[BUG] 3.10 Preheat menu loops when power on with loaded filament

[Lyr3x]

Printer type - [e.g. MK3S, MK3, MK2.5S, MK2.5, MK2S, MK2]
Mk3s+
Printer firmware version- [e.g. 3.8.1, 3.8.1-RC1, ...]
3.10 / 3.9.3

SD card or USB/Octoprint
SD

Describe the bug
When power on the printer with loaded filament the preheat menu is appearing again and again (inconsistent delay between). This only occurs if filament is loaded before power on the printer. As soon as I remove the filament once the issue is gone until the next power switch. Filament sensor is detected correct and the lever etc. is working flawless.
Flashed 3.9.3 back and the issue is gone. I repeated the flash process multiple times.
I also tried multiple materials and colors.

Tried to factory reset before and after flashing 3.10 with no success. Sensor voltages and states are correct. When unplugging the filament sensor the pop up does not appear.

I can't identify a hardware issue as with 3.9.3 everything is working without any issues. The preheat menu appears also if autoload is disabled.

When unloading and reloading the filament while the printer is powered on, the sensor is working as expected.

To Reproduce

1. Load filament
2. Power off the printer
3. Power on the printer
4. Preheat/Autoload Menu appears again and again

Expected behavior

1. Load filmanet
2. power off -> on
3. Nothing

If you need any more information, i am happy to provide them ASAP.

[wavexx]

@Lyr3x can you make a video of the behavior?

If I understand correctly, the Preheat menu shows up, you exit the menu, but after a few seconds it pops up again?

[Lyr3x]

Sure. Follow the link: https://photos.app.goo.gl/trkSfidprWqBfEJdA

I printed over the weekend with 3.9.3 without the issue. Flashed prusa3d_fw_MK3S_MK3S+_3_10_0_4481.hex for the video and as you can see, the issue is instantly present.

[wavexx]

Might be an issue with the fsensor version detection?

Start from reset with the filament unloaded. Under the support support, just under the date what does the "Fil. sensor v.:" say? Then go to voltages and note down the value. Then insert the filament just to trigger the sensor (ignore autoload - go back to voltages). What does it read now?

Then re-do the same, but with the filament inserted and reset. Detected version under support? Initial voltage?

[Lyr3x]

Reset with unloaded filament:
Fil. sensor v.: 0.4 or newer
voltage unloaded: 4.4V
voltage loaded: 0.4V
IR Sensor state: 0

Reset with loaded filament:
Fil. sensor v.: 0.4 or newer
voltage loaded: 0.4V
IR Sensor state: 1
voltage unloaded: 4.4V

[wavexx]

Looks good. I assume you don't see any weird fluctuations in voltage right if you leave it on for several seconds? (sadly, the display doesn't refresh fast enough to see fast peaks, but that would be a good indication).

Too me it the behavior would match if the was a random disconnect from the cable, maybe. Although why this changes with 3.9 is a bit weird but I need to look at a diff again to be sure.

I know it sounds stupid considering 3.9 works, but just as an obvious suggestion for the above, did you try to unplug/replug the sensor cable on both the board and extruder just to reseat the contacts? Connector or cable issues are always annoying.

[Lyr3x]

The voltage of the IR unfortunately sensor stays stable at 0.4 or 4.4v. Any way to get more/faster debug output of via terminal?

Too me it the behavior would match if the was a random disconnect from the cable, maybe. Although why this changes with 3.9 is a bit weird but I need to look at a diff again to be sure.

I absolutely thought the same and checked the wiring, measured continuity and unplugged the sensor (both at the sensor and on the MB). I also checked if with heating and with fans on something is changing.
The fact that 3.9.3 is working fine (I flashed now 20+ times back and forth) doesn't look like a typical wiring issue. I am an engineer but not a C++ developer and do not understand the diff regarding the sensor, so i cant really support there.

I'll quickly check the wiring again and see if i can observe something new. I'll measure the sensor voltage by hand and see if i can see any fluctuation

[wavexx]

Since you're going to do that, measure directly at the connector on the einsy and move the extruder cable bundle by hand (just in case the cable itself has issues). If you have a min/max feature on your meter you could catch a potential issue without having to look at the display.

Another question sprung to mind: so if you insert the filament after the printer has been turned on, then cooldown until all fans are off, does it also trigger randomly? Or it's only triggered from a cold-start with the filament pre-inserted?

[Lyr3x]

We checked with 4 eyes: stable 4.39V and little more "unstable" 0.38-0.39v which looks fine for me and can even be my low cost meter. Nothing changes if moving the cable bundle.

I do not even need to hot load the filament, just trigger the sensor. If filament is inserted AFTER the printer is turned on i don't see the issue. Only if I power on while the sensor is triggered. Does that make sense?

[wavexx]

Yes it does, maybe some initialization is skipped. So if you poweron with the filament inserted (issue present), then remove and re-insert the filament the issue is also gone?

[Lyr3x]

Exactly.

[wavexx]

Ok this narrows down the issue a lot. For some reason I cannot trigger the issue, but with these set of conditions there are not many places where this can happen.

[Lyr3x]

I am more than pleased if I can test anything (like changes on a branch) or provide info on short notice.

[diluno]

I'm experiencing this problem as well! Already tried flashing the firmware again - now I guess I'll downgrade to V3.9.3.

Also, when I select a filament type in the preheat menu, it just keeps on "loading". It just doesn't stop oozing filament, even after selecting "Yes" when asked if it loaded correctly.