Shrinking on Y axis

alangibson

Hello all.

I just did a 200mm wide print that ended up measuring only 192mm wide. The part has several holes along the width. I measured their deviation from what they should be and the error seems to accumulate along the width. It's not just missing 8mm, it really 'shrunk'.

I know my rod lengths are perfectly equal (ordered from Seemecnc), so I'm already out of ideas as to what could be causing this. Does anyone have any suggestions of what I should check next?

My board is a DuetWifi and my Z probe is a Duet IR sensor.

Phaedrux

I'm assuming this is a delta?

Post your config.g, homedelta, config-override, bed.g, and the results of M122 and M98 P"config.g" please.

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

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

alangibson

Yes indeed, it is a delta. I've attached the files you asked for.

I've had some success dialing it in by setting the axis steps/mm with M92. Right now, my 200 mm test shape is 202. So that's progress. The weird issue I'm seeing at this point is that I can never hit 200. I'm always at 198 mm or 202 mm.

Thanks in advance for your help.

M122.txt
M98.txt
homedelta.g
config-override.g
config.g
bed.g

Phaedrux

Rather than change the steps per mm, which should be mechanically accurate and calculatable from the physical states, you could use axis scaling to make the correction in firmware.

https://duet3d.dozuki.com/Wiki/Gcode#Section_M579_Scale_Cartesian_axes

However, I suspect that something is off mechanically. Loose bolt, stretched belt, binding carriage. Something like that.

Did the top of your M122 get cut off? I was hoping to see the firmware version string in there as well.

G31 X0 Y0 P465 Z0.333
You're missing the X Y offset of the probe from the nozzle.

M558 P1 You're missing some additional parameters here as well. Hopefully the defaults are suitable.

Macro file "CONFIG.G" not found.

Can you try lowercase config.g? M98 P"config.g"

alangibson

@phaedrux

Rather than change the steps per mm, which should be mechanically accurate and calculatable from the physical states, you could use axis scaling to make the correction in firmware.

Thanks. I'll set everything back where it was and give this a shot.

"On a Cartesian RepRap you can get prints exactly the right size by tweaking the axis steps/mm using the M92 G Code above. But this does not work so easily for Delta and other RepRaps for which there is cross-talk between the axes."

I guess that would explain why I can't seem to hit 200mm.

However, I suspect that something is off mechanically. Loose bolt, stretched belt, binding carriage. Something like that.

I noticed that I had removed the belt tensioner springs at some point and forgot to replace them. I put them back on but they didn't make a difference. Everything else seems solid.

I just printed out an 8 pointed asterisk shape that is off by the same amount in every direction. That makes me think that it really is a scaling issue somewhere.

Did the top of your M122 get cut off? I was hoping to see the firmware version string in there as well.

M122 is complete. My (ancient) firmware version is 1.17e (2017-02-10)

You're missing the X Y offset of the probe from the nozzle.

I'll add this.

Can you try lowercase config.g? M98 P"config.g"

Still nothing.

M98 P"CONFIG.G"
Macro file "CONFIG.G" not found.

It's echoing upper case, but I'm definitely entering M98 P"config.g"

Phaedrux

@alangibson said in Shrinking on Y axis:

My (ancient) firmware version is 1.17e (2017-02-10)

Ah ok, that explains it a bit.

alangibson

@phaedrux

I've added the following line to the epilogue of config.g and I've now got an exactly 200mm test shape

M579 X1.03 Y1.03

The odd thing about this is that the math says the adjustment should be 1.015 (1.015 == 200/197). But using 1.015 makes no difference in final size at all.

Thanks for your help.

o_lampe

@phaedrux said in Shrinking on Y axis:

@alangibson said in Shrinking on Y axis:

My (ancient) firmware version is 1.17e (2017-02-10)

Ah ok, that explains it a bit.

But it doesn't explain why the off-scale prints kept unnoticed for so long?
@alangibson what did you set as 'rod-length' ? The actual rod length or did you add the diameter of the Mag-balls too?

alangibson

@o_lampe

But it doesn't explain why the off-scale prints kept unnoticed for so long?

I normally print small things where dimensional correctness doesn't really matter, so I just never noticed it most likely.

what did you set as 'rod-length' ? The actual rod length or did you add the diameter of the Mag-balls too?

I'm using the 290 mm ball-cup arm kit from Seemecnc. I used their published center-to-center value of 290 in my config.

https://www.seemecnc.com/products/290mm-ball-cup-arm-set-of-6pcs

o_lampe

@alangibson

Did you ever 'measure' the length yourself? Just trying to think about everything...

alangibson

@o_lampe

Did you ever 'measure' the length yourself?

I did some smallish calibration shapes (20mm I think) a very long time ago. At that size the problem is not visually recognizable and I would have considered deviations less than 1mm to be OK for my purposes.

It's only when I recently got into printing large center punching templates that I got to a place where accuracy matters enough for me to even notice the problem.