Pressure Advance Calibration

mihaitintea

Hello

If I run the python file and generate the PA test gcode file, I see that it contains G1 Xxxxx Yyyyy Eeeee commands whose E values are very large, e.g.:

G1 X-47.400 Y3.200 E15.7653

whereas if I look into any gcode file generated by Cura, I only see very small E values in the G1 commands, e.g.:

G1 X9.063 Y1.1 E0.16886

Why does this happen, and how can I safely run/print the generated PA test gcode file ? (i.e., won't my extruder/hotend be blown up because of such enormous E values in the G1 commands ?)

Regards
Mihai

OBELIKS

These are all long printing moves, as the models usually have small ones.
You can always check in gcode.ws

jens55

Is pressure advance supposed to change with different extrusion thickness ? I was always under the impression it was relatively constant at least with the same filament.
I had run calibration for PLA 0.8 mm nozzle and 0.2 mm layer thickness and got a number of 0.45.
I subsequently printed out a model at 0.4 mm layer thickness and ran into some odd issues that looked a lot like pressure advance issues but dismissed that because I had just calibrated that.
I ran another pressure advance calibration but this time with 0.4 mm layer thickness and got a figure closer to 0.2 for pressure advance.
Does this mean that in order to run a long bowden tube (where pressure advance is most critical) I have to calibrate every combination of filament, nozzle size, layer thickness and possibly nozzle temperature ?
If so then the case for a direct extruder setup got a lot stronger ....

deckingman

@jens55 I must confess that although I've played around with various parameters in relation to pressure advance, layer height vs nozzle diameter isn't one of them. I'd say that it's eminently possible that using a layer height of 20% of the nozzle diameter could offer a restriction to the filament flow as it exits the nozzle (because the nozzle tip is closer to the part being printed). Which would of course affect the pressure in the hot end, making it higher than if the layer height was greater.

That being the case, then I'd expect the same problem to manifest itself, regardless of whether the extruder is direct drive or Bowden (because the cause of the pressure increase is downstream of the extruder).

jens55

@deckingman said in Pressure Advance Calibration:

That being the case, then I'd expect the same problem to manifest itself, regardless of whether the extruder is direct drive or Bowden (because the cause of the pressure increase is downstream of the extruder).

All I have ever read about pressure advance seems to indicate that a direct extruder only uses minimum amounts ... or put into other words, just a smidgen to optimize things.
Going from 0.45 to 0.2 in a bowden setup becomes the difference between a good print and a failed print. The print I have been dealing with recently ended up with holes in the outer wall layer (still testing this finding) at the point where a layer started.
I would very much like to have confirmation if my understanding of this is correct before I contemplate a major printer configuration change !

deckingman

@jens55 Common perception is that pressure build up happens mostly in the Bowden tube, due to the filament buckling and twisting, which is why direct extruders are less prone to it's effects. Nobody really knows for sure but it seems a reasonable hypothesis. I was just surmising that if you are seeing a big difference between layers heights, then it might be because the filament flow as it exits the nozzle is being restricted with small layer heights. In which case, the cause of the pressure might be something other than the filament buckling and twisting. So if that is the case, then a direct extruder may exhibit the same behaviour. It's just conjecture on my part.........

jens55

@deckingman , thanks for your input. I agree with your thinking.
I am hoping that somebody with first hand experience might chime in as we are both guessing.

Munce31

Hi All

I've been using this python script to fine tune my pressure advance on my zesty nimble extruder. It has been a massive help in this process.

Just one suggestion - I don't know if it is possible - however, I would like to see what pressure advance setting it is currently at during the print. Preferably to be displayed like a pop, similar to how the Report PA macro works in https://forum.duet3d.com/topic/6181/tuning-macros-menus-accel-jerk-retraction-pressure-advance. Is it possible to add something to this python script to report current pressure advance? It would make fine tuning the pressure advance just a little easier by taking a chunk of the guess work out of it.

Thanks,
Vince

wilriker

@Munce31 You could manually add M117 <current PA value> into the GCode file after every change of PA.

Munce31

@wilriker said in Pressure Advance Calibration:

@Munce31 You could manually add M117 <current PA value> into the GCode file after every change of PA.

Yeah I figured I could do that as well, but that potentially could be quite laborious. I was thinking more along the lines of something like M117 <current PA value> being added directly into the python script so that it can be automated.

Phaedrux

The Report PA Value macro just sends the command for pressure advance without any parameters so that the current value gets echoed back. Doing a find and replace in the gcode file shouldn't be too hard to add it.

OBELIKS

@Munce31 If you are using the script @jens55 posted, you can add

print("M117 %.3f" % pressure_advance)

after

print("M572 D0 S%.3f" % pressure_advance)

Munce31

@OBELIKS said in Pressure Advance Calibration:

@Munce31 If you are using the script @jens55 posted, you can add
print("M117 %.3f" % pressure_advance)
after
print("M572 D0 S%.3f" % pressure_advance)

Hi @OBELIKS thanks for providing that. It'll make it easier to watch the test print and see what pressure advance it current is running at.

zzing

I was reading though some of this to take a look at my own values.

What is quite odd is that the point at which the fast areas and slow areas are equal in width is at 1.75 - which seems crazy high according to the documentation. I am using a bowden tube of capricorn of length ~600mm. It has a bondtech extruder and a smart effector/v6 combo on the output.

I was using the original script from the start.

The other relevant settings from config:

M92 X80.00 Y80.00 Z80.00 E830.00                  ; Set steps per mm
M566 X1200.00 Y1200.00 Z1200.00 E40.00             ; Set maximum instantaneous speed changes (mm/min)
M203 X18000.00 Y18000.00 Z18000.00 E5000.00        ; Set maximum speeds (mm/min)
M201 X3000.00 Y3000.00 Z3000.00 E120.00            ; Set accelerations (mm/s^2)
M906 X1000.00 Y1000.00 Z1000.00 E500.00 I30        ; Set motor currents (mA) and motor idle factor in per cent

I also just figured out this:

M92 E820
M592 D0 A0.001246 B0.000406 T230

Do these seem reasonable? The max instanteous speed sounds low compared to one post above.

This is an upgraded mini kossel.

wilriker

@zzing said in Pressure Advance Calibration:

M566 X1200.00 Y1200.00 Z1200.00 E40.00             ; Set maximum instantaneous speed changes (mm/min)
Do these seem reasonable? The max instanteous speed sounds low compared to one post above.

These values really are rather odd. How did you find such a low max inst. speed value for E? I have mine at 1800 and think I am still rather on the low end. Also this 40 of yours will probably lower total print speed.

zzing

I am not sure if I set that or it was configured from the tool to build a config. With all those other values I could probably use realistic values for a delta machine.

I can retest then.

bot

If a user has a very geared down extruder, low E jerk is required (if using pressure advance). IIRC, extruder jerk is only used when pressure advance is in effect. Otherwise, it has no effect on print time. E Accel has an effect on print time when set too low, with or without pressure advance.

My extruder can't use correct amounts of pressure advance without a low E jerk value of ~90.

SnakeSP

I have very strange results when trying to calibrate PA.
Here are the speed, acceleration and jerk settings:
M203 X15000.00 Y15000.00 Z15000.00 E6000.00
M201 X2000.00 Y2000.00 Z2000.00 E2000.00
M566 X1200.00 Y1200.00 Z1200.00 E1200.00
I'm using BMG cloned extruder (if matters) with bowden of 550 mm and ID 2 mm.
Used python script to print patterns with the following settings:
extrusion_width = 0.5 (nozzle is 0.4 actually, 0.5 is used in slicer as an extrusion width)
layer_height = 0.2
filament_diameter = 1.72 (measured with calipers)

travel_speed = 150
first_layer_speed = 15
slow_speed = 5
fast_speed = 70

I tried PA settings from 0.1 to 1.0 at first and there were no signs of PA working. Further experiments showed that noticable result is there only in range 1.5 to 1.9. But this looks insane... Or this is normal?
Also i noticed that the speed is limited to 63-65 mm/s (due to too low extruder settings possibly?).

Phaedrux

@SnakeSP Perhaps try this method of tuning pressure advance.

https://duet3d.dozuki.com/Wiki/Pressure_advance#Section_Methods_of_finding_the_right_amount_of_pressure_advance

SnakeSP

@Phaedrux
I tried the first one, with the python script.
Will try the second one, with the cube, but need to understand how to make Simplify3d to align seam at the center of the side, for now i can only align it to one of the corners.

One more thing still bothers me: i have not yet found the proper settings for BMG extruder (speed, acceleration, jerk) there are very different number said to be best in every topic i've read and there is nothing about it on Bondtech site. Can you point me where can i find such info please?