Latest Firmware for TAZ5 - advanced development version

Development Software
41

There has to be errors, especialy when enabling the checksum for error correction fixed the one problem… Or you changed the object and the errors where only on one special gcode (for example if it has curves with many points).

The noise may be OK, as long as the print looks good. I will dryrun your gcode tomorow morning and see what happens. When you unload the filament check for backlash in the extruder gear by moving it by hand. Check every 90° position of the big gear. If it’s more than a very tiny amount that could also cause the noise.

42

Sebastian,

Thank you for helping me get to the bottom of this. I appreciate the time you are taking.

Let me know what the results of your dry run are. In the meantime, I will check all you suggested and get back to you.

J

43

I did the dry run, now I’m sure the noise is due to excessive slack in your heringbone gears. I get the same noise when I run your print without filament, so there is no load to dampen the hard steps of the extruder. As soon as I apply some “breaking” by putting a finger in the nut of the extruder bolt, the noise is gone.
This will not harm your printer, but it will increase the wear of your gears. Do the “wobble” test as I described before, and if there is some play, readjust the position of the stepper by loosening the 4 screws (easy when you take of the extruder from the mount plate). After loosening them, press the small gear to the big one with one hand and tighten the screws simultainously. Recheck the play by turning the gears by hand, repeat if necessary. Have in mind that printed gears are nevery perfectly round, so there might be more loose pisitions. I tend to search for the loosest position and clamp the stepper to eleminate the slack in this position.

44

Sebastian,

Thank you so much!

I looked at the extruder stepper, and man- one of the bolts was totally loose, and there was a lot of play between the stepper and the big gear…I pressed them together and that fixed it - mostly. There is still sometimes a less-pronounced version of the vibration. Maybe, as you said the gears are out-of-round and don’t contact with each other perffectly in all arrangements.

I really do appreciate your help. I’m going to give the firmware a real shot and I’ll report back.

Best,
J

45

Sebastian,

So far, I am liking the firmware. Seems to work nicely.

I Noticed 1 thing - someone else on here mentioned the same thing a few days ago:

If I use my standard software retraction, I get fairly sizable gaps at the start of un-retracts. Almost like it isn’t unretracting.

When I switch to FW retraction (by replacing all of my gcode retracts with G10/G11 commands, the prints are perfect. For my specific settings I use this post-processing code:

{REPLACE "G1 E-0.8500 F1750\n" "G10\n"}
{REPLACE "G1 E0.0000 F1750\n" "G11\n"}
{REPLACE "G1 E0.0000 F525\n" "G11\n"}

So, I’ll use FW retraction which is no problem. Just curious why you think this might be happening when using SW ret?

Best,
J

46

That’s quite strange… Even with FW retract, the retract/prime movements are executed in the same way as it would do it with G1 codes. If you have an object were the problem happens repeatable at the same place, it would be very helpful if you could post the isolated gcode part where it happens (the retract ±5 code lines).
I will think about the problem, but to be honest I have no real idea at the moment… It may have something todo with the seperate extruder ISR.

47

I tried the firmware, and I really like the improved top surface finish. In the end, it was back to the LB default firmware due to the same issue with gaps or thinning of the layer after when extruding after the retraction. Turning off retraction resulted in great prints, but I started getting spider webs/hairs when moving over open spaces. I experimented with the retraction amount, and even .5mm left gaps in the side… I might give it one last try with .1mm retraction.

It may also be my extrusion temp of 245C… could probably get away wtih a few degrees lower for the ABS.

48

Kcchen, Sebastian-

Try using the firmware retraction and let me know your results. I found the same as you - using SW retraction, even set extremely low led to gaps.

I am including three files.

1-An STL of a test I made up
FirmwareTest.STL.zip (139 KB)
2-gcode with Firmware retraction
FirmwareTestFWretGcode.txt (512 KB)
3-gcode with Software retraction
FirmwareTestSWretGcode.txt (525 KB)
Using the advance firmware for both. K75, PLA. All the specs are in the headers of the gcode files.

Best,
J

49

I can reproduce your results. First 2.5mm of the walls are printed with FW retract, than I switched to SW retract in the same gcode file. That’s realy, realy strange! I will have a look into the code…
20160621_183704.jpg

50

I’m glad I was able to um ‘help’ :slight_smile:

51

You can follow the progress on this issue here: Marlin Github

52

Bug fixed! Sorry for the problem, here is the new hex file. This time, there should be no print errors, with or without FW retract :blush:
Marlin_LIN_ADVANCE.rar (105 KB)

53

Thank you!

BTW…the increase in top surface quality on my prints is quite noticeable!

J

54

Any chance we can get the whole sketch? Or can you modify for dual extruder? :slight_smile:

55

You can find the whole thing (not configured for the taz) here: Github

56

It’s now also fixed in the main Marlin Repo, so you can use the main Branch from Github

Please report back all strange things, and also if it is now working with retraction. :slight_smile:

My latest configured version is in the attachment.
Marlin.rar (1010 KB)

57

Yes I was getting the same gaps but they went away after switching to fw retraction. Glad the software retract bug is fixed!

58

Thanks for posting the update. First calibration cube print with software retraction looks good… Top finish looks great, ABS corners look sharp,

Which files from the Github repository do you modify to run on the TAZ? Is it just configuration.h?

59

Configuration.h,configuration_adv.h and one additional pin definition in pins_rambo.

It’s easy if you take an existing config file as a reference.

60

Thanks!

I was going to do a diff on the files to see the parameter changes.