Linear advance, what to look for

[dc42]

@johnjohn1990 said in Linear advance, what to look for:

Thank you for your reply how on earth did u come to that conclusion regarding retraction? Ive been trying to find a way to "calculate" retraction...

Like you I built a printer with a long Bowden tube, and I was getting a small zit on the print at the start of every travel move. I struggled with this for hours before I eventually tried 8mm retraction, which solved it. The Bowden tube was about 650mm long.

Another of my printers has a 330mm Bowden tube and needs 4.5mm retraction. So my rule of thumb is based on experience of these 2 printers.

What I believe is happening is that when you extrude, the back pressure from the hot end causes the filament to compress by taking up a helical path inside the Bowden tube. When you retract, you have to pull the filament straight again before it is actually pulled out of the hot end. The longer the Bowden tube, the more the filament can shorten by taking up that heilcal path. So it stands to reason that retraction needed should increase linearly with the length of the Bowden tube. Then you need to add the retraction that you would need if there is no Bowden tube, which is normally reckoned to be in the range 0.5 to 2mm.

When I switched to Capricorn tuning, I was able to reduce retraction by 1mm, and with pressure advance enabled I reduced it by a further 1mm.

[johnjohn1990]

@dc42 said in Linear advance, what to look for:

@johnjohn1990 said in Linear advance, what to look for:

Thank you for your reply how on earth did u come to that conclusion regarding retraction? Ive been trying to find a way to "calculate" retraction...

Like you I built a printer with a long Bowden tube, and I was getting a small zit on the print at the start of every travel move. I struggled with this for hours before I eventually tried 8mm retraction, which solved it. The Bowden tube was about 650mm long.

Another of my printers has a 330mm Bowden tube and needs 4.5mm retraction. So my rule of thumb is based on experience of these 2 printers.

What I believe is happening is that when you extrude, the back pressure from the hot end causes the filament to compress by taking up a helical path inside the Bowden tube. When you retract, you have to pull the filament straight again before it is actually pulled out of the hot end. The longer the Bowden tube, the more the filament can shorten by taking up that heilcal path. So it stands to reason that retraction needed should increase linearly with the length of the Bowden tube. Then you need to add the retraction that you would need if there is no Bowden tube, which is normally reckoned to be in the range 0.5 to 2mm.

When I switched to Capricorn tuning, I was able to reduce retraction by 1mm, and with pressure advance enabled I reduced it by a further 1mm.

I went away from capricorn again cause my filament godt stuck in the tube once in a while, if i wasnt fast enought to pull the filament out when hot...
But i still use capricorn from the y splitter down to the hotend due to my i2k coupler and disc in the hotend is smaller than the diameter of a regular tube..

I only get a blob on startpoints after retracts endpoints are perfect, just need to be at 40-45mm/s retract speed..
Ive just always struggled with retraction.. always..
It was first when i got my 3mm olsson block i actually found a retract setting that worked perfect.. thats when i started to get back into 3d printing again.. i was thinking of selling all of it due to nothing would work just a tad as expected and i had more downtime than uptime..

Thats one of the reasons i went away from e3d cause they are so super sensitive to retract...

But that seems very valid to me...
I believe youre having quite a high backpressure when printing? Cause i see no difference in 195 or 210c with my trusted pla, except the shine its got..
And the higher the temperature, better retracts.. which has always been the opposite for me.. but i guess that makes sense, in some weird way i cant explain..

All the printers i have fixed for ppl i have based everything on own experiences just like yours with this..
Give me a stock cr10 and i will make it a mean machine thats comparable to a prusa..
We all have something were good and bad at.. and definately retraction has always been a pain for me...

If pla was just as easy as abs i would be super happy

[johnjohn1990]

@dc42

I am so gratefull for your input with this! 8mm exactly, removes the points absolutely perfect!

Though unfortunately, it underextrudes after retracts on small areas now... for example on the furnace of a benchy or similar smaller areas, with retracts in between..

[johnjohn1990]

This is how all my smaller consecutive retract startpoints look... look in the bqvk of the boats infill, you can see the startpoints, this also applies when its doings walls(infill are just easier to see)

Which is what i have battled all my time with e3d, just to point that out..

Lowering retract speed makes more blobs... setting up to 8.5mm and 9mm just makes under extrusion worse and does not improve blobs..
Higher temperature makes blobs smaller but underextrusion worse..

But my retract points are unspeakble good with 8mm and 40mm/s retract speeds! Cant even feel them, i cant even get my camera to focus on them..

The pla also says blobbing noises whenever it primes before doing those areas where it underextrude and it doesnt do overhangs that great now.. (kinda like too high temperature)..

But if i turn the temp down, it will just make the blobs even worse...

[bot]

What is the resolution of your E axis? IE, what are your E Steps/mm? I've found that increasing the resolution of the E axis helps with underextrusion at the beginning of moves.

You can do some simple calculations to see if this may be the case. You need to calculate the number of microsteps/full steps the extruder motor would move with a given layer height/extrusion width/distance combo. EG: 400 E steps/mm, 1.75 mm filament, 0.4 mm nozzle/extrusion width, 0.2mm layer height. At these settings, a tool path of .2x.4x1 mm long would take only 13 microsteps -- not even one full step of the motor. In some cases, the motor may not have the power when microstepping to move, so there could be no movement for longer than 1 mm of toolpath extrusion.

If you do this calculation, and you find that you have plenty of full steps for the distance you're seeing problems for, then you can disregard this as a possible cause of your problems.

[johnjohn1990]

@bot said in Linear advance, what to look for:

What is the resolution of your E axis? IE, what are your E Steps/mm? I've found that increasing the resolution of the E axis helps with underextrusion at the beginning of moves.

You can do some simple calculations to see if this may be the case. You need to calculate the number of microsteps/full steps the extruder motor would move with a given layer height/extrusion width/distance combo. EG: 400 E steps/mm, 1.75 mm filament, 0.4 mm nozzle/extrusion width, 0.2mm layer height. At these settings, a tool path of .2x.4x1 mm long would take only 13 microsteps -- not even one full step of the motor. In some cases, the motor may not have the power when microstepping to move, so there could be no movement for longer than 1 mm of toolpath extrusion.

If you do this calculation, and you find that you have plenty of full steps for the distance you're seeing problems for, then you can disregard this as a possible cause of your problems.

Why didnt we learn this in the school in the 90's lol

The bondtech uses 140steps/mm, so that gives 280 with my .9 degree 1.69A 65ncm omc motor.. its running at 1100ma, only gets about handwarm after a 24 hour print...

Ive always used this type of extruder without issue... its the same as the gears on a prusa mk3...

I know it works... but i have noticed that the teeth that drives the gears have started to wear a bit... i do not know if this could be cause.. that would cause the "idling" gear to not interlock properly and loose its tension..
And that is an issue that is rather new... i dont know if the clogged hotend i had and powerfull motor have stripped some of the teeth away.. there were metal particles in the extruder...

But never the less, 8mm with 35mm/s 2400 E jerk, 1200 xy jerk, 3000 acc on xy and e, seems to work pretty okish so far...

Also i lowered my temp all the way to 185-190, and that seems to be a good temperature for 60mm/s nice and beatifull .2 layers..
I have always been able to print with a very low temperature ever since i made the pt100 4wire instead of 2..

But what youre writing makes sense.. sort of.. cause then all 1.8 degree china extruders shouldnt be able to print as they do, once setup correctly...
My e steps are fine... depending on the filament i throw in it, i need anything from 0.9 to 1.0 in extrusion, so i believe it is completely fine it cant suddenly change out of the blue, at least it shouldnt..?

[johnjohn1990]

Should i try putting my geared extruder on? I can quickly swap it from my other machine..?

[johnjohn1990]

@bot

Oh my...
This is on the edge om embarassing..

When i took my bondtech apart it was full of metal dust.....

The teeth are completely smashed.. round on one side and heavily worn on the other

So i put my geared extruder back on, dialed up jerk on E, and it looks and sounds so much better now than before

[johnjohn1990]

Everything makes so much sense now..

I had a few clogs with some special filament so it might have taken its wear on the gears... then all my retraction testing... then my 240mm/s pa print probably took the last bit of life out of it..
They are after all pretty old those gears..

[SteveYYC]

It sounds like you are having the same issue I was having. I have a fairly long bowden tube and couldn't see any change in my pressure advance settings. It turned out the problem was I wasn't going far enough into the pressure advance values. I wound up needing to go up to 1.485 to get the improvement I was looking for.

To save time, I used a script-driven pressure advance object like this one:
https://www.thingiverse.com/thing:3442071

It uses post processing to increase the pressure advance for every Z layer so you can test a large range of values in just one print - whichever layer looks best you just measure height with a caliper to figure out the pressure advance value you need.

[johnjohn1990]

@steveyyc said in Linear advance, what to look for:

It sounds like you are having the same issue I was having. I have a fairly long bowden tube and couldn't see any change in my pressure advance settings. It turned out the problem was I wasn't going far enough into the pressure advance values. I wound up needing to go up to 1.485 to get the improvement I was looking for.

To save time, I used a script-driven pressure advance object like this one:
https://www.thingiverse.com/thing:3442071

It uses post processing to increase the pressure advance for every Z layer so you can test a large range of values in just one print - whichever layer looks best you just measure height with a caliper to figure out the pressure advance value you need.

Thats a pretty awesome one i will definately bookmark that!

Though i actually just printed my most neatly looking piece of a boat last night..

It seems that it was all due to a failing extruder so messed up i didnt see this earlier.. i feel so dumb...
But thats jusy how it is.. cant change what has happend

But i will play around with pa when i get it stable for more than this one filament

And thank you for your help everyone seems that a little bit of info from all of you made it all work!

Im currently running 8mm @ 40mm/s and its perfect.. when i cant feel retract points when i run my finger over them, i will say its approved