bigtreetech / BIGTREETECH-TouchScreenFirmware

support TFT35 V1.0/V1.1/V1.2/V2.0/V3.0, TFT28, TFT24 V1.1, TFT43, TFT50, TFT70
GNU General Public License v3.0
1.32k stars 1.65k forks source link

[BUG] Print job stops randomly due to PR #2824 #2850

Closed kisslorand closed 1 year ago

kisslorand commented 1 year ago

Description

With latest master print jobs are stopping randomly. I identified the culprit as PR #2824.

Others are also confirming this freeze issue: https://github.com/bigtreetech/BIGTREETECH-TouchScreenFirmware/issues/2835#issuecomment-1737045630

Display froze after about 15 seconds Restarted... lived 30 seconds this time (I have persistent info enabled) Reset to default values... I see different failure modes

  1. Screen freezes, doesn't respond anymore
  2. Notification area is flashing, but screen doesn't respond anymore
  3. Notification message changes when tapping the screen (from "busy" to "SD card inserted" and "SD card removed", cycles to theses 3 messages when I tap the screen, but I cannot exit

The author of the PR #2824 himself admits the presence of this bug here: https://github.com/bigtreetech/BIGTREETECH-TouchScreenFirmware/issues/2835#issuecomment-1737157504

The freeze should not be present but after some time TX and RX KPI drops to 0 in my TFTs. The bug is present even with unbuffered mode

Steps to reproduce

  1. Start a print
  2. Pray
  3. At some point print freezes...

Expected behavior Print finishes successfully.

Actual behavior Print stops at random point.

Hardware Variant

TFT: BTT TFT35 V3.0 MB: BTT SKR-2 with STM32F429VGT6

TFT Firmware Version & Main Board Firmware details

TFT: latest master Marlin: latest bugfix

Proposed fix

I am asking @bigtreetech to urgently revert PR #2824 as it defeats the main purpose of a 3D printer, to print an object.

Remarks

Those who do not have a backup or just want a more updated FW with print hesitation guard feature (not present in this FW) can download it from here.

kisslorand commented 1 year ago

I spent some more time on this bug and I thought it might be my rig, some memory corruption so I did a full reset, reinstalled everything. At first it seemed everything is back to normal the FW is working fine. Here's the video printing a cylinder in vase mode without ARC, many many tiny segments at a high speed so height and layer number are changing rapidly.

https://drive.google.com/file/d/16nBcgebuQoHMPYeNKCiaOuTfkyMqCEsP/view?usp=sharing

After some repeated testing (with Advanced_OK on and off) suddenly the print stopped, there was no advancement of the progress (height, layer number, progress bar), only the timers were working.

https://drive.google.com/file/d/1yvT9lwoZLY1_NyKkqC5R9Z4OAcY9WrGc/view?usp=sharing

After a while a more interesting thing happened, "No printer attached!" message appeared and the print started at an insane speed. I managed to capture it on video. https://drive.google.com/file/d/14O3ADOWLXOJUiEAoc3cSpt4pknTv5xDG/view?usp=drive_link

kisslorand commented 1 year ago

OK, new data. After several attempts I could reproduce the issue even with PR #2824 reverted, so I take back that PR #2824 is the culprit.

If in the future anyone else reports this issue than we already know where NOT to look to, it's not PR #2824. I will close it know as I highly doubt a regular user would stress the TFT as I did to bring this bug to the surface.

I am sorry if I caused trouble to anyone with this.

github-actions[bot] commented 8 months ago

This issue has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.