How to speed up print with high linear advance.
-
@Turbo With a PA value of 0.86, the extruders will behave in way that looks alarming. Depending on the extruder, it may sound alarming too. My E3D Titans made horrible clacking noises, the Bondtech BMGs do not. Once you increase extruder jerk, you'll probably end up with a significantly lower PA value. But my advice would be to use whatever PA value gives the best print quality. If you are not comfortable with how the extruders look or sound, print at a lower speed, rather than dial back PA from it's optimum value.
-
@deckingman So i adjusted jerk to 2400 after doing some research, and not wanting to over-do it. Now calibrating Pressure Advance seems to do nothing, with ranges between 0-2, and anything past 1 starts to make a terrible noise like you described.
-
@Turbo If you settled on a value of 0.86 before with very low extruder jerk then I would expect that you need a much lower value than that with 2400 jerk. So I'm not surprised that alarming things happen with values >1. What do you mean when you say PA isn't doing anything and how are you setting it?
-
@deckingman Im using a PA test gcode. I made a post on using the marlin generator with RRF, that's what I'm using again. When I say it doesnt do anything, thats really all I'm saying. as in theres no difference between 0 PA and 2 PA, and everything in between. I'll run it in between 0 and 1 real quick with .1 stepping and send a picture.
-
@Turbo Let's take a step back. In your OP you said that you had dialled in PA and arrived at a value of 0.85 which fixed the issues but slowed the print down. I suggested that the slow down was likely due to the low extruder jerk that you were using. So can we establish if printing the same part, with the same PA value, but with the new extruder jerk setting improves the print speed?
-
@deckingman correct, however it begins to underextrude/not enough material and lines don't connect.
15867946996626117362152457791302.jpg
This is a test with 0-0.5 values. 0.05 stepping. Because 0-2 and 0-1 showed no change, similar to this one. Not the best pic but it wouldn't let me upload a higher quality one.
Cant stop tuning wont stop tuning.
Dbot, Custom i3, Voron 0&2 -
If PA is doing nothing but make a terrible noise with jerk set to 2400, your jerk is likely set way too high and your extruder can not actually instantly change speed to 40 mm/s.
2400 is likely almost ten times the amount of extruder jerk that is reasonable and/or would allow you to use PA appropriately.
*not actually a robot
-
@bot I'm using 2400 based off @deckingman jerk because I also use a Titan extruder. Not sure how they would be drastically different. I'm also using full size nema 17s.
-
@bot said in How to speed up print with high linear advance.:
If PA is doing nothing but make a terrible noise with jerk set to 2400, your jerk is likely set way too high and your extruder can not actually instantly change speed to 40 mm/s.
2400 is likely almost ten times the amount of extruder jerk that is reasonable and/or would allow you to use PA appropriately.
We'll have to disagree on that. But it likely depends on the extruder - you use very high gearing don't you?
-
Which aspect do you disagree with? 2400 is completely unnecessary. There is no scenario, except retraction, that ever requests the extruder to move 40 mm/s during a print.
-
@Turbo Difficult to tell from the pic but it looks like under extrusion in the middle of the moves
@bot said in How to speed up print with high linear advance.:
Which aspect do you disagree with? 2400 is completely unnecessary. There is no scenario, except retraction, that ever requests the extruder to move 40 mm/s during a print.
Extruder Jerk interacts with PA - if you don't believe me, ask DC42. With a Titan or BMG and 400 steps per mm, I can categorically say that if extruder jerk is set to 600, it will slow things down. I haven't just made that up - that's from real use case testing.
-
@deckingman You must use very high amounts of PA.
Even then, 2400 is way more than is necessary. In what situation, of PA, would the extruder need to instantly reverse by 40 mm/s? None that I can imagine.
*not actually a robot
-
@bot said in How to speed up print with high linear advance.:
@deckingman You must use very high amounts of PA.
Yes - like the OP. Maybe it's a feature of multi input hot ends because my Bowden tubes are only about 200 mm long - but then I have 5 of them.
-
@bot said in How to speed up print with high linear advance.:
@deckingman You must use very high amounts of PA.
Even then, 2400 is way more than is necessary. In what situation, of PA, would the extruder need to instantly reverse by 40 mm/s? None that I can imagine.
I don't profess to know how the interaction of extruder jerk and PA work - only that there is that interaction.
-
@deckingman What it looks like to me is not enough pressure advance. However, there is no change between 0 PA and 3 PA and everything in between. Kinda lost on why its not making a difference, especially when i can clearly hear the motor upping the pressure advance with each line.
Cant stop tuning wont stop tuning.
Dbot, Custom i3, Voron 0&2 -
@deckingman it's a very simple interaction if you care to think about it.
A given X/Y feedrate at a specified layer height/extrusion width combo has a nominal E axis movement speed. Typically, this is no greater than 5 mm/s. PA simply applies an advancement or adjustment based on the current instantaneous actual E acceleration happening at the moment. If we imagine that the nominal E speed of the move would be 5 mm/s, and the PA is set such that during acceleration it requires TWICE as much plastic to flow, it will have to jerk to 10 mm/s instantly. If we have our jerk speed set lower than 5 mm/s, it will slow down the XY acceleration of the print move (and therefore the required E acceleration of the print move, because they are coordinated).
*not actually a robot
-
What print speeds are you using for calibrating PA? What extrusion width and layer height combo are you using?
*not actually a robot
-
@Turbo said in How to speed up print with high linear advance.:
@deckingman correct, however it begins to underextrude/not enough material and lines don't connect.
15867946996626117362152457791302.jpg
This is a test with 0-0.5 values. 0.05 stepping. Because 0-2 and 0-1 showed no change, similar to this one. Not the best pic but it wouldn't let me upload a higher quality one.
OK. Taking a gander at that picture, it looks like the lines are fat at both ends but thin in the middle. What normally happens is that at the start of a move, the carriage accelerates but the melt rate of filament lags behind, so you get under extrusion at the start of a move. Then at the end of a move, the carriage decelerates but because pressure has built up, filament continues to get forced out at a faster rate, so you get over extrusion at the end of a move. PA compensates by advancing the extruder at a faster rate at the start of a move and retarding the extruder so that it slows down more at the end of a move. But as I said, your image seems to show that the lines are thin in the middle, which might indicate that you are simply exceeding the melt rate that the extruder is capable of.
-
@bot
From the generated .gcodeCreated: Mon Apr 13 2020 09:13:37 GMT-0700 (Pacific Daylight Time) ; ; Settings Printer: ; Filament Diameter = 1.75 mm ; Nozzle Diameter = 0.4 mm ; Nozzle Temperature = 210 °C ; Bed Temperature = 60 °C ; Retraction Distance = 2.6 mm ; Layer Height = 0.2 mm ; Z-axis Offset = 0 mm ; ; Settings Print Bed: ; Bed Shape = Rect ; Bed Size X = 305 mm ; Bed Size Y = 188 mm ; Origin Bed Center = false ; ; Settings Speed: ; Slow Printing Speed = 1200 mm/min ; Fast Printing Speed = 4200 mm/min ; Movement Speed = 7200 mm/min ; Retract Speed = 1800 mm/min ; Printing Acceleration = 2400 mm/s^2 ; Jerk X-axis = firmware default ; Jerk Y-axis = firmware default ; Jerk Z-axis = firmware default ; Jerk Extruder = firmware default ; ; Settings Pattern: ; Linear Advance Version = 1.5 ; Starting Value Factor = 0 ; Ending Value Factor = 0.5 ; Factor Stepping = 0.05 ; Test Line Spacing = 5 mm ; Test Line Length Slow = 20 mm ; Test Line Length Fast = 40 mm ; Print Pattern = Standard ; Print Frame = false ; Number Lines = true ; Print Size X = 98 mm ; Print Size Y = 75 mm ; Print Rotation = 0 degree ; ; Settings Advance: ; Nozzle / Line Ratio = 1.2 ; Bed leveling = G29 ; Level bed ; Use FWRETRACT = false ; Extrusion Multiplier = 0.97 ; Prime Nozzle = true ; Prime Extrusion Multiplier = 2.5 ; Prime Speed = 1800 ; Dwell Time = 2 s ; ; prepare printingCant stop tuning wont stop tuning.
Dbot, Custom i3, Voron 0&2 -
@bot said in How to speed up print with high linear advance.:
@deckingman it's a very simple interaction if you care to think about it.
A given X/Y feedrate at a specified layer height/extrusion width combo has a nominal E axis movement speed. Typically, this is no greater than 5 mm/s. PA simply applies an advancement or adjustment based on the current instantaneous actual E acceleration happening at the moment. If we imagine that the nominal E speed of the move would be 5 mm/s, and the PA is set such that during acceleration it requires TWICE as much plastic to flow, it will have to jerk to 10 mm/s instantly. If we have our jerk speed set lower than 5 mm/s, it will slow down the XY acceleration of the print move (and therefore the required E acceleration of the print move, because they are coordinated).
...........and segmented moves such as circles where PA is no applied to each segment but at the end of a curve?
But I don't want to continue this because I have my real life experience of empirical testing and what happens in fact, which you don't believe. Therefore, you must think that I am just making this all up for some unknown reason, or that I am a liar. Either way, there is no point in my continuing.
