Duet wifi stepper noise in signal
-
@ret said in Duet wifi stepper noise in signal:
[...] I copied the exact same values used in the gcode I posed, just converted them from reprap firmware to marlin, and converted the units where needed. [...]
Jerk does not work the same way as in marlin, afaik. IIRC, you essentially have to use half the jerk value in RRF than you do in marlin. This is because, and my memory might be failing me here, marlin uses half the jerk value as the start/stop speed, and the jerk value itself is used for cornering speeds or something.
Anyway, it's obvious that copying and pasting the settings isn't working, so you should start trying different values to see what changes things.
-
Is there a Makergear user group somewhere that you could check in with and see if others are having similar issues using the Duet?
-
@bot I'm aware its different, just not by what degree. Please re-read my post as ive not just copied and pasted and started asking for help. I've gone through many many many different configurations with the jerk and accel from really low to super high with no significant quality improvements. Please don't get me wrong here, i do appreciate everyones help on this, i'm just very frustrated with this issue. If i in fact need to reduce my jerk by half, ill be running at around X120 Y120 Z15 E30 which in my opinion, as of what everyone else settings are, is on the very low end of things. I have in fact tried half of my original settings from marlin with very very jerky slow prints, which makes sense.
@Phaedrux I've sent messages out to a few people who've put the duet on their printer and their settings near match mine and the marlins default values. They both say they love their duet, which makes me wonder if I've either got a faulty board, or am having unrealistic expectations for the duet.
-
Are you positive that your maximum speeds and accelerations are the same as for marlin? Or maybe marlin doesn't coordinate the E axis well with the X/Y/Z axes, and so you have to increase your accel/jerk for the E extruder a lot, without adjusting the X/Y values.
If you want, I'll lend you a Duet WiFi to see if you have a faulty board. Just pay for postage and promise to send it back.
-
@bot Thank you so much for that kind offer, but I couldn't possibly take one of your printers offline for that.
On a side note-
Pressure advance makes my top layers worse in tight back and forth top fill, and doesn't make my corners any less rounded/bulged. But, it does make my extruder skip a lot if over 0.25, which makes sense. Any idea why that could be, maybe tied into the underlying issue? Printing without Pressure advance improves my quality on top and corners, but still nowhere near what I'd consider "good"Besides that, do you have any other suggestions to try?
-
@ret I actually have a couple spare wifis that are not being used, so it wouldn't take a printer out of service.
I can't think off-hand anything worth trying to improve things. Just keep coming up with theories and testing them, eventually you'll get it! haha... that's my strategy.
-
Not sure if it’s been said but when I had issues like this I noticed that the config override file was not set/being used like I thought. It’s worth making sure you either have the right mcodes in place to run the override file or intentionally not use it.
-
@ret said in 0_1556294322048_config.g
M587 @ ; Configure access point. You can delete this line once connected
As the comment says remove this from you configuration file once you have successfully connected your printer to the wifi.
@ret said in 0_1556294322048_config.g
M350 X16 Y16 Z16 E16 I1
I can see a few different settings but it doesn't appear that you have tried 256?
@ret said in 0_1556294322048_config.g
M566 X240 Y240 Z30 E120 ; Set jerk (maximum instantaneous speed changes) (mm/min)
This is conservative as others have said. Fair enough you may have tried many different settings but an iterative correction process may settle on a completely wrong value with values this far off. I've a comparatively flexible Ormerod running on:
M566 X900 Y900 Z12 E800
...and another with a reprap 3:1 geared extruder running at E3600. This firmware (as with most) will slow the moves to fit the acceleration and instant speed of the slowest axis.
I can however see that you have copied the values from Marlin and multiplied by 60.
@ret said in 0_1556294322048_config.g
M201 X900 Y1000 Z30 E2000
Acceleration not far off my Ormerod settings. If you're running a geared extruder that isn't struggling try E3600.
@ret said in 0_1556294322048_config.g
M906 X1400 Y1000 Z750 E1000 I30
Lots of different motors on your system? Others have reported surface finish artefacts when running the motors with a current limit significantly lower than the motors rated current. Set the motors to 80% (unless the values above are already higher than this) of the rated current an feel how warm they get.
The marlin firmware seems to have the motors limited to 0.75A? Are they high inductance motors? I had an issues on an Up printer where the machine would run really rough of the current limit was too high. It was basically the resistance of the coils that limited the current not the chopper driver and it didn't seem to like that!
@ret said in 0_1556294322048_config.g
; Axis Limits
M208 X0 Y0 Z-5 S1 ; Set axis minima
M208 X220 Y255 Z203 S0 ; Set axis maximaThis won't help you at all but I liked one of the recent changes that allowed the above to be condensed into one line as follows. I find it much more readable.
M208 X0:220 Y0:255 Z-5:203
@ret said in 0_1556294322048_config.g
M570 H0 S1200 ;heater faults for 1200 seconds
M570 H1 S1200I'm suspecting something is seriously amiss with your heaters if you have the need to have a heater fault existing for 20 minutes before the system shuts down?
Others have raised the concern about config override but there is no M501 in your configuration file, so that shouldn't be being read in.
Are your PID values from a Duet auto tune or also copied from the Marlin config?
-
Ok, just seen the extruder is driven by a 5.18 ratio gear box, and counted the teeth on my extruder and it is a 3.09 ratio. There for comparative instant speed settings would be between 480 and 2150. Comparative accelerations values between 600 and 2150.
In a nut shell acceleration on the E axis is fine, but perhaps room for instant speed increase. It shouldn't be terrible as is.
-
@ret
What is the name/type of your power-supply, and is it a switching one? (If you do not have it already you could throw a high-end power-supply at it like a "daitron" sps or a toroidial one with additional capacitor-smoothing-banks... Just maybe something to think about? On the other hand, since the trinamic-driver are anyway switching themselves while measuring, they should be able to also cope with a less perfect input... hm...) Just writing this because I had problems with a cheaper one in the past, but I also made a build for somebody where because of the size the cables where very long ca. 3.5meters and I had a lot of emi-problems in the beginning... but if in your case it works fine with one board and not so fine with the other it is for sure the config of those stepper-drives, a lot of values there at least for a beginner like me for those trinamic-drivers
-
@bot thanks for the encouraging words
@Alex-cr My override is empty, so its not used. Thank you for sharing though!
@DocTrucker I have indeed tried 256 and it doesn't make quality better or worse unless i start printing fast and then my extruder skips. I have tried extruder jerk form 1000-9999 in thousand increments. So far, values under 1000 prove to be the best. I don't believe my E is the limiting factor here since print times or motion doesn't seem to be different from 60 jerk up. I bumped up my accel to 3600 with no difference in quality. Every stepper is different, yes. I have tried every stepper current in incements of 100mA up from 750mA to 1A and found them to be most quiet and effective where they're at. I'm not sure their inductance. As for the heater faults, I simply threw that in there because its the values makergear supplied with their U1 running a duet and it runs the same hotend. PID's are from the auto tune. They're really not any better than the stock values, but I left them.
@LB My printers isn't running a very cheap, nor the best PSU. I do however believe its very adequate(Mean well SE-450-24). When running, it never drops below 23.8V(24V PSU), even under the most strenuous conditions. Vripple is also negligible .
-
@ret said in Duet wifi stepper noise in signal:
I have tried every stepper current in incements of 100mA up from 750mA to 1A and found them to be most quiet and effective where they're at. I'm not sure their inductance.
...and the 1.4A you are running the X axis at. Do you not know the specification of your motors or have you forgotten?
-
Also ensure the chassis is ground to the protective earth and there is continuity between the steppers and earth.
I didn't doubt that you have tried many different settings, just highlighting that aspects of your current setup are a way from average, for a very average looking cartesian machine.
There options here are limited:
- You've missed the sweet spot.
- Motors are odd balls that work well with the RAMBO board but not the Duet - but we know nothing of motor spec.
- Wiring and or electrical noise issues. Duet may be susceptible to specific emc noise.
- Minor board fault, although I've not been on these forums enough to see another case where a board fault has caused surface artefacts rather than catastrophic inability to print.
-
I'd add motor swaps to known specification units to the list for fault finding.
-
And run prints without:
- pressure advance.
- bed level correction.
Not to do so permanently, just to see if it has an effect.
-
Sorry, forgot to talk about the X. That's a 2A motor that seems to get real noisy and hot over 1.5A, so I stepped it down to 1.4 where it seems to run its best.
Here are the motors:
X: MS17HA4P4200
Y: MS17HD2P4100
Z: e43h4b-2.33-a20 similar to the link
E: 1040252It looks like all of the steppers are all under 4mH, except Y which is 8.9. I have read this writeup and they suggest against that high of an inductance. I have put the original Y motor back on it which is 1124141 with 4.8mH, no difference in quality besides more ringing. The original XY steppers had a lot of vibration and noise, so i swapped them out. I print mostly without pressure advance since it makes quality worse and have never used bed level correction. I've gone through and verified every stepper is grounded to the chassis, which is grounded. I did happen to come across something odd that I've never tried before. I attached a probe from the ground on the duet, to the chassis of the power supply which is grounded . Put my meter in ACV as a cheap ripple/noise meter and I get a constant ~3.5V. I then put it in DC and get about 0.5-1V. THroughout different places on the printer probing, I get different voltages. This could all be normal and mean nothing, just throwing it out there.I'm no electrical engineer, just a hobbiest with electronics, so I don't know exactly what this could mean, if anything.
-
Fair enough, that's off the list then!
Regards grounding most switch mode PSUs have a capacative link between the DC negative and protective earth. I've found my ethernet connected v0.6 and v0.8.5 units don't connect well to the ethernet when there is no link between protective earth and dc negative. In an ideal world I would pull my finger out and build the circuit recommended by another forum user that creates a bit of a seperation between protective earth and dc earth but lets voltages through that are over a threshold.
Edit: Getting decent low resistance earths throughout a machine can be tricky. Especially so on an alloy extrustion machine but think yours is assembled/pressed steel? Tho be honest if the motor bodies are grounded I doubt it is worth spending much more effort on it. That said some people have grounded the hotends but think that was to ease temerature reading noise. I'm also assuming that grounding your extruder drive stepper will ground you extruder. Similar to before though, I doubt it will be the magic bullet.
