Slowdowns / Pauses (PID algorithm and PWM freq.?...)

[deckingman]

@Argo There is something wrong there somewhere. Is your bed heater 12/24V and if so, is your power supply up to the job? 10Hz for a bed heater is usually fine and if you use an SSR, it is preferable.

[Argo]

@deckingman

I've tested it with a Delta 250w and Meanwell 350w PSU. Both are 24v. The bed heater is also 24 volts.

Printer is (was) a Prusa Mk3 printer.

After I discovered that the slowdowns are related to the bed heater, I disabled PID, enabled Bang-Bang and the issue was gone. Then I remembered that I lowered my PWM frequency from 100hz to 10hz which started the issue. So I set it back to 100hz, re-enabled my PID tuning and the problem was gone.

I don't see what else could cause this?

[deckingman]

@Argo I think it must in some way be related to you having a 24v heater (although I have no idea why). I've always used 10Hz and still do - in fact until very recently that was the default frequency for bed heaters because they are often used with SSRs. But then, for the last "n" years, I've always used mains powered bed heaters.

[Argo]

@deckingman

Are you already using the new PID tuning? I think it was implemented with RRF 3.2

[deckingman]

@Argo said in Slowdowns with new PID algorithm and low PWM freq.:

@deckingman

Are you already using the new PID tuning? I think it was implemented with RRF 3.2

PID tuning has been around since Adam was a boy The algorithm changed in RRF3.2 but I've been using it for years. Having said that, I have not PID tuned the bed heater with the latest algorithm.

EDIT. I lied - just checked my config.g file and I see this

M307 H0 R0.104 C2186.5 D28.16 S1.00 B0 V0; auto tune 08/01/2021 using RRF 3.2. 

[Argo]

@deckingman

Yes, that's the new PID
The old PID didn't have the "R" value if I remember correctly.

"V0" You are using a SSR? Then maybe your other assumption is right that it may have something to do with whether is a 24v (12v?) heater or not. But that's also only a assumption on my side.

[deckingman]

@Argo said in Slowdowns with new PID algorithm and low PWM freq.:

............ The old PID didn't have the "R" value if I remember correctly.

That's correct. Being a sad old git, I always comment out the old PID values when I enter a new one. That way one can revert back easily if something doesn't work as expected.

So I also have ......

;M307 H0 A230.8 C2252.5 D20.3 S1.00 B0 V27.3 ; auto tune 06/08/2019

EDIT. In the above case, the "V" was irrelevant because it's a mains heater with SSR.

[Argo]

I did some research and found Prusa dev blog article about the topic:

"Prusa Dev Diaries #2: How we made the heatbed silent:"
https://blog.prusaprinters.org/dev-diary-2-how-we-made-the-heatbed-silent_30946/

I don't know though what is transferable to Duet hardware.
They are also describing slow downs with specific frequencies.

[Argo]

The slowdown are still there.

After changing the PWM frequency to 100hz I could probe without slowdowns but during a print they still happen. Then I removed the "Q" flag as my old Duet 2 config also didn't use it. But it still happens...

https://www.youtube.com/watch?v=1WkCpfci4O4
(don't mind the sewing machine in the background)
The slowdowns always happen at corners. For the next test I'll disable pressure advance.

M122:

m122
=== Diagnostics ===
RepRapFirmware for Duet 3 Mini 5+ version 3.3beta2 running on Duet 3 Mini5plus Ethernet (standalone mode)
Board ID: ZDWF8-8296U-D65J0-40KMU-J603Z-77MM8
Used output buffers: 1 of 40 (13 max)
=== RTOS ===
Static ram: 99720
Dynamic ram: 101012 of which 200 recycled
Never used RAM 39916, free system stack 110 words
Tasks: NETWORK(ready,225) ETHERNET(notifyWait,569) HEAT(delaying,299) CanReceiv(notifyWait,943) CanSender(notifyWait,358) CanClock(delaying,339) TMC(notifyWait,100) MAIN(running,363) IDLE(ready,20) AIN(delaying,260)
Owned mutexes:
=== Platform ===
Last reset 00:16:22 ago, cause: software
Last software reset at 2021-03-14 14:00, reason: User, GCodes spinning, available RAM 39916, slot 1
Software reset code 0x0003 HFSR 0x00000000 CFSR 0x00000000 ICSR 0x00000000 BFAR 0xe000ed38 SP 0x00000000 Task MAIN Freestk 0 n/a
Error status: 0x00
Aux0 errors 0,0,0
Aux1 errors 0,0,0
Supply voltage: min 0.0, current 24.2, max 24.5, under voltage events: 0, over voltage events: 0, power good: yes
Heap OK, handles allocated/used 99/1, heap memory allocated/used/recyclable 2048/158/122, gc cycles 0
Driver 0: position 24640, ok, SG min/max 0/432, read errors 0, write errors 1, ifcnt 68, reads 8503, writes 18, timeouts 0, DMA errors 0
Driver 1: position 19797, ok, SG min/max 0/432, read errors 0, write errors 1, ifcnt 65, reads 8290, writes 18, timeouts 213, DMA errors 0, failedOp 0x06
Driver 2: position 2159, standstill, SG min/max 0/312, read errors 0, write errors 1, ifcnt 59, reads 8503, writes 15, timeouts 3, DMA errors 0, failedOp 0x71
Driver 3: position 0, standstill, SG min/max 0/386, read errors 0, write errors 1, ifcnt 50, reads 8252, writes 12, timeouts 256, DMA errors 0, failedOp 0x41
Driver 4: position 0, standstill, SG min/max 0/314, read errors 0, write errors 1, ifcnt 59, reads 8506, writes 15, timeouts 0, DMA errors 0
Driver 5: position 0, assumed not present
Driver 6: position 0, assumed not present
Date/time: 2021-03-14 14:17:03
Cache data hit count 1406805765
Slowest loop: 158.68ms; fastest: 0.11ms
=== Storage ===
Free file entries: 9
SD card 0 detected, interface speed: 22.5MBytes/sec
SD card longest read time 7.2ms, write time 148.3ms, max retries 0
=== Move ===
DMs created 83, maxWait 158038ms, bed compensation in use: mesh, comp offset 0.000
=== MainDDARing ===
Scheduled moves 10999, completed moves 10963, hiccups 0, stepErrors 0, LaErrors 0, Underruns [1, 0, 18], CDDA state 3
=== AuxDDARing ===
Scheduled moves 0, completed moves 0, hiccups 0, stepErrors 0, LaErrors 0, Underruns [0, 0, 0], CDDA state -1
=== Heat ===
Bed heaters = 0 -1, chamberHeaters = -1 -1
Heater 0 is on, I-accum = 0.4
Heater 1 is on, I-accum = 0.8
=== GCodes ===
Segments left: 1
Movement lock held by null
HTTP is idle in state(s) 0
Telnet is idle in state(s) 0
File is doing "G1 F1800.000" in state(s) 0
USB is idle in state(s) 0
Aux is idle in state(s) 0
Trigger is idle in state(s) 0
Queue is idle in state(s) 0
LCD is idle in state(s) 0
SBC is idle in state(s) 0
Daemon is idle in state(s) 0
Aux2 is idle in state(s) 0
Autopause is idle in state(s) 0
Code queue is not empty:
Queued 'M106 S249.9' for move 10985
1 of 16 codes have been queued.
=== Filament sensors ===
Extruder 0 sensor: ok
=== CAN ===
Messages queued 9142, send timeouts 9138, received 0, lost 0, longest wait 0ms for reply type 0, peak Tx sync delay 0, free buffers 17 (min 16)
Last cancelled message type 4514 dest 127
=== Network ===
Slowest loop: 148.43ms; fastest: 0.03ms
Responder states: HTTP(0) HTTP(0) HTTP(0) HTTP(0) FTP(0) Telnet(0), 0 sessions
HTTP sessions: 1 of 8
- Ethernet -
State: active
Error counts: 0 0 0 0 0
Socket states: 2 2 2 2 2 0 0 0
 

config.g

; General preferences
G90                                         ; Send absolute coordinates...
M83                                         ; ...but relative extruder moves
 
; Network
M550 P"Pruset II MK3s"                        ; Set machine name
M552 S1                                     ; Enable network
;*** Access point is configured manually via M587
M586 P0 S1                                  ; Enable HTTP
M586 P1 S0                                  ; Disable FTP
M586 P2 S0                                  ; Disable Telnet
M575 P1 S1 B57600							; Panel Due
 
; Drive Mappings
M569 P0.0 S1 V0                             ; Drive 0: X Axis
M569 P0.1 S0 V0                             ; Drive 1: Y Axis
M569 P0.2 S1 V0                             ; Drive 2:  Z Axis
M569 P0.3 S1 V0                             ; Drive 3: E Axis
M569 P0.4 S1 V0                             ; Drive 4: E1 used for Z1
 
; Motor remapping for dual Z and axis Limits
M584 X0 Y1 Z2:4 E3				; two Z motors connected to driver outputs Z and E1
M671 X-37:287 Y0:0 S10			; leadscrews at left (connected to Z) and right (connected to E1) of X axis
 
; Micrpstepping and Speed
M350 X32 Y32 E16 Z16 I1                     ; Configure microstepping with interpolation
M92 X200.00 Y200.00 Z400.00 E415.00         ; Set steps per mm
 
; Speeds, Acceleration and Jerk
M566 X240.00 Y240.00 Z24.00 E270.00 P1      ; Set maximum instantaneous speed changes (mm/min)
M203 X7200.00 Y7200.00 Z723.00 E1200.00     ; Set maximum speeds (mm/min)
M201 X1250.00 Y1250.00 Z1000.00 E500.00     ; Set accelerations (mm/s^2)
M204 P1250.0 T1250.0   			            ; set print and travel accelerations (mm(s^2)
 
; Motor currents
M906 X800.00 Y850.00 Z700.00 E550.00 I15    ; Set motor currents (mA) and motor idle factor in percent
M84 S30                                     ; Set idle timeout
 
; Printer geometry
M208 X0:250 Y-4:215 Z-0.5:205               ; Axis Limits
M564 H0							            ; allow unhomed movement
 
; Endstops for each Axis
M574 X1 S3								 ; Set endstops controlled by motor load detection
M574 Y1 S3                               ; Set endstops controlled by motor load detection
 
; Stallgaurd Sensitivy
M915 X S-2 F0 H200 R0		                ; Set X axis Sensitivity  
M915 Y S0 F0 H200 R0		                ; Set y axis Sensitivity 
 
; Input Shaper
;M593 F58.5 P4 S0.2  ; Input Shaping 3.3/3.4
M593 F58.5 
 
; Z-Probe Super PINDA
M574 Z1 S2                                      ; Set endstops controlled by probe
M558 P5 C"^io3.in" I1 H1 F570 T6000 A50 S0.005  ; PINDA 
 
; Mesh Grid and Sheets
M557 X24:233 Y15:195 P8                         ; Define mesh grid for probing
;G31 P1000 X23 Y5 Z1.530 	                    ; Textured Sheet
G31 P1000 X23 Y5 Z1.215 					    ; PEI Sheet
 
; Heatbed Heaters and Thermistor Bed 
M308 S0 P"temp0" Y"thermistor" T100000 B4725 C7.060000e-8       ; Set thermistor + ADC parameters for heater 0 Bed
M950 H0 C"out0" T0					                        ; Creates Bed Heater
M307 H0 R0.276 C365.2 D12.50 S1.00 V24.0 B0						; PID 3.2
M140 H0										                    ; Bed uses Heater 0
M143 H0 S120                                                    ; Set temperature limit for heater 0 to 120C Bed
 
; HotEnd Heaters and Thermistor HotEnd      
M308 S1 P"temp1" Y"thermistor" T100000 B4725 C7.060000e-8       ;define E0 temperature sensor e3d
M950 H1 C"out1" T1					                            ; Create HotEnd Heater
M307 H1 R2.301 C162.7:122.2 D7.10 S1.00 V24.3 B0
M143 H1 S285                                                    ; Set temperature limit for heater 1 to 285C HotEnd
M302 S185 R185
 
; Fans
M950 F1 C"out5" Q100				; Creates HOTEND Fan
M106 P1 T45 S210 H1                 ; HOTEND Fan Settings
M950 F0 C"out6" Q100				; Creates PARTS COOLING FAN
M106 P0 H-1                         ; Set fan 1 value, PWM signal inversion and frequency. Thermostatic control is turned off PARTS COOLING FAN
 
; Tools
M563 P0 D0 H1 F0                            ; Define tool 0
G10 P0 X0 Y0 Z0                             ; Set tool 0 axis offsets
G10 P0 R0 S0                                ; Set initial tool 0 active and standby temperatures to 0C
 
; StealthCop calibration
M17
G4 P150
G1 X1 Y1 Z1 	
G4 S2
M84 ; disable motors
 
; Filament Sensor Logic
M98 P"0:/macros/02_Functions/StartupFilamentSensorCheck"  ; Runout Sensor Logic: Startup with filament = runout sensor active   Startup without filament = autoload active
 
 
 
 
 
 
 
 
 

[deckingman]

@Argo Yes, your extruder jerk is really low and there is an interaction between extruder jerk and pressure advance. So if disabling PA cures the pauses, whack the extruder jerk up. It does no harm to set it really high as there are very few instances when it would ever come into play. I have mine set to 3600 (you have 270). I'd also whack the extruder acceleration up too. If it's too low, then it could slow down any XYE print moves. Similarly to extruder jerk, I set extruder acceleration really high (6000) so that it's out of the way.

Edit. .......and changing jerk policy to 0 might help.

[Argo]

@deckingman

Disabling PA does not cure the problem. When it happens it always pauses directly on the apex of a corner.

With my Duet 2 I never had problem with jerk policy set to 1.
The problems come and go with this new board though.

Timeline with issues so far:

• Duet 3 Mini arrived - flashed latest stable 3.3.2 release: homing brutally crashes. In the forums i found a thread about a sensorless homing bug in 3.2.2. Version 3.3b1 fixes the issue
• 3.3b1 flashed: extruder is skipping steps / under-extrudes a lot. So I took the source code of 3.2.2 and took the one file that fixes sensorless homing from 3.3b1 and compiled my own working 3.2.2 and reported the bug on the forum. Printer worked like a charm (and I should've better stopped here )
• 3.3.b2 was released so I wanted to give it a try. Same issue as with 3.3b1. Flashed back to 3.2.2 and everything worked as expected. Then I got some instructions on how to test 3.3b2 for specific issues so I reflashed 3.3b2 and the underextrusion issue was gone, which I can't really explain why.
• New issue since then with 3.3b2 and 3.2.2: Slowdowns during probing and short pauses during a print. Adjusting to a higher PWM frequency fixed the probing slowdowns but the pauses during the print are still there.

Something is still off here. I was also expecting that there was a hidden gcode somewhere in the Slicer start gcode that enables debugging during a print but setting PWM to 10hz outside a print could reproduce the slowdowns during G29 probing. I also have no daemon.g with some messy code that might draw CPU power from the board.
Speaking of CPU power. Is there a way to see how much CPU power is used? I never really bothered with debugging but maybe that might help.

[deckingman]

@Argo I appreciate that you used jerk policy 1 before with Duet 2 but have you tried setting it to 0 with this board/firmware combination? Does the slow down correspond with layer change? Alternatively, do you use Z hop during retraction? - if so try disabling it. I'd be inclined to disable all the "embellishments" like input shaping, mesh grid, stall guard to see if that makes a difference and of so, then re-enable them one at a time.

[Argo]

@deckingman

Thanks for your help. Of course I tried your suggestion before posting.
Right now I've disabled / changed the following features:

• input shaper (DAA) off.
• dynamic fan speed (slicer)
• jerk policy sest to 0

The issue does not happen during a layer change. It happens always at the outer perimeter at the on a test Voron Cube.
It also happens at the exact same spot when printing the outer perimeter.

Right now I'm suspecting a Slicer issue/bug to be honest. The Voron cube has many overhangs paired with 45° corners.

The problem with those problems I'm having that they can be quite similar but have nothing to do with each other, like the problem with the probing slow downs when the PWM frequency is low. My thought was slow down, ok, when the print is having slow downs it must be the same issue but now I think it isn't.

I'll do another print and disable all overhang detection features in the Slicer software and report back.

EDIT.

Yes. The slow down during the print was a actual bug with the Slicer software and it's overhang detection and has nothing to do with the probing slowdowns when having a low PWM bed frequency. Right now I'm using 100hz for heater and fans. I don't know if that makes sense for the electronics of the board but probing does not slow down anymore with that frequency.

[deckingman]

@Argo It's an easy trap to fall into. Problem A is caused by xyz, problem B looks just like A so one naturally thinks the cause is the same xyz,

So, in summary you just have weird behaviour when probing and when the bed PWM frequency is set to 10Hz. Increasing the bed PWM frequency to 100 Hz fixes the problem. Is that correct?

If so, then I'm not likely to be able to help because I don't use a Z probe (long story but I use a kinematic mount and have things arranged such that the nozzle acts as a switch for homing Z - the bed is mechanically flat and level so I don't use any form of firmware compensation).

But to help others help you, I suggest you provide more details. i.e. the board type, firmware version, what type of probe you use, etc. And maybe a short video clip showing the behaviour at 10Hz and again at 100Hz.

[Argo]

@deckingman
The pause during print has been fixed. It was a bug in the Slicer software

Taking your advice to sum the issue up:
Slowdown during probes with low PWM bed frequency. After read the recommendation in the forum to use 10hz it began to have slowdowns during the probing sequence.

Printer: Prusa MK3*ish printer. With Super PINDA probe. Mk3 heated bed 24v.

Board: FIRMWARE_NAME: RepRapFirmware for Duet 3 Mini 5+ FIRMWARE_VERSION: 3.3beta2 ELECTRONICS: Duet 3 Mini5plus Ethernet FIRMWARE_DATE: 2021-03-10 13:41:12

Bed PWM frequency at 10hz:

https://www.youtube.com/watch?v=OEk7ahEcNFs

By hearing at the tone you notice tiny little hiccups / slowdowns. Those are gone going up to 100hz PWM for the bed heater.

[droftarts]

@Argo I’d guess the bed pwm is interfering with the probe. You can turn off heaters when probing, using the B parameter of M558. See https://duet3d.dozuki.com/Wiki/Gcode#Section_M558_in_RepRapFirmware_Num_3

Ian