# Hi! Can someone explain to me in simple words,

Hi! Can someone explain to me in simple words, what junction deviation is and/or what it does? Or is it just the SAME as jerk in marlin/repetier?

@Rene_Jurack If youâve read grblâs explanation, Iâm not sure how much better to explain it.
In the simplest terms I can think of : higher junction deviation means the machine shakes more, lower junction deviation means the machine shakes less.

NaaaâŠ Not THAT simple, please Ok, maybe I need to be more precise: I did read the explanation and the way I understand it is, that e.g. if the printhead has to move a 90Â° corner, it would âtake a shortâ and therefore I get no perfect 90Â° cornerpath but a âcurveâ. Or am I completely wrong here?

@Rene_Jurack No, it always follows the exact right path. But depending on how steep the change in direction is, junction deviation determines by how much to slow down. So itâs pretty much how_much_to_slow_down = change_in_direction_angle x junction_deviation.

Ahh, ok, thx! So it is the same as âjerkâ in Marlin/Repetier. Does it have a unit? Or how do I interpret the figures? I have a very rigid and fast printer, acceleration of 10.000mm/sÂČ and more is possible, and now I want to adjust junction deviation as well. Standard setting was 0.1 but I donât know if I should try in 0.1 steps or in bigger ones like 10-20sâŠ

Itâs not exactly the same, but it essentially does the same thing, just differently. It doesnât have a unit, itâs an arbitrary ratio.

Ok. So it depends on the âchange_in_direction_angleâ how high I could set this parameter? Like printing a big circle could work with a high setting, but with printing a triangle could result in loosing steps with the same high setting?

Pretty much. Just try different values and find what works for you. You can explore by changing orders of magnitude ( divide by 10, multiply by 10 )

Ok, thank you very much for the explanation! One last thing: Where is junction deviation generally settled? Letâs speak like a rule of thumb for non specialized printers, e.g. some simple mendelmaxes or standard coreXY/H-BotsâŠ Is it more like âgo search between 0,1 to 3â or more like â10-50â?

Very very fast and rigid printers do values around 0.1 to 0.5, CNC mills do values around 0.001 , or 0.01 if very very rigid

Thank you, @Arthur_Wolf . I will try some settings this evening. IIRC, currently it is set to 0.3 without issues. Is there a test-STL for testing this or should I try a big triangle?

@Rene_Jurack Anytihng with plenty of sharp angles will work fine.

RenĂ© please go on asking such questions here. Soon I will start tweaking config for my delta and then I do not need to ask anymore

currently testing 10.000mm/sÂČ with 0.5 junction deviation @ 100mm/s. Seems to work. Am I right, if I say: the higher the junction deviation, the better the printresults? Given, that the printer can handle high junction deviation in terms of no ghosting / no ringing / not loosing steps etcâŠ

The higher the junction deviation, the faster the prints. How that impacts print quality is going to depend on the machine a lot.

So, currently printing at 100mm/s / 10.000mm/sÂČ and junction deviation of 0.5mm with no issues so far. Print-quality looks better, but sharp corners are now a bit ârounderâ than with junction deviation of like 0.1 or so.