Magnetic Filament sensor

[cfletch]

I just installed a magnetic sensor, pcb ver 1.7b.

configured with; M591 D0 P3 C"e0_stop" S1 R70:130 L24.8 E3.0 in config.g file.

Upon powering up the green led flashes 3 times, but then continuously emits a pattern of a single green flash followed by a single red flash.

if I issue a M591 D0 command the response is: Duet3D rotating magnet filament monitor v1 on pin e0stop, enabled, sensitivity 24.80mm/rev, allow 70% to 130%, check every 3.0mm, no data received

I have a Duet2wifi board running Firmware: RepRapFirmware for Duet 2 WiFi/Ethernet 3.3 (2021-06-15)

What advice can you give me to resolve this?

Thanks,
Carroll

[alankilian]

@cfletch OK, you've got good LED blinkage so you've got a working sensor with the power connected properly.

The only think left is the signal wire.

As a test for you, I pulled out the signal wire and powered-on and I got the exact-same message you did.

(Signal wire disconnected)
M591 D0
Duet3D rotating magnet filament monitor v1 on pin e0stop, enabled, sensitivity 28.80mm/rev, allow 60% to 160%, check every 3.0mm, no data received
 
(Signal wire connected)
M591 D0
Duet3D rotating magnet filament monitor v3 on pin e0stop, enabled, sensitivity 28.80mm/rev, allow 60% to 160%, check every 3.0mm, version 3, mag 128 agc 112, no calibration data
 

So, check CAREFULLY the signal wire and make sure you've got s good connection at both ends.

Also check to make sure it's plugged-in to the correct endstop place.

[cfletch]

@alankilian I had resistance checked the cable before installation, but I'm always suspect of the connection on those tiny wires, so I checked them again. They were OK, but I tried another set of wires just to be sure, but the error remains. I did try both E0 and E1 connections while making sure the configuration was correct but neither worked. I did notice that the limit led on the board was flickering synchronously with the led on the filament sensor. I took a look at the board for cold solder joints, etc. but nothing stood out. any other ideas?

Thanks for the quick response,
Carroll

[alankilian]

@cfletch said in Magnetic Filament sensor:

the limit led on the board was flickering synchronously with the led on the filament sensor.

Interesting. I've never noticed that.

Indeed, my limit LED flickers with the filament monitor LEDs.
When I disconnect the sensor wire it stops, so I think your sensor is connected correctly. (I learned something new today that might help the next person debug their sensor. Thanks!)

@dc42 any ideas?

[infiniteloop]

@cfletch said in Magnetic Filament sensor:

any other ideas?

Does the green LED of the sensor flicker if you push some filament through it? If not, the idler may not work properly.

A major cause of trouble is the housing: I had to print it myself, with FFM, of course. Well, that's not overly precise, the case design has been constructed with a resin printer in mind. With FFM, we face three problems:

1. the filament guides may be too narrow
2. the idler might be stuck
3. wrong distance between magnet and sensor

The latter one is tricky: if it's too small, the magnet will not rotate. If it's too large, the signal is too weak. In both cases, you get no values from the sensor, just as described in your op.

If the axis of the magnet assembly rotates when passing filament through the MFM (check this with a point on it's axis, a felt marker helps), the distance between magnet and sensor could be too large. Then, the bottom side of housing lid should be sanded in order to reduce this distance.

[cfletch]

@infiniteloop The alternating green and red pattern that I see is occurring at a rate of about 110 flashes per minute. So, I'm not sure I would see the other pulses. Also, I disconnected the board and distanced it from the sensor. When I powered it up I got the correct 6 red flashes that indicates no magnet sensed. I did check to see if the metal shaft rotated as you suggested and it does not.

Thanks,
Carroll

[infiniteloop]

@cfletch Sorry for being a bit late …

@cfletch said in Magnetic Filament sensor:

I did check to see if the metal shaft rotated as you suggested and it does not.

To investigate this, you should remove the PCB and the housing lid, keeping just the bottom part of the housing with idler, spring and hobbed assembly. Does the axis rotate if you push some filament through? But be careful, the idler assembly likes to jump out of its box

If the filament isn't moving smooth, re-drill the filament path with a 2 mm drilling bit.

Next, look if the idler assembly looms above the top of the housing body. You can check that with a ruler. In this case, the two printed idler parts need manual treatment. With my prints, I had to remove some material around the tiny axes (these pins who center the ball bearing).

Rework your printed parts until there are no more kinks and/or superfluous material left. Re-assemble all mechanical parts (i.e. without the electronics) and verify proper function: the filament must pass without larger resistance, and the axis must rotate accordingly.

Finally, if the mechanics work, you might have to look for the proper distance between magnet and sensor …

[cfletch]

@infiniteloop I did as you suggested and the sensor will now detect the motion of the filament. When queried the probe responds with: Duet3D rotating magnet filament monitor v3 on pin e1stop, enabled, sensitivity 24.80mm/rev, allow 70% to 130%, check every 3.0mm, version 3, mag 133 agc 66, measured sensitivity 23.04mm/rev, min 87% max 107% over 27.4mm. This was obviously a very tiny print, it did pause at one point and reported an overfeed issue. I didn't expect that!

Thanks very much for your help. I'm certain I wouldn't have figured it out on my own.
Carroll

[infiniteloop]

@cfletch Glad that you got the sensor running. With my (self-printed) housing, I had the same blues just some days ago.