Filament sensor debouncing?

zBeeble

So... I looked at my previous filament sensor post --- and this deserves a new post.

I bought some cheap filament sensors on Amazon. AFICT, they produce a positive spuriously but briefly. Now ... I could fix this with some sort of capacitor circuit --- but that is just not my wheelhouse.

Now... my filament sensor is about 15 to 20 cm from the extruder --- so I'm fairly confident that a 1 second debounce could be added to software and/or to the M command that sets it up --- and that this would solve my problem.

Equally, yes, I could not buy sensors on Amazon, but since I have, I'm hoping that this little feature request my save my (and other) butts in hindsight.

OwenD

@zBeeble
You can check for bouncing in filament-error.g
Note: On some firmware versions the name could be filament_error.g
Something like this is sufficient if it is a simple switch.

; check if switch is bouncing
G4 P10 ; delay 10ms to debounce
if sensors.filamentMonitors[0].status="ok"
	echo "switch bounce detected - error cancelled"
	M99 ; break out if sensor value is zero again (bouncing)

;error persists so we pause print
echo "Filament error - pause print"
M25

zBeeble

@OwenD OK. Awesome. I don't even know where to look for this functionality in the docs, but I'm ecstatic that it's there. My mind is immediately thinking about a retry loop for certain types of actions...

But back to the project --- my firmware is 3.4.1. Nearly current. filament-error.g?

zBeeble

@OwenD Are we using python-esque indentation is scope here?

OwenD

@zBeeble
Yes should be filament-error.g
See M591 and Events for more info

OwenD

@zBeeble said in Filament sensor debouncing?:

Are we using python-esque indentation is scope here?

Yes.
See here for more info on conditional code and meta commands

dc42

@zBeeble what problem is the bouncing causing? Bouncing should occur only when the state changes.

zBeeble

@dc42 bouncing may not entirely be the right term. Maybe noise? I think the code will be correct, tho.

Basically, the filament enstop is "detected" ... but the filament went out --- and the LED is on. I can only assume that the sensor was only "off" for a very short time... as "off" is the runout condition.

I can feel a little microswitch in the sensor (it "clicks" when I push through a new filament). Given my recent experience with mouse switches, I can only assume the same thing --- that noise due to the switch being "on" (or contacting) but moving a bit --- noise is causing a momentary sensing of "off"

zBeeble

@OwenD Just got verification that this works (at least to sense the filament not running out).

... I increased it to 100ms to be pessimistic, but that would be correct... the printer is still (for a little while) in filament. It should run out for real (testing the other branch) before this print is through.

zBeeble

@zBeeble So... after some time, I finally did run out of filament, and that, then, verified both branches ... "This works" ...