For anyone still using endstop switches...

mrehorstdmd

@LB These are cheapo $3 each opto sensors from China. You'll never know where any of the parts come from, and if you order them on different days the parts may come from different suppliers. Fortunately, these things are so simple that it doesn't really matter who makes the parts or assembles the boards. There's not much to go wrong.

These aren't like bed proximity sensors that are going to be sensitive to temperature or metallic composition of the bed or optical properties of the bed. It's an opaque blade blocking an IR light beam. It's hard to screw that up.

These are the ones I used in my printer and in the sand table: https://www.amazon.com/Biaobiao-Optical-Switches-Endstop-Comparison/dp/B07JMDLD84/ref=sr_1_9?dchild=1&keywords=optical+endstop&qid=1610988408&sr=8-9

Adamfilip

I have been having layer shift issues, and the idea of homing XY on each layer is something i never considered but is brilliant.

I just have regular switch endstops.

A Former User

@mrehorstdmd

I fast-read (not all and not slowly) your blog and I like very much your idea to have some possibility to have the printer print a layer and not only home after/before each layer BUT to "record" the values of deviation to tell the pinter-machinist if there is a problem in one of the axis. I could imagine this in for of e.g. an "error-evaluation-macro" (or call it what you want) or option so to e.g. once in a month or so take the extra-time that takes into account to see if belts or gears or who knows what has to be tightend, loosend, slowed, etc. pp. on the printer because the error is bigger then what one wants for the printer.

In contrast to a G28 between each layer what can be done in almost any slicer I can think of, this recording-feature would have to be part of the duet-firmware, because it has to compare should-axis-position with real-trigger-position and write that deviation down in a *.txt or *.csv file or so.

What do you think? Could this done by any sort of macro? (Maybe good to avoid a firmware-wish in the first attempt to get this working). I think it would be helpful to be able to once in a while check how precise the printer is!

mrehorstdmd

@Adamfilip If you have layer shifting problems, rehoming on every layer isn't the solution. You need to find and fix the actual problem and not put a bandage on it.

mrehorstdmd

@LB I don't know if the distance traveled during a homing move is captured somewhere by the firmware. You'll have to ask one of the experts.

My printer doesn't seem to get out of adjustment as far as I can tell, but maybe a CNC mill with a lot more force on the tool would benefit from such a check, especially if it is belt driven.

Adamfilip

@mrehorstdmd

Yes an I have tried

I have replaced steppers with longer versions with more torque
adjust current for the steppers (they are slightly warm)
added cooling on stepper drivers
Lubricated linear rails and checked for binding
checked belt tension
Checked Pulley screws
Prints are dimensionally accurate

I am running out of ideas on what to check/fix to reduce the likelyhood of layer shifts.

so now im onto ideas like homing every layer or adding cheap S42B closed loop controllers.

mrehorstdmd

@Adamfilip In addition to all the things you listed, layer shifting is often due to too high jerk and/or acceleration settings. I have also heard of a few cases caused by too much mechanical force required to pull filament from the spool.

There are probably hundreds of posts in this and other forums on the topic of shifted layers.

Does the machine use steel core belts?

Adamfilip

@mrehorstdmd

My machine is not setup for speed.
I typically print at 40-60mm/sec

here are my acceleration settings

M566 X600 Y600 Z20 E800 ; Set maximum instantaneous speed changes (mm/min) JERK
M203 X9000 Y9000 Z1500 E10000 ; Set maximum speeds (mm/min)
M201 X1000 Y1000 Z250 E5000 ; Set accelerations (mm/s^2)
M906 X1800 Y1800 Z1500 E1300 I30 ; Set motor currents (mA) and motor idle factor in per cent

I dont use Steel belts., I once tried them but i use regular 6mm GT2 belts.

mrehorstdmd

@Adamfilip The acceleration and jerk look reasonable. X and Y motor currents look a little high- what type of motors are you using? Sometimes excessive motor current can cause bad behavior...

Is this a corexy machine?

When it shifts is it always at 45/135 degree angle? That would indicate one motor is not turning...

A Former User

@Adamfilip

Do you have "external" steppers -> In that case the cable that transmits the stp/dir/en pulses could be the problem. If only one of the pulses is missed you loose something oviously.
Microstepsetting too high -> the ABSOLUT-torque for 1 full-step get´s more and more divided with microsteps into incremental-torque-steps. If your incremental torque drops below of what you need in "extreme"-situations you loose a step. This is realted to the next 3.
Edit: already written down here: Check accel/jerk

Have you checked also the above?

Adamfilip

I have recently upgraded to stronger 60mm long Nema 17 with more torque, and 2AM max current, thats why it looks high

previously I was using aprox 1000ma

I dont recall if it was diagonal or not, will examine closer on next occurrence
yes machine is CoreXY

using Built in stepper drivers of Duet2Wifi