How to fix layer shift caused by jerk of the print bed?

Hi,

It took some time for me to get familiar with RepRapFirmware gcode styless and DWC usage.
I like Fisher's performance and printing quality.

Metal Fisher in action. It's still in work in process stage.
All frame parts are laser cut steel plate, painted in black.
Aluminium print bed, anodized in orange.


Tower of pi
[youtu.be]

Moai glasses holder
[youtu.be]

I have two major issues I need to solve.

Firstly, auto calibration creates a tilted print bed. Y tower side is 0.1 to 0.15 mm higher than X and Z tower roughly.


Three or four iteration of four points G32 produces deviation within 0.02.
The strange thing is calibrated print height (H) always shorter than actual distance between the nozzle and print bed.
Adding 0.3 to 0.4 mm using M665 fixes it.

Do I add Z probe hight to H after G32?
Below is config.g's lines.

;// Z probe and compensation definition
M558 P4 X0 Y0 Z0 H4.2   ; Set Z probe type and dive height
G31 X0 Y0 Z-0.3 P200    ; Set the zprobe height and threshold (put your own values here)

I suspect there is imblance in six ball joints of the carriages because I've replaced three types of diagonal rod since I bought Fisher Beta.
Original rod -> laser cut aluminium rod -> igus ball end rod -> back to original rod.

Secondly, it seems the spring force is not too strong enough to press the bed against three mounts.
It often generates subtle or outstanding layer shift in middle of printing, which affects overall printing quality.

Thanks in advance.

Edited 1 time(s). Last edit at 09/23/2016 10:06PM by janpenguin.

I replaced the standard fisher bed with an aluminium and glass bed with heater and because of the added weight needed to replace the springs to fix the contact issue.

The g31 command sets the offset for a height. By the time the z probe reads 200 the nozzle will be lower than 0 so this offset corrects that. You probably need to improve the bed contacts before changing that as I suspect that's the source of your problem.

I'd also recommend you try the 6 factor calibration it seems to give more reliable results.

---

DC42 Kossel 330mm x 2meters
My Thingiverse Creations

@DADIY

The aluminium bed was a conduct that three balls share one resistance path.
When I checked the surface resistance of the aluminium bed, it was infinite. I never suspected the contact of three holes where the ball seats which I drilled after painting the bed.

After rebuilt Fisher and checked all the wires, I was checking resistance of three balls of the bed, all the ball should be disconnected but they were close to zero ohm by multimeter reading. I used black electrical tape as non-conductor so that the ball would touch exposed aluminium wall in the hole. To reinforce the bed springs, I used ball pen spings.

I found the right Z probe setting for my Fisher.

;// Z probe and compensation definition
M558 P4 X0 Y0 Z0 H4     ; Set Z probe type and dive height
; G31 X0 Y0 Z-0.3 P200    ; Set the zprobe height and threshold (put your own values here)

G31 X0 Y0 Z-0.1 P1000    ; Set the zprobe height and threshold (put your own values here)