Out of calibration, # of steps ? - need firmware setup help

[kristen 0]

I upgraded my printer with the direct drive modification and replaced the pulleys and belts with custom ordered GT2 parts, however my prints are now coming out a bit undersized. 98.5% of what they should be in the X and Y directions.

I'm looking at Ginge's custom firmware builder and trying to figure out if I need to increase or decrease the steps per unit. I'm guessing I should multiply the X and Y steps per unit by 101.52% to get me properly calibrated. (was 78.7402, changed to 79.9393)

http://marlinbuilder.robotfuzz.com/

Can someone in the know check over these settings (I do have a heated bed and an ulticontroller).

http://imageshack.us/photo/my-images/801/w4my.png/]

 

[illuminarti 18]

IIRC, I think that if you have GT2 pulleys, then, by definition, the steps per mm is 80 - which would be pretty close to your measured value.

You don't have to rebuild the firmware to change the steps per mm - you can do it from the Ulticontroller (Control -> Motion, I think) , or via gcode:

M92 X80 Y80 ; set steps per mm

M500 ; save to EEPROM

Then power-cycle the printer to make sure they get used. M503 should display all the settings back, to make sure they're being used.

If you want to build Marlin, then I think the settings you showed look ok, although I'd probably turn off the two 'software experimental' options.

 

[kristen 0]

Spot on !

The Ulticontroller proves awesome yet again. Control -> Motion ->X-steps/mm, then back one screen and "Store memory"

That is going to make fine calibration pretty sweet.

 

[illuminarti 18]

Just beware that depending on the version of Marlin you're using, it's possible that any changes to the steps-per-mm might not get picked up until you save the settings and then power-cycle the printer. I'm not certain about it, but there were some bugs about picking up and using changed settings like that, and I fixed several a few months ago. So, just to be safe, and to avoid going insane trying to test things, I recommend saving and then rebooting. (At the very least, you will have to explicitly save the changes you make in the UC if you want them to persist across reboots).

 

[calinb 11]

Yeah--I do scaling with xsteps and ysteps all the time using the Ulticontroller. There is no single "calibration" that works with all filament materials, slicers, models and processes (which I tune for the model), or even all bed orientations of asymmetric models. I rarely find it necessary to tweak zsteps in order to realize my accuracy goals but I have a log of xsteps and ysteps for nearly every model I print or plan to print repeatedly.

I've found that the test models I printed when I first obtained my Ultimaker were a waste of time. Now I only print (or partially print) tests of objects that I actually need. Generally speaking (and given that it squirts molten plastic), the UM is quite accurate right out of the box. Any further improvements in accuracy must be hard-won on a print by print basis, or at least extrapolated between very similar prints and processes.

Sure--when modifying some element of the UM design, as you have done, some calibration tests make sense, but I'd probably still print something that's useful too and take measurements from it.

BTW, one recurring challenge is to find a good accuracy compromise between inside and outside dimension on prints. Typically, settings are a trade-off between the two!

I often must be picky because I'm printing tools and commercial mechanical products and often attempting to hold dimensions within a very few mils. I'm not printing garden gnomes. Sometimes I use my machine tools to finish projects that I print, where I can often hold dimensions to 0.001 inch.

 

Spot on !

The Ulticontroller proves awesome yet again. Control -> Motion ->X-steps/mm, then back one screen and "Store memory"

That is going to make fine calibration pretty sweet.

 

[Daid 306]

Proper steps with GT2 belts would be 80, not 79.939