Pressure advance user feedback

  • Following on from some other threads, it seems that the optimum setting for pressure advance on my machine is a lot higher than would seem "normal". There are a number of reasons why this may be the case, not least of which is that I am using a 3 input mixing hot end. However, the "normal" figures are based on what may be limited feedback from users and there area couple of "gotchas" that may have affected initial findings and test results.

    So I though it might be useful if people could (when convenient) revisit their pressure advance settings and post their findings here so that we can come up with a better range of numbers that new users can refer to as an initial starting point. It would help too if things like the hot end make, extruder type and Bowden tube lengths could be included, then we can maybe find some sort of correlation between machine parameters and pressure advance settings.

    "Gotcha" number 1 is that if the extruder Jerk setting is too low, the extruder acceleration will be limited and will slow down the entire print (not just the pressure advance). I use E3D titans and ran into this problem when my jerk setting was at 600mm/sec. Doubling it to 1200 fixed the issue.

    "Gotcha" number 2 isn't really a gotcha but from the extensive testing I've recently been doing, it seems that the effect of pressure and thus the effectiveness of the compensation are more difficult to evaluate at low printing speeds. At least that is what happens on my machine. So I've been evaluating at print speeds of 90mm/sec upwards. Once a setting has been found at high speeds, it will work equally well at lower speeds but it's harder to see the difference. At least, that is what I've found…..

    Evaluation is difficult as it is just a subjective visualisation of print quality. What works best for me is to print a number of cubes from 10 to 40 mm square, spaced a reasonable distance apart (10 to 40mm) and printed with 100% solid infill at a fast speed. Then, as well as visual check, I run my finger around the top perimeter and feel for roughness. Without compensation, pressure builds up and at the end of the move the print head starts to decelerate but because of the pressure build up, the filament is forced out at the same speed even though the extruder decelerates. This results in over extrusion which gives this rough finish at the ends of the moves. When the compensation is set right, this roughness disappears and can belt felt as well as seen. Printing cubes like this with varied spacing also allows one to evaluate and set the retraction amount at the same time.

    So, with my Diamond hot end, fed with 3 E3D Titans and using 3 Bowden tubes of 160mm in length, the optimum pressure advance setting for me is 0.5 to 0.6. This works at all speeds, on small moves as well as long ones and using a single filament or all 3 together. Observing the extruder gear behaviour it looks crazy, but it prints very nicely.

    Please post any results and findings here as well as test methodology if you have a better one.


  • I am just waking up so I will add in more details later but my findings are the results are very visible at lower speeds 42mm/s (at least when you have high jerk settings) 3600 for me. Also another important factor is already having things extremely fine tuned with pressure advance turned off before enabling it.

    The effects can be seen in the icicle retraction test but when tuned well for this test they do not correlate well for tests like 3d benchy. You can even see the infill pattern being affected when switching it off and on.

    Ill have to try some higher speed cube prints like you have done and see how that affects other tests like icicle and benchy.

  • I can give you my settings now Ian. I might be able to try a bit of tuning using the method above at some point.
    Machine 1 Corexy, titan extruder, pancake 20mm stepper, mounted direct on the x-carriage, genuine e3d volcano, 0.4mm nozzle 36w heater (by resistance), I use for ABS, PLA, PETG, for flexibles I haven't settled on a value is too finely balanced.

    M92 X200 Y200 Z200			; Set axis steps/mm
    M92 E440:440				; Set extruder steps/mm
    M906 X800 Y800 Z800 E700            	; Set motor currents (mA)extruder current just enough will click not strip
    M201 X2000 Y2000 Z200 E1000            	; Accelerations (mm/s^2)
    M203 X15000 Y15000 Z3000 E3600       	; Maximum speeds (mm/min)
    M566 X600 Y600 Z30 E20                           ; jerk values
    M572 D0 S0.0125	; pressure advance 

    Machine 2 Delta with flex3drive, 32mm motor via 90cm cable drive, genuine e3d v6, 0.4mm nozzle, 30w heater (claimed), ABS, PLA.

    M92 X200 Y200 Z200			; Set axis steps/mm
    M92 E3800:3800				; Set extruder steps per mm (flex3drive)
    M906 X1100 Y1100 Z1100 E800:800 I60	; Set motor currents (mA) and increase idle current to 60%
    M201 X3000 Y3000 Z200 E200		; Accelerations (mm/s^2)
    M203 X15000 Y15000 Z15000 E3600		; Maximum speeds (mm/min)
    M566 X1200 Y1200 Z1200 E100                        ;jerk values
    M572 D0 S0.025				;set extruder pressure advance

    I use a slightly higher pressure advance value here, although the extruder performs as a direct drive, there is slight latency in the drive cable.

    I say claimed with the v6 heater, as the genuine volcano kit I got a few days ago from e3d has a "30w" heater with a resistance of 6.2ohms so only 26W in reality at 12v.

    Hope this helps.


Looks like your connection to Duet3D was lost, please wait while we try to reconnect.