Nozzle Drags on Print post Calibration

[davec3275]

Okay, I did some digging and have yet to find my particular issue in-depth, so here goes. I recently used OpenDact to calibrate my Max V2 and now as I print 2 peculiar things happen. First, there are moments when the nozzle sounds as if it is grinding or dragging on the print while moving a long distance (from one corner of the model to the next. This has happened on a few occasions with different prints. It is quite unnerving.

Also, since calibration I noticed some issues where it looks as if an extra layer is placed in the path of the nozzle on a previous layer that causes the head to hit it and "bump" over it. I will say I have NOT calibrated the extruder quite yet, so it could just be a remnant/drip.

It took 21 tries to get OpenDact to mark the print "good". One of the last set of changes was to change my Steps per MM from 80 to 81.1974. I used the latest OpenDact in manual mode with a digital dial gauge (had to invert the numbers since as the plunger was depressed the numbers on the display shrunk - after zeroing the gauge, when the arms were homed the dial read -18.9).

Any help would be appreciated.

[Polygonhell]

You should not have to change the Steps per MM, it's a function of the belt geometry and the number of teeth on the pulleys.
Changing it just scales everything.

You can add some ZLift to avoid hitting the print during a travel move, but unless you are over extruding, it really shouldn't be much of an issue.

[davec3275]

Ah, okay. I'll change it back then.

But if it causes issues like that, why is it a recommended setting change during calibration using OpenDact? Or am I missing something?

[Polygonhell]

Can't comment on why open DACT would suggest changing it, I haven't used it, and stopped following the discussion at some point.
I also don't agree with changing diagonal rod length, unless it's a very small amount from the nominal value.

Both of these numbers are physically constrained by the selection of parts, and can't be changed during assembly.
The Z offsets of the towers obviously varies based on the end stop screw positions, the delta radius will vary with assembly to some extent and the tower rotational offsets are heavily influenced by the quality of the build.

I suspect the first two values are introduced as variables to try and accommodate other errors (tower lean etc.) that cannot be directly corrected in the firmware, the problem is even if introducing them gives a better mathematical solution, the solution is only based on consistent Z position in a single plane, and that's insufficient since all of the variables effect all three axis. IMO your better off only changing variables that directly equate to variations in the build.

[mhackney]

+1 Polygonhell - I seriously do not advocate changing the steps per mm of the motion components (you DO have to calibrate this for the extruder - see below).

+davec3275 - unfortunately, I think OpenDACT has complicated things enormously by mucking with steps/mm for the axes. I would highly recommend learning how to use David Crocker's online calibration. It is identical to the auto calibration built into his RepRapFirmware branch (and there is now a Smoothie implementation of it) and it WORKS.

Even wit perfect extrusion, a little Z lift for rapid travels is a good idea - it has no real negative side effects and does prevent brushing against the inevitable imperfections (like a little "hershey kiss" where the extrusion ended).

It is important to calibrate the extruder because it is nearly impossible to measure the perimeter of the cogged gear and this is a vital measurement. There is also enough variability in the machining of these and the filament diameter and hardness that calibrating it is essential to get good results. Luckily, it is also very easy to do.

[davec3275]

Thanks for the information and suggestion. I'll reset everything back to default and then give this other calibration method a try.