Pressure advance not working for me (SOLVED)

deckingman

I'm going back to basics with my latest multi-input hot end. I developed a method of evaluating pressure and setting pressure advance a few years back. Essentially, it involves printing a series of 300mm long lines in the X direction with a 10mm non-print Y move between each line, at increasing speeds and observing the line thickness at the start and end of each line as well as any filament that gets laid down during the non print move.

I've tried values of PA up to 0.6 but nothing makes any difference. Observing the extruder gear, I can't detect anything happening at the start and end of the lines, over and above what normally happens without PA. Previously, when PA was working, the extruder acted in a pretty alarming way and with values in the order of 0.4 and upwards, it would even run backwards at the end of those lines.

So my conclusion is that PA simply isn't working with my machine. I'm running RRF3.3 RC3 because RRF3.3 final breaks my machine https://forum.duet3d.com/topic/24815/3-3-final-destroys-my-corexyuv-but-rc3-is-ok/6

It's a mixing hot end. The tests are being done using tool 0 which is defined thus:-

M563 P0 S"Black" D0:1:2:3:4:5 H1 ; Define tool 0 (drive 3 not used)
G10 P0 X0 Y0 ; Set tool 0 axis offsets
G10 P0 R0 S0 ; Set initial tool 0 active and standby temperatures to 0C
;M568 P0 S1 ; Enable mixing for tool 0
M567 P0 E1.00:0.00:0.00:0.00:0.00:0.00 ; Set mixing ratios for tool 0

So essentially, it's just using extruder zero at 100%.

All 6 extruders are connected to expansion boards - 3 per board.

I note that the release notes for 3.4 state under "known issues",

"Duet 3 + CAN-connected expansion boards] If a sequence of short movements is commanded involving only motors connected to expansion boards, then the moves might be sent too quickly to the expansion boards, eventually resulting in lost moves".

So might that be the reason why PA is not working for me??

deckingman

Any comment on this? Is it likely to be looked into? Is it likely related to other known firmware issues?

dc42

@deckingman it doesn't sound to me that you are doing a sequence of very short moves; so I don't think that known issue is relevant.

I think the reason why PA is not very effective in your setup is that your hot end has 5 inputs that merge together at the top, followed by a mixing chamber that presents significant resistance to filament flow. So when you are printing with just one filament, the extra pressure that PA tries to add by moving that filament additional distance will largely be absorbed by the other 4 filament inputs. In effect you need to apply 5 times as much PA to compensate for filament compression in the Bowden tube, because you need enough pressure to compress all 5 filaments. I recall writing this a few years ago, and that you disagreed with me.

IMO in a mixing setup, PA would ideally be applied in full to all filaments, instead of in proportion to the mix ratio - just as retraction is applied to all filaments.

If you were to mix all 5 filaments in approximately equal ratios and use a large amount of PA, I think you might get better results. Using a vary large PA with just one filament printing is less likely to work well because there is a limit to how much PA a single extruder can apply.

Another possible reason is that between the top and bottom of your hot end, you have a mixing chamber that probably imposes significant resistance to filament flow. Therefore, although applying PA may be effective at compensating for filament compression in the Bowden tube, it may be less effective at providing additional pressure at the top of the nozzle. Your 6th filament feeding into the bottom of the mixing chamber (which AFAIR you said you were planning to move to the top) might be usable as a means to apply PA at that point.

deckingman

@dc42 No, you misunderstand.

What I'm saying is that the extruders are not behaving as they always have done. I'm talking about visual observation of the extruder gear. There was no discernable difference in the action of that gear between having PA set to zero and PA going all the way up to 0.6. Whereas before, with PA set to such high values, the extruder(s) would do some pretty crazy things and even run backwards at the end of such a long, fast move. This wasn't happening. I could not detect any discernable difference in the action of the extruder at the start and end of those long, fast moves regardless of whether PA was set to zero or 0.6.

Having said that, I've just managed to get some sort of print done and from what I could tell by looking at the extruder gears during that print, it appeared that PA might have been working as it should. I have a theory why PA did not work in my earlier tests but I need to test that out.

I'll get back to this in a few days time.

Edit. BTW, this latest version of my hot end has no mixing chamber - so it isn't fundamentally much different to Diamond 5 colour (except that it has 6 inputs).

dc42

@deckingman I'm sorry I misunderstood. I can't think of any reason why PA would sometimes not be applied when it should. The 3.3-post-RC3 upgrade notes at https://github.com/Duet3D/RepRapFirmware/wiki/Changelog-RRF-3.x-Beta-&-RC#reprapfirmware-33-post-rc3 don't have anything that looks relevant to me except for the short-fast-segments issue that you mentioned.

dc42

PS - another user has reported PA not working, see https://forum.duet3d.com/topic/24848/pressure-advance-seems-not-to-work-using-ideamaker, so perhaps there is a firmware issue after all.

A Former User

@dc42 said in Pressure advance seems not to be working for me:

PS - another user has reported PA not working, see https://forum.duet3d.com/topic/24848/pressure-advance-seems-not-to-work-using-ideamaker, so perhaps there is a firmware issue after all.

It's me! it doesn't work at all with latest 3.3... I thought it was Ideamaker's fault...so you are telling that maybe could be a FW issue maybe? uh...could you check please?

EDIT
I've tried to slice the object with Simplify 3D and now I'm starting the print, I'll let you know...if it doesn't work neither with S3D I'm afrayd it's a FW issue

EDIT OF EDIT
With S3D it works

deckingman

@dc42 said in Pressure advance seems not to be working for me:

PS - another user has reported PA not working, see https://forum.duet3d.com/topic/24848/pressure-advance-seems-not-to-work-using-ideamaker, so perhaps there is a firmware issue after all.

David,

The theory that I need to test is that when PA seemed not to be working, firmware retraction was set to zero. The subsequent print where PA seemed like it was working had retraction set to 2mm. That's the only difference I can think of between those two "prints". So what I need to do is repeat my first test with retraction set to some non-zero value to see if that is indeed the reason or if it's a red herring. But we have visitors this weekend so it'll be a couple of days before I can do that test.

deckingman

I've marked this as solved although it remains a mystery to me. It appears that PA is doing what I expect during "normal" printing but not when testing using my "made up" gcode file. Here is a sample of the test file if anyone can spot anything that I can't.

G1 X330 U330 A330 E16 F3600; 60 mm/sec no retract
G1 Y30 V30 B30 F6000
G1 X30 U30 B30 E16 F3600; 60 mm/sec with retract
G10
G1 Y40 V40 B40 F6000
G11
G1 X330 U330 A330 E16 F4200; 70 mm/sec no retract
G1 Y50 V50 B50 F6000
G1 X30 U30 B30 E16 F4200; 70 mm/sec with retract
G10
G1 Y60 V60 B60 F6000
G11

Essentially it's just a series of 300mm moves in X at increasing speeds, while extruding 16mm of filament, interspersed with a 10mm move in Y without any extrusion. Alternate moves have either firmware retraction or no retraction. I can't (visually) detect any change in extruder behaviour (or the start/end line thickness) from zero PA up to silly numbers like 0.8. Yet if I print something like a 200 x 200mm cube, I can easily see the change in extruder behaviour and start/end line thickness with varying amounts of PA. So there must be something wrong with my test gcode but I can't see what.

dc42

@deckingman I have added this to my investigations list.