bad bed adhesion on one side after rff3 upgrade

[sungod3k]

Hi,

I attempted a large print to that goes to the limits of my deltas bed. This was the first big print since I upgraded to rff 3 but it's possible that this issues existed before that.

I have a normal kossel style delta and use the rff configurator which works great. The g32 results mostly come out great (Calibrated 6 factors using 16 points, (mean, deviation) before (-0.147, 0.076) after (-0.000, 0.071)). I usually don't use mesh level correction, but looking at it now I cant see any trend just some small up and downs.

With this big print I'm noticing that one side (right corner) is much lower than the opposite site (left corner). Over the distance of 20cm id say at least 0.3mm. So much that one side gets smushed into the bed while the other side is so high up that it's not connecting correctly.
The model is laying on the bed in a way that the lowest corner points towards one of the towers.

The machine is not the most accurate construction but not that bad that g32 wouldnt be able to correct for it.
I wonder how i could correct for that, or if rff3 corrects slightly different than rff2 did?

[Phaedrux]

I'm not aware of any changes in RRF3 that would account for that.

Can you post your config.g and config-override.g?
Can you post your homedelta.g and bed.g?
Can you post the entire gcode file or at least the first 50 or so lines of it?
Can you send M122 and post the results?
Can you send M98 P"config.g" and post the results?

[sungod3k]

@Phaedrux said in bad bed adhesion on one side after rff3 upgrade:

sure
homedelta.g config-override.g config.g bed.g FirstLayerTestCircle.gcode M122.txt M98.txt

[Phaedrux]

Do you use mesh compensation? Do you have an example heightmap image? Does it show a tilt?

[sungod3k]

@Phaedrux

Havent used mesh bed levelling in some time now because it always worked without it.
I did some test with mesh bed activated, but it didnt make a difference.

[Veti]

M308 S0 P"bedtemp" Y"thermistor" T100000 B4138 ; configure sensor 0 as thermistor on pin bedtemp
M308 S1 P"e0temp" Y"thermistor" T100000 B4138 ; configure sensor 1 as thermistor on pin e0temp

your thermistor setting is incorrect. find the correct beta value in your thermistor documentation.

[Phaedrux]

Can you show the heightmap on a more oblique angle? Dead on makes it hard to see the surface relief.

[wummelfrosch]

which sensor (Z-probe) is used ? maybe this is the fault ? I had such a similar problem with the optical sensor of DC on the green gfk print surface. The values were sometimes good and then again bad and partly implausible ! Only after switching to BlTouch did I have reproducible values.

[sungod3k]

@Veti youre right. i corrected the beta, but i dont see a difference in heated behavior

[sungod3k]

@wummelfrosch i use the smart effector, has served me well so far. and reproducibility is good as far as i can judge it.

[Phaedrux]

http://boim.com/DeltaUtil/CalDoc/Calibration.html

See if you can identify any issues by comparing your heightmap to these.

[sungod3k]

@Phaedrux

[sungod3k]

@Phaedrux yes, i remember this post. it looks like the examples where one tower is different from the others. but endstop adjustment during g32 should be simple enough, or what could throw off the calibration algo that much?

[Phaedrux]

I'm afraid that's the extent of my delta knowledge.

Best I can suggest is to go through the setup and calibration guides carefully and ensure the build is accurate.

https://duet3d.dozuki.com/Wiki/ConfiguringRepRapFirmwareDeltaPrinter
https://duet3d.dozuki.com/Wiki/Calibrating_a_delta_printer

Overall it appears to be fairly level, but oddly bumpy and maybe bowled a bit.

[Veti]

that looks a bit like a ripple effect that is caused by backlash.
check the belts and pulleys.

[Phaedrux]

Yes maybe backlash. @sungod3k can you provide a screen shot showing an isometric view?