproduce the wobbling effect during a print.

[V3DPrinting]

I have done a manual bed leveling, presuming the mesh bed leveling was having issues due to the fact that I'm using a 3 mm glass instead of 4 mm (broken the last one this weekend ). So an additional negative value to compensate again the homing position.

It hasn't changed anything and the first layers are still more than the 200 layer height defined.

I am now printing the same pan as at the beginning of the thread, trying to reproduce the X and Y wobbling. But I print it in PLA.

[V3DPrinting]

I have printed yesterday evening the pan in PLA as said.

Print was failed, layer height was inconsistent :
First layers were too big (more than 300 microns) as in previous tests, instead of 150 microns
Then layers were too small or null, so I had a nozzle jam, filament grinding ...

But the print ended without errors and the total height reported by the printer was good or so (I have a +0,5 mm Z end script) for a 32mm height model.

I have checked the gcode file and layer height is 150 microns

https://www.dropbox.com/s/vha6z0k1ch076en/sauce PLA.factory?dl=0
https://www.dropbox.com/s/0gtc76kp0i0ef3q/sauce PLA.gcode?dl=0

The printer has been switch on since two days now (except 15min today to clean the extruder).

I will make a reprint, but I fear I need to replace the board.

[deckingman]

@v3dprinting A couple of suggestions

1. I still think there could be a mechanical issue such as a loose pulley on the motor shaft. When you use level compensation during a print, maybe the bed moves down where it needs to but does not move back up again. That would explain the higher layers. I don't use any form of flatness or level compensation myself so I can't be sure but I believe that it tapers off. If that is the case, it would explain why the issue corrects itself when you get further into the print.

If you are absolutely sure the pulleys are all tight and there are no mechanical issue then:

2. Try the Z motor on a different driver such as a spare extruder drive if you have one. This will tell if there is a fault with the stepper driver.

3. Try a different motor if you can - maybe swap for one of the other axes. If the probelm with Z is cured but you get new problems on the axis you swapped motors with, then it's a motor issue.

[V3DPrinting]

@deckingman
I will do the checks suggested.

I don't think of mechanical issues as the first layer is good, when the head is not crashing into the bed.

If I have a loose pulley or belt, then I won't crash the head, as the build plate has to be raised, it exerts an effort on the nut, leadscrew, pulley and stepper, so the play will be compensated and maximum.
As a result the build plate would be lower than expected.
This is true when homing, doing the mesh leveling procedure or printing.

Also, at the beginning of the problems, Z axis was working fine, but X and Y moves were shaky, creating wobbling patterns on long straight walls (with no acceleration).

That said, the problem seems to be more global to the electronic and linked to the temperature of some components.

[deckingman]

@v3dprinting It seems to me that from the outset you have made up your mind that it's a faulty board. You might be right but it's a very unusual Duet problem. I don't mean this in a bad way but I think you ought to be a bit more open minded. I bought one of the very first gen 2 boards and have been visiting these forums daily ever since. 99% of similar issues have turned out to be mechanical or some other non-Duet related problem. You could of course be one of those 1%.

Anyway, I think you should check the pulley screws and so forth, rather than dismiss any mechanical issues out of hand. Then try swapping the driver and motor as I suggested.

Also, if you think it is temperature related, try rigging up a fan to blow cold air over the board. That might help to prove it or otherwise.

[V3DPrinting]

@deckingman said in Multiple issues with Duet Ethernet since 2 weeks:

Also, if you think it is temperature related, try rigging up a fan to blow cold air over the board. That might help to prove it or otherwise.

I already have one from the beginning with this board.
And on top I use the PSU fans to extract the hot air from the case and the PSU outside of the case with a custom duct.

As said I will check the belt and pulleys for Z axis, do the tests on drivers and steppers.

But I already done it for X and Y axis, without any finding and I still don't explain why I had some wobbling for the first symptoms and not anymore.

Other finding is, during the last two failed prints, I had grinding issues (with Bond tech QR 3.0), but the nozzle wasn't clogged and there was some room below the nozzle due to the inaccurate z layer height.

That's why I keep saying there is an electronic problem.

But don't be mistaken, the Duet is a great board and I am very happy with it.

[deckingman]

@v3dprinting said in Multiple issues with Duet Ethernet since 2 weeks:

Other finding is, during the last two failed prints, I had grinding issues (with Bond tech QR 3.0), but the nozzle wasn't clogged and there was some room below the nozzle due to the inaccurate z layer height.

That's why I keep saying there is an electronic problem.

The extruder is grinding the filament and you think that is an electronic problem?

[dc42]

How stable is your power supply voltage? The M122 report will tell you the minimum and maximum voltage measured since power up or last time M122 was run.

[V3DPrinting]

@deckingman said in Multiple issues with Duet Ethernet since 2 weeks:

The extruder is grinding the filament and you think that is an electronic problem?

Yes I do.
The bondtech extruders hardly grind filament. Nozzle wasn’t clogged at all and settings were correct. At least with the temperature defined when calibrated.

I suspect the feed rate not to be accurate and pushing too much filament. Exactly like with the Z axis making layers more than the one defined during the first 3 or 4 mm.

[V3DPrinting]

@dc42 said in Multiple issues with Duet Ethernet since 2 weeks:

How stable is your power supply voltage? The M122 report will tell you the minimum and maximum voltage measured since power up or last time M122 was run.

This is the extract of the M122 done on Dec 6

Supply voltage: min 24.1, current 24.3, max 24.4, under voltage events: 0, over voltage events: 0

So it seems the PSU is working fine

[V3DPrinting]

@deckingman said in Multiple issues with Duet Ethernet since 2 weeks:

@v3dprinting A couple of suggestions

1. I still think there could be a mechanical issue such as a loose pulley on the motor shaft. When you use level compensation during a print, maybe the bed moves down where it needs to but does not move back up again. That would explain the higher layers. I don't use any form of flatness or level compensation myself so I can't be sure but I believe that it tapers off. If that is the case, it would explain why the issue corrects itself when you get further into the print.

If you are absolutely sure the pulleys are all tight and there are no mechanical issue then:

I have checked all the Z axis motion and bearing : pulleys were properly tight (I remembered having fixed them with Loctite), belt was properly tensioned, no noticeable play in the leadscrew.

2. Try the Z motor on a different driver such as a spare extruder drive if you have one. This will tell if there is a fault with the stepper driver.

I have inverted Y and Z as Y was good and Z having problem generating the right layer height.
I have, of course, put the right M584 X0 Y2 Z1 E3:4 command, to remap the drivers.

3. Try a different motor if you can - maybe swap for one of the other axes. If the probelm with Z is cured but you get new problems on the axis you swapped motors with, then it's a motor issue.

Not a motor stepper issue.

I have printed another square 100 * 100, 10*10 section.
The print is very good from a dimensional point of view : from the previous tests I should have expected issues on Y and having Z fine, but the result is the part is within the usual tolerances (+0, +0,2mm).

The only issue is I have some wobbling on both X and Y axis. See pictures.

It might be of importance, but the printer had been swithed off since Dec 7th evening, and just powered up to load filament and start print.

So I confirm the issue is linked to the board.

As the printer had been standing idle 16hours, I will make a reprint to validate the issues are linked to the fact the printer is switched on.

The PSU is steady with no significant deviation since powered up 17 hours ago
See the M122 output done right now

10:42:20M122
=== Diagnostics ===
Used output buffers: 3 of 32 (11 max)
=== Platform ===
RepRapFirmware for Duet 2 WiFi/Ethernet version 1.21 running on Duet Ethernet 1.02 or later
Board ID: 08DDM-9FAMU-JW4S4-6JKD6-3SS6L-12Z3U
Static ram used: 16152
Dynamic ram used: 100464
Recycled dynamic ram: 2168
Stack ram used: 1224 current, 6472 maximum
Never used ram: 5816
Last reset 18:14:54 ago, cause: software
Last software reset at 2018-12-12 16:27, reason: User, spinning module GCodes, available RAM 7760 bytes (slot 3)
Software reset code 0x0003 HFSR 0x00000000, CFSR 0x00000000, ICSR 0x0441f000, BFAR 0xe000ed38, SP 0xffffffff
Error status: 0
Free file entries: 10
SD card 0 detected, interface speed: 20.0MBytes/sec
SD card longest block write time: 89.1ms
MCU temperature: min 29.0, current 31.2, max 35.8
Supply voltage: min 24.1, current 24.3, max 24.4, under voltage events: 0, over voltage events: 0
Driver 0: standstill, SG min/max 0/1023
Driver 1: standstill, SG min/max 0/1023
Driver 2: standstill, SG min/max 0/1023
Driver 3: standstill, SG min/max 0/1023
Driver 4: standstill, SG min/max not available
Date/time: 2018-12-13 10:42:19
Slowest main loop (seconds): 0.159806; fastest: 0.000050
=== Move ===
MaxReps: 4, StepErrors: 0, LaErrors: 0, FreeDm: 240, MinFreeDm 151, MaxWait: 1206640321ms, Underruns: 0, 0
Scheduled moves: 2, completed moves: 2
Bed compensation in use: mesh
Bed probe heights: 0.000 0.000 0.000 0.000 0.000
=== Heat ===
Bed heaters = 0 -1 -1 -1, chamberHeaters = -1 -1
Heater 0 is on, I-accum = 0.0
Heater 1 is on, I-accum = 0.4
=== GCodes ===
Segments left: 0
Stack records: 1 allocated, 0 in use
Movement lock held by null
http is idle in state(s) 0
telnet is idle in state(s) 0
file is idle in state(s) 0
serial is idle in state(s) 0
aux is idle in state(s) 0
daemon is idle in state(s) 0
queue is idle in state(s) 0
autopause is idle in state(s) 0
Code queue is empty.
=== Network ===
Responder states: HTTP(1) HTTP(0) HTTP(0) HTTP(0) FTP(0) Telnet(0) Telnet(0)
HTTP sessions: 1 of 8
=== Network ===
State: 5
HTTP sessions: 1 of 8
=== Expansion ===

[dc42]

I see that you are still running firmware 1.21, which has been superseded first 2.0 and the 2.01, with 2.02 in the late Release candidate stage.

The patterning in your print looks like a motion or extruder issue to me. If it only happens after the printer has been running for some time, then it's probably temperature-sensitive. Does the patterning happen along both arms of the L, or only one?