search

MarlinFirmware / Marlin

Marlin is an optimized firmware for RepRap 3D printers based on the Arduino platform. Many commercial 3D printers come with Marlin installed. Check with your vendor if you need source code for your specific machine.
https://marlinfw.org
GNU General Public License v3.0
16.17k stars 19.21k forks source link

[BUG] Anycubic kossel Linear Plus is much to slow #15466

Closed CorFlotscj closed 4 years ago

CorFlotscj commented 4 years ago

Hello,

i update my Anycubic kossel Linear plus with the Marvin firmware and used some different Config.....h from thinkervers as setting. the general problem is that the printer seam now to be much slower. I'M usind cura for slicing, which with the standard firmware was alway close to printing time, but now prints take 1.5 or even twice the time that cura expects.

(xyz cube with 3 shell 0.2 layer 20% infill, Cura ~22min, real 37min)

in increase acceleration and jerks already with no effect. (auto leveling and al the other suff works very well, just the speed is very bad, can someone give me a hint what i can still tweak to get it inline? best regards Martin

i asked this question in a lot of other places and dind got anything usefull so i thoght i got to the roots...

AnHardt commented 4 years ago

Try less segments/second and/or a larger planner buffer.

hapklaar commented 4 years ago

Is the print really taking longer than before, or just longer than what Cura predicts? You can enhance this by enabling acceleration control and jerk control, so Cura can use these values in its prediction.

CorFlotscj commented 4 years ago

Hi again, i tryed the tipps segments/second :

larger planner buffer. set was 16/16 test with 32/64 (more war not possbile, even with those i was warned that is could be instable

best results were 36,5 min from 38min -> no real improvment

i also flashed back to standard Anycubic 26,3 min with the same gcode. with same setting Cura 3.6 and 4.2 are close to gether 25 to 26 minutes. but there fo i got some ghosting :\

is there may be a mismatch of units for accelration and speed?

edit: what also was noteble is that the sound when moving from home to buildplate is much louder with the Marlin firmware. didn't stop it but from feeling the time should be roughly equal is there some setting about the stepper motors that could have an effect?

CorFlotscj commented 4 years ago

in the meantime i've gone thru the conficuration.h and found a solution for me

it with the new way /calulation of the DEFAULT_AXIS_STEPS_PER_UNIT i've modifined with below setting, back to 80/80/80/96 as hard setting and not calulations. and got the predicted 26min...

but i would have expecet if there are problems with this also the sizing should be effect but this was very good with Marling compared to before

maybe someone with more inside then me has an idea...

/**

Glodigit commented 4 years ago

I've noticed this too. I put 120mm/s speed for all values in Cura but doesn't exactly seem like it's going at that speed.

AnHardt commented 4 years ago

i tryed the tipps segments/second :

160 (wish were from the profile) to 200 (standard) to 300

No. Wrong direction. Less not more!

There is no Marling-Firmware it's "Marlin"

If you don't see a change in size of the part after flashing new steps/mm-values you do have EPROM active and did not reset that to the new numbers. Probably all your measured numbers are meaningless.

han9988 commented 4 years ago

I had some unpredicted results when I changed segments number on my delta so I thought the segements is still a work-in-progress. What do you think @thinkyhead ?

boelle commented 4 years ago

@CorFlotscj still an issue? did you remember to overwrite eeprom? ie send M502 and then M500

sjasonsmith commented 4 years ago

I recommend that you slice a file, then use the exact file with both the old and new firmware. Time the print with both firmwares, to see if there is really a performance change. It is hard to act on "it feels slower" without having some solid numbers to compare.

There are so many variables that could impact performance, that without printing the same file with both firmwares, it is hard to know whether something in Marlin changed, or if it is something about the file/slicer.

I have a Kossel LP, but haven't used it for a while. With the Aug 24th state of bugfix-2.0.x I could get it moving much faster than the machine was rigid enough to handle.

CorFlotscj commented 4 years ago

Hello, i do have it still/again. For some time the print timing was back to that cura expected

But since yesterday its back to too long. I did have this with my setting from above, and another try with 80 segments per secound, but the "new" step calculation. (basical what AnHardt suggested). both worked fast.

But i had some issuses with the extruder setting. so if change the speps /mm for Extruder and did auto leveling, as i had also not the best with a nother test print with bad adhesion and to much material extruded.

after this adjsutements timing of the same file took longer. ~1.5 times londer

today i reflashed marvin fully with als levelung (except the Autoleveling, only those inside Delta calibation) and beeing still only at ~50% real speed.

flashed the AC standard "beta" firmware and the print is as fast as exepcted....

can board issued cause this? that there are some physical damages? or that some calculations with the leveling take to long? also during of the "test belt" the display "fizzeled" and showed only "nonsence"

I'm very confused.

BR Martin

AnHardt commented 4 years ago

today i reflashed marvin

Whatever marvin is - go there and ask for help.

Miss typing "Marlin" two consecutive times is looking careless to me. If you are that careless with your Configs and observations i don't wonder about anything.

End of communication.

Glodigit commented 4 years ago

I'd test it too if it weren't for bug #15495.

boelle commented 4 years ago

https://github.com/MarlinFirmware/Marlin/issues/15495 was fixed 4 days ago

so maybe there is news here too?

Glodigit commented 4 years ago

Can confirm. 120mm/s looks so much faster now.

boelle commented 4 years ago

oki, will close this one as "fixed" then

we can always reopen

github-actions[bot] commented 4 years ago

This issue has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.