New experimental firmware 1.19beta10
-
I just tried tuning the heaters, my E3D V6 with 30W heater got this result:
Heater 3 model: gain 454.8, time constant 203.3, dead time 5.0, max PWM 1.00, mode: PID
Computed PID parameters for setpoint change: P15.9, I0.078, D55.9
Computed PID parameters for load change: P15.9, I1.103, D55.9pre-beta 10 result:
M307 H3 A465.8 C207.5 D7.0 S1.00 B0However it failed to tune my heat bed on first run:
Auto tune of heater 0 failed due to bad curve fit (G=59.2, tc=65.8, td=0.0)one second run it got
Heater 0 model: gain 53.4, time constant 52.2, dead time 1.0, max PWM 1.00, mode: PID
Computed PID parameters for setpoint change: P177.6, I3.406, D122.1
Computed PID parameters for load change: P177.6, I58.798, D122.1
The 2 blue peaks are from the heat bed PID tuning, look similar to me
The yellow peak is the heater after new PID tuning, look AMAZING -
Lars, thanks for trying the leadscrew adjustment. To help me diagnose this, please do the following:
1. Create a copy of bed.g as a macro. In that macro, replace S3 on the last G30 command by S-1.
2. First run the macro, then G31. Repeat until the corrections are out of range. Post the output.
I take it you mean G32 not G31…
20:40:32 M98 P0:/macros/4 - Leveling Check Probe Bed probe heights: -0.663 -0.068 -0.086, mean -0.272, deviation from mean 0.276 20:39:49 G32 Error: some computed corrections exceed configured limit of 0.50mm: -0.010 0.012 0.836 20:37:26 M98 P0:/macros/4 - Leveling Check Probe Bed probe heights: -0.665 -0.068 -0.088, mean -0.273, deviation from mean 0.277 20:36:37 G32 Leadscrew adjustments made: -0.229 0.135 0.409, points used 3, deviation before 0.197 after 0.000 20:35:37 M98 P0:/macros/4 - Leveling Check Probe Bed probe heights: -0.287 0.152 0.078, mean -0.019, deviation from mean 0.192Front left lead screw (-60,-10) is connected to Z (drive 2)
Front right lead screw (560,-10) is connected to Y (drive 1)
Back centre lead screw (250,560) is connected to X (Drive 0)Macro "4 - Leveling Check Probe":
T0 M401 G30 P0 X40 Y40 Z-99999 ; probe near a leadscrew G30 P1 X250 Y460 Z-99999 ; probe near a leadscrew G30 P2 X460 Y40 Z-99999 S-1 ; probe near a leadscrew and calibrate 3 motors M402 G90 G1 X250 Y250 T-1 ```bed.gT0
M401
G30 P0 X40 Y40 Z-99999 ; probe near a leadscrew
G30 P1 X250 Y460 Z-99999 ; probe near a leadscrew
G30 P2 X460 Y40 Z-99999 S3 ; probe near a leadscrew and calibrate 3 motors
M402
G90
G1 X250 Y250
T-1M584 Y5 X7 U8 V6 E3:4 Z0:1:2 ; Set up steppers for x, y, u, v, z and e, use P4 once it works to hide V
M671 X250:560:-60 Y560:-10:-10 S0.5 ; Set up lead screw positions for bed leveling -
Thanks Lars, that gives me something to look at. You are probing close to leadscrews 2,0,1 in that order, so we would expect the corrections to be close to the height errors for points 1,2,0 in that order. They are:
Height errors in order 1,2,0: 0.152 0.078 -0.287
Adjustments: -0.229 0.135 0.409So the corrections look way out. I'll check and simulate the maths.
Which variety of Duet do you have?
-
Thanks Lars, that gives me something to look at. You are probing close to leadscrews 2,0,1 in that order, so we would expect the corrections to be close to the height errors for points 1,2,0 in that order. They are:
Height errors in order 1,2,0: 0.152 0.078 -0.287
Adjustments: -0.229 0.135 0.409So the corrections look way out. I'll check and simulate the maths.
Which variety of Duet do you have?
Duet WiFi 1.0 + DueX5
Hi lars, is the speed problem still present in beta10? Please respond in the thread on beta10 if is is.
I do believe I know what caused the speed problem. In cura (2.6.2) there is a setting of max Z speed in mm/s. This was for some unknown reason set to 5mm/s for extruder 2. This translates a M203 Z5 (not at the start of the print but after first switch to t1) but should have been generated to M203 Z300 as on the duet the M203 command max federate is in mm/min.
If you stop the print you still got this federate when starting a new print. It might need to drop 10mm in Z with a max speed of 5mm/min at the start of the print… This should be easily solved by setting this parameter to 0 in cura, it will then not issue any M203 …
Would it make sense to revert max speed (and other configuration settings) to the original configuration after printing from file is done/stopped?
-
I think I found a bug in 1.19beta10 with my delta.
config.g contains:
M665 L265.000 R109.801 H289.779 B105.0 X-0.308 Y-0.499 Z0.000 M666 X0.281 Y-0.325 Z0.045 A0.00 B0.00After a reboot, [c]M665[/c] returns the above parameters, but [c]M666[/c] returns:
Endstop adjustments X0.00 Y0.00 Z0.00, tilt X0.00% Y0.00% ```and the bed height is incorrect when I try to print. Manually entering [c]M666 X0.281 Y-0.325 Z0.045 A0.00 B0.00[/c] after a reboot sets it correctly and the bed height is corrected when I go to print. Which reminds me… I'm curious what "good" numbers for endstop correction, tilt, and tower position correction are. I'd like some idea where my printer is on the range between perfect and "you really ought to fix that"... -
Do you have a config-override.g file, and a M501 command near the end of config.g?
Those numbers look good to me.
-
Lars, I found the problem: I missed a minus sign when transposing a formula from Maxima to the source code. So in the 3 leadscrews case it was computing the second leadscrew correction with the wrong sign.
Please try 1.19beta10+2 which I have just released.
-
Heater tuning
Old version 1.18.1
M307 H0 A311.7 C649.8 D3.1 S1.00 B0
M307 H1 A423.4 C203.6 D5.1 S1.00 B0
1.19 Beta 10
M307 H0 A240.2 C387.9 D1.1 S1.00 B0
M307 H1 A469.9 C253.0 D4.8 S1.00 B0 -
Can this new independent Z motor leveling be used for dual Z Cartesian printers like the Prusa, OrdBot, clones, etc, or is this a Delta only feature? Seems like it could be used to automatically get the X-axis of the bed mechanically leveled, only requiring the Z motors to track Y-axis movement…
-
Do you have a config-override.g file, and a M501 command near the end of config.g?
No. I took the post-calibration M665 and M666 numbers from config-override.g, put them in config.g, and deleted config-override.g. No M501 command at the end of config.g. (I tried the M501 / config-override.g methodology in the past but ultimately prefer to have it all in config.g).
Those numbers look good to me.
Good to hear!
-
Lars, I found the problem: I missed a minus sign when transposing a formula from Maxima to the source code. So in the 3 leadscrews case it was computing the second leadscrew correction with the wrong sign.
Please try 1.19beta10+2 which I have just released.
Works now!
-
I have another oddity to report with 1.19beta10:
Sometimes when a (successful) print ends, I get a "Can't delete file 0:/sys/resurrect.g" error. It's not surprising it can't delete it (because I don't have a resurrect.g file), but I never saw that error until recently (I'm pretty sure coincident with going from beta8 to beta10). I also don't know where the [c]G10 P1 S0[/c] came from - it's not in my gcode, and the only [c]G10[/c]s I have in my Duet config files are [c]G10 P0 R0 S0[/c] and [c]G10 P1 R0 S0[/c] in my config.g file when I set up my tools. Here's the console output from a print I ran last night:
11:12:00 PM G10 P1 S0 Can't delete file 0:/sys/resurrect.g Done printing file 11:11:56 PM Can't delete file 0:/sys/resurrect.g Done printing fileHere are the last lines of gcode from that print:```
G1 X0.053 Y9.146 F540
G1 E-8.0000 F6000
; layer end
T0 ; Select Extruder 0 if not already selected
G90 ; Switch to absolute mode
G1 X0 Y0 Z250 E-10 F6000 ; Retract filament and raise printhead
M104 S0 ; Turn off heater
M84 S10 ; turn off motors after 10 secondsBTW, do you prefer feedback like this in this forum or as an issue in github? -
Thanks, Yonkiman. I prefer issues reported here, especially for betas.
Regarding G1 P1 S0, do you have a stop.g file, and if so, is it in that?
I'll suppress the "Can't delete" message in the next beta.
-
Chrishamm has released DWC 1.17+2 which includes the changes needed to work with the new height map. So will those of you running firmware 1.19beta10 please upgrade it to 1.19beta10+2 if you haven't already, and upgrade DWC to 1.17+2. The mesh bed compensation height map should then display correctly.
-
Regarding G1 P1 S0, do you have a stop.g file, and if so, is it in that
Yes, I have the file; no it is not in that file. File only has the default 2 comment lines. I actually did a search for the word "G10" on my entire SD card, and it only appeared in the tool setup in config.g (as reported above) and one or two binary files.
Speaking of which…once we successfully flash the firmware, is it safe to remove all the non-text (i.e. .bin/.zip/.rar) files from our sys directory?
-
Speaking of which…once we successfully flash the firmware, is it safe to remove all the non-text (i.e. .bin/.zip/.rar) files from our sys directory?
Yes, although the only binary files that should be in there are the two extra .bin files for the Duet WiFi.
-
Hey David, did you see my question nested up a few posts about dual-Z motor cartesian printers?
-
Can this new independent Z motor leveling be used for dual Z Cartesian printers like the Prusa, OrdBot, clones, etc, or is this a Delta only feature? Seems like it could be used to automatically get the X-axis of the bed mechanically leveled, only requiring the Z motors to track Y-axis movement…
Yes, it's for Cartesian and CoreXY printers, not deltas.. See https://duet3d.com/wiki/Bed_levelling_using_multiple_independent_Z_motors.
It can also be used to tell you what adjustments to make to manual bed levelling screws, but I haven't documented how to do that yet.
-
Did some more testing for the leveling. There might be some other bug. Feels like its still adjusting in the wrong direction…
I changed the probe order to match the reporting of G32 and moved the points a bit.16:55:14 M98 P0:/macros/4 - Leveling Check Probe Bed probe heights: -0.136 -0.441 -0.471, mean -0.349, deviation from mean 0.151 16:54:26 G32 Error: some computed corrections exceed configured limit of 0.50mm: -0.007 0.505 0.553 16:52:50 M98 P0:/macros/4 - Leveling Check Probe Bed probe heights: -0.138 -0.439 -0.461, mean -0.346, deviation from mean 0.147 16:52:09 G32 Leadscrew adjustments made: -0.294 0.260 0.269, points used 3, deviation before 0.170 after 0.000 16:51:36 M98 P0:/macros/4 - Leveling Check Probe Bed probe heights: 0.186 -0.203 -0.174, mean -0.064, deviation from mean 0.177Front left lead screw (-60,-10) is connected to Z (drive 2)
Front right lead screw (560,-10) is connected to Y (drive 1)
Back centre lead screw (250,560) is connected to X (Drive 0)Macro "4 - Leveling Check Probe":
T0 M401 G30 P0 X250 Y420 Z-99999 ; probe near a leadscrew G30 P1 X420 Y80 Z-99999 ; probe near a leadscrew and calibrate 3 motors G30 P2 X80 Y80 Z-99999 S-1 ; probe near a leadscrew M402 G90 G1 X250 Y250 T-1 ```bed.g:T0
M401
G30 P0 X250 Y420 Z-99999 ; probe near a leadscrew
G30 P1 X420 Y80 Z-99999 ; probe near a leadscrew and calibrate 3 motors
G30 P2 X80 Y80 Z-99999 S3 ; probe near a leadscrew
M402
G90
G1 Y250
T-1M584 Y5 X7 U8 V6 E3:4 Z0:1:2 ; Set up steppers for x, y, u, v, z and e, use P4 once it works to hide V
M671 X250:560:-60 Y560:-10:-10 S0.5 ; Set up lead screw positions for bed leveling -
Lars, thanks for that. The corrections it wants to make now look correct - they are similar to the height errors but with opposite sign and somewhat larger magnitude - which is expected because you are probing well inside the leadscrew positions.
As you say, it looks like the corrections may be getting applied with the wrong sign, at least for the 2nd and 3rd screws. I'll simulate your figures on my Cartesian printer (which has only 1 Z motor) and see whether there is a problem with positive corrections.
EDIT: I found the problem. When it is setting the motor direction, it sets the directions of all Z motors together instead of individually. I will fix this shortly as I am keen to get this feature working properly.
