Page MenuHomeAleph Objects Inc

auto-cal reset/freeze m.145
Closed, ResolvedPublic

Description

I flashed a Redgum to 2.0.0.145, reset to factory settings, and attempted auto-calibration for backlash.
On the first attempt, after completing the auto-calibration I was left with a blank blue screen and the machine was unresponsive.
I power cycled the machine and checked to see if any settings had been saved, all values still read zero.
I reset and saved factory defaults just to be sure and then ran auto-cal again. This time after completing auto-cal, the printer just restarted and didn't save any settings.

Reverted FW to 2.0.0.144 and was able to complete auto-cal without issue.

Event Timeline

logan created this task.Jul 22 2019, 9:12 AM
logan triaged this task as Normal priority.
logan created this object with edit policy "LulzBot Hardware Products (Project)".

Okay. I'll look into this. However, .144 is the current FW for RedGum, so we should be good.

This seems to be caused by commit b2b7c583eb418d008c275d9279efdda0cc15ad09, but I cannot for the life of me figure out why. Will continue investigating. Just wanted to get this logged.

Nevermind previous comment. It seems like I don't really understand what a merge does. It looks like more code got brought in than just that commit.

This will be fixed in .146 and forwards.

Flashed the unit I have to m.147 just now and ran auto-calibration. Issue persists, printer resets upon completion and no data is saved.

marcio added a subscriber: alexei.Jul 24 2019, 1:00 PM

@logan: That's odd.

@alexei: Can you clear out the build dir on buildbot? This may be another one of those cases where stray files from a previous build is gumming up the works. When I do a fresh build on my machine, I don't see this issue anymore.

Okay. My fault entirely. I tested a fix and confirmed it worked. But that fix introduced a problem when I tried compiling for Quiver, which is a 32-bit boards, so I tweaked my fix for Quiver prior to pushing .147, but never actually tested it again on Workhorse. It turned out the tweak actually rendered the original fix ineffective.

Ultimately this boils down to me not really understanding the problem (and the fix) as much as I thought I did. So this will require further thought.

@logan: Could you retest this under .148?

logan added a comment.Jul 29 2019, 3:38 PM

Absolutely, as soon as my workhorse is done with the print it is running.

logan added a comment.Jul 31 2019, 9:05 AM

@marcio Tested under m.148, auto-calibration completes successfully and data is saved.

logan closed this task as Resolved.Jul 31 2019, 3:02 PM