[BUG] wifi_upload blocks temperature processing, so hotend and bed can overheat

[vlsi]

Did you test the latest bugfix-2.1.x code?

Yes, and the problem still exists.

Bug Description

Uploading a big file via wifi can cause overheat of the hotbed and hotend

Bug Timeline

No response

Expected behavior

wifi upload should not stop processing critical tasks like temperature monitoring

Actual behavior

wifi_upload seem to stop all the other processes

Can it be caused by a spinloop in https://github.com/MarlinFirmware/Marlin/blob/f6de698fc8bf284c76a2ae09b1c7b5041c74d495/Marlin/src/lcd/extui/mks_ui/wifi_upload.cpp#L678-L681 ?

It looks like:

1. A file might take a lot of time to transfer
2. The transfer performs multiple packets, so it iterates many times in the loop
3. Every time it calls watchdog_refresh, so I guess it resets the watchdog timer, so watchdog never fires

I guess the fix would be moving hal.watchdog_refresh(); outside of the loop.

Steps to Reproduce

1. Start preheating
2. Start uploading a big file (e.g. 500M)

Version of Marlin Firmware

2.1.1

Printer model

FlyingBear Reborn 2

Electronics

No response

Add-ons

No response

Bed Leveling

ABL Bilinear mesh

Your Slicer

Cura

Host Software

None

Don't forget to include

• [X] A ZIP file containing your Configuration.h and Configuration_adv.h.

Additional information & file uploads

R2S_2.1.1_GD32F407.zip

[vlsi]

It is still relevant

[ellensp]

Temperature control is on a timer interrupt, ie they keep happening...

Did you actually verify this, or is it speculation?

[vlsi]

Could you please hint where timer interrupt-based temp control is activated?

The "overheat during wifi transfer" issue is very known in Reborn 2 community.

Here's how it looks like: https://youtu.be/7MxLkfBDbUM (not mine)

I have pesronally run into the issue as well

[vlsi]

Here's a typical precaution: https://t.me/reborn2support/42670

[ellensp]

Video contains no useful information. Only more speculation.

I haven't checked the firmware source code from their website yet, but I strongly suspect that Flying Bear disabled the thermal runaway protection for the units coming out of the factory. This is not likely to be one of those rare cases where a mosfet fails "on" as the hot end stayed cold after powering on.

Motherboard listed In Config BOARD_MKS_NANO6 Is not even in stock Marlin

[EvilGremlin]

BOARD_MKS_NANO6 do not exist in vanilla marlin, so your issue is kinda invalid in the first place. But MKS_UI is buggy orphaned code in general, so yeah, no wonder. Basically, don't use LVGL_UI if you aren't willing to fix it yourself.

[ellensp]

reborn 2 firmware https://drive.google.com/drive/folders/15Dw9_xIY9br-wts5MLN6GH2n5i6SZ7ik Configs are different from provided....

[ellensp]

also from the comments on that video "Had another thermal runaway event while preheating. This time not using the wifi."

[vlsi]

Video contains no useful information.

Which information could help to diagnose and analyze the issue?

Basically, don't use MKS_UI if you aren't willing to fix it yourself

Many users might be hit by the bug, so I wonder if something can be fixed/improved in Marlin itself so the upcoming mks-based builds still work. For instance, there might be a trivial "disable heater" at the beginning of wifi_upload and "re-enable it after the upload is done.

reborn 2 firmware

For reference, here's an exact firmware https://github.com/vlsi/reborn2-marlin/tree/2.1.1-rb2_autolevel (see commit https://github.com/vlsi/reborn2-marlin/commit/527c4aa9a8b28f3440138f3f2de7a8f897271f6c)

Unfortunately, the official firmware sources contain many irrelevant changes (like random whitespace change, a random addition of /* C-declarations for C++ */ comments), so I've made a small cleanup in https://github.com/vlsi/reborn2-marlin/commit/6952091136c4d1bdd5a01292130e6322b5484122 so the diff is easier to analyze.

[EvilGremlin]

Yep, but no one willing to fix many failings on LVGL_UI yet, so general advice - just don't use it until it's fixed.

[vlsi]

@EvilGremlin , I am willing to add "disable and re-enable heater to wifi_upload.cpp". Will that be accepted?

[thisiskeithb]

@EvilGremlin , I am willing to add "disable and re-enable heater to wifi_upload.cpp". Will that be accepted?

Since thermal runaways were reported with Wi-Fi disabled, that’s likely not going to fix the issue. Something else is causing these thermal runaway events.

[vlsi]

Something else is causing these thermal runaway events.

wifi transfer makes it 100% reproducible

[thinkyhead]

One thing you can do to look for problems is to enable PINS_DEBUGGING and then send M43 to the machine to get a pins report. It's a little trickier to check on timer conflicts, but that is also possible. If neither of those is the case, then the next thing to look for is weird differences between a firmware that works, and some nearby version of Marlin that doesn't work.