Hey! This thing is still a Work in Progress. Files, instructions, and other stuff might change!

Marlin Junction Deviation Calibration toolkit for Cura

by DrywFiltiarn Mar 2, 2019
Download All Files

Thing Apps Enabled

Please Login to Comment

Printed this one from .01 to .1 on my s5, but I really don't see any difference?

Anyone able to get this to show any difference when using Prusaslicer 2.1? I've used the settings described to get the single wall, no top, and printing at 60mm/s. Even ensured that min layer time was set down to 1s, so it would not slow down the print. I've manually added the incremental M205 Jx into every 20th layer, and can see that it increases on the LCD, but I'm not seeing any difference in print quality. Even went up to .2 on my ender3

Could it possibly have something to do with how Prusaslicer generates Gcode when set to Marlin output?

I can't see any difference either - and I'm using Cura. As you've mentioned, LCD updates, but no visible difference in quality of print in the different sections. Out of curiosity, do you have linear advance enabled?

No linear advance set for me

Hi there,

So i got a question, i configured everything as per your instructions but i got a question, in marlin your can define the Junction Deviation default value "#define JUNCTION_DEVIATION_MM 0.02", so my question, you state that you have to enable Junction Deviation by un-commenting #define JUNCTION_DEVIATION but do we also have to set a default value? if we set a default value does it get overwritten by your script during printing?

I am asking cause i have set a default value of 0.02 in marlin 2.0 and then printed your test file by using the script but to me the test file looks the same in every step (start value 0.02 with 0.005 increments).

So i am not sure if the script is overwriting the default marlin value.

Ive been playing around with this lately. Neat tool, thanks for the work Dry!

Kotik, if you havent, you need to check min layer time or min speed so it doesnt slow it down so much that nothing makes a difference.
I noticed it was printing way slower than my target speed so set min layer time to 1sec to basically shut it off.

Many people make the same mistake on the temp towers. You often see people with the 180c section looking the best.
It may look the best at the 10mm/sec it printed at due to the min layer time cap, but things will prob be quite different when they print something larger that actually runs the speed they set.

Im running the latest 2.0 bugfix and there are limits on the setting, looks to me 0.01 to 0.3

Send: M205 J0.32
Recv: Error:?J out of range (0.01 to 0.3)
Recv: ok

So my test print aborted.

The script defaults to Starting Value of 0.02 and Stepping size of 0.05, but should be a stepping size of 0.005

Hi, nice work! We have to know your acceleration settings, because from what i've seen the JD function it's based on acceleration too. Thanks!

What acceleration values do you use? Which ones should I use for this test? Ringing shows for me even at JD 0.025 with aluminum tensioners and tightened belts.

Do I need to disable jerk/acceleration control in Cura?

Would advice to do so anyway personally, as Cura’s overriide values tend to be higher then firmware default values. Personally I prefer to do these configs purely in the firmware.

Do keep in mind you do need to update firmware on your machine with a recent Marlin with JD support and you also need to activate this.

This seems like a nifty tool that I will definitely test out one evening!
Just wondering about linear advance. Is it best to turn that off, will it interfere with the results if it's on, mask them sort of?

I'm currently running a set of 4 of these tests using 50/60/70/80mm/s printspeeds and NO linear advance. From what I see so far after completing 2 of the prints, without the linear advance there is always corner bulging taking place no matter the junction deviation set it seems, investigating further, Might also alter the model to test different angles of cornering then just the -90 and +90 degrees to see if that can give different results that can be more conclusive on what JD value would be best.

I have used this to calculate my value, and so far it seems to be a good starting point at least.

I have done some testing on a normal square cube, with LA on, and I had to alter the JD value quite a bit to actually see any difference, so I also suspect LA to affect the results.

Comments deleted.

Linear advance does partially impact outcome I think, as that also does some magic for bulging corners on the object being printed. If you can I would say try it with both linear advance on and off and let me know if you see any difference in the outcome. I will also test this for myself. This first version is just a starting point really. It will be improved based on user input (and I hope also with some feedback from the Marlin crew) and further testing I will be doing myself.

I do see options to add a configuration setting to enable or disable linear advance for this test. For this moment it will need to do additional research and testimg how certain things relate. Also the impact of printspeed and acceleration (or even s-curve) are factors I will have to look into further. Although these factors are more or less covered as you have the freedom to select your preferred speeds and accelerations for doing this test.