Oversized objects after upgrade to smart effector

Nuvag

Printer: Kossel XL from Think3DPrint3D with all metal parts (except carriages for v-rollers).

When printing a 45mm cube the actual size was 44.6, now when using the smart effector its 45.5.

I replaced the robodigg metal effector and Traxas arms with the smart effector and magnetic arms, additioally i didn't reuse the old E3DV6 but installed a new one.
Did not change the carriages or the frame, but had to uninstall the bed for wiring and removed the black paper between aluminium heatspreader and glass printbed.
First setup was autocalibrated by using the mini ir probe, now using the smart effector.
Delta radius was reduced from about 177 to about 171.
Changed from a E3D Titan to the zesty nimble, but tuned extrusion so its identical to before, so its not caused by massive overextrusion.

I wonder what could cause this rather large change and how i can get it as close as possible to the requested size?
I know i could fix this using Slicer settings, but would expect that could and should be fixed in firmware (?)

srs5694

Chances are your delta arm length setting is wrong – it's just a little bit too long, if I recall the relationship correctly. It's my understanding that auto-configuration systems can't detect this error, so it's important that you get it right manually. This is set via the "L" option to the "M665" command. If you're using a Duet controller, you'd set this in config.g (although it can be overridden in config-override.g, if your config.g is set up to load it, so you might need to set it in both files to be sure). In some other firmware systems, you'd edit the firmware configuration file, rebuild the firmware binary, and re-upload it.

dc42

Suggestions:

use 6 or 8 factor calibration, not 7 or 9 factor
make sure the L parameter in M665 matches the length of the rods as written on the stickers
check the tower steps/mm, it affects XY part scaling as well as the Z scaling
if necessary, use the M579 command to adjust the scaling

Nuvag

Thanks for your input.
L parameter was off, i was mistaken in the Delta radius beeing the same as rod length and would be auto calibrated.

Now its 44.8 when commanded to do 45.
I'm currently using 10 factor calibration, would reducing it to 8 factors be adviceable?

dc42

RepRapFirmware doesn't support 10-factor calibration.

Dougal1957

Nuvag think you may be confusing number of points with the calibration factor which is set by the S Parameter on the very last G30 line in you bed.g file and that needs to be set to either S6 or S8 8 is my preferred option.

Doug

Nuvag

You are right, i was mistaken its 10 points 6 factor, sry.

Dougal1957

try 8 factor and you could increase the number of points as well I use 16 on mine or whatever the max that Davids Bed.g wizard gives

Doug

ps S8 also compensates for tower tilt

Nuvag

16 probing points instead of 10 -> no measurable difference
8 factor compensation -> no measurable difference

Ended up using M579 for scaling which fixed the issue for now, thanks for your help!

Dougal1957

as long as you applied the M666 and M665 then the M579 is good to get to your goals lol