Duet3 v3.3 6HC Magnetic Filament Sensor = 'sensorError'

verdeel

@alankilian Thank you. I'm going to give it a try once I manage to get my hands on some capacitors that are roughly the same capacity. I've already tried to get the sensor cable away from the stepper cables, but it keeps happening. The sensor cable is bundled together with the z-probe from the smart effector and the fans, heater.

PetrKroupa

@verdeel You can solder capacitor directly to sensor. But you will probably lose waranty. Just for info. My waranty is also gone BUT sensor works.

verdeel

Before starting with soldering capacitors to anything I decided to investigate further. Before making any changes I sometimes had 20 or more sensorErrors just during mesh leveling.

First I installed a new wire for the sensor, completely outside of the machine, away from all other cables incl. the other cables going to the smart effector. Test print of about 45 minutes, no errors.

So my issue was being caused by the wiring. I still had some cable for analog sensors, with braided shielding. But because those wires were pretty thick I just removed the shield from that wire, and put it around my new filament sensor wire, guided it through the nylon sleeving that goes to the smart effector and did another test print. Again, no sensorErrors!

I'll have to see how it behaves the next couple of weeks, but to me it seems that the error was caused by my own wiring being to close to the other wiring for the smart effector.

Surgikill

I'm having the same issue, I'm on a Duet 2 WiFi. I'll try the capacitors to see if it is fixed.

Surgikill

@surgikill said in Duet3 v3.3 6HC Magnetic Filament Sensor = 'sensorError':

I'm having the same issue, I'm on a Duet 2 WiFi. I'll try the capacitors to see if it is fixed.

Didn't fix it. I put a 50V 10uf capacitor on the sensor, no dice, still getting sensor error sporadically.

PetrKroupa

@surgikill - 10uF is small. I have two. 1x 100nF and 1x 220uF.

dc42

@petrkroupa I don't believe the sensor is very sensitive to power supply fluctuations, unless your sensor is faulty. I think it more likely that you have a bad connection in the +3.3V or ground feed to the sensor, or you are getting static discharge from the hot end to the sensor connections.

PetrKroupa

@dc42 - Two sensors on two different printers? I'm not saying it's not possible but it's unlikely. Capacitors have resolved the issue on both printers. If the cause is elsewhere, I really have no idea where. However, since then, both sensors have worked perfectly. See for yourself. For 23 hours of printing, the tolerance is 97-104%. I do not think it is possible to achieve better values.

Duet3D rotating magnet filament monitor v4 on pin 121.io1.in, enabled, sensitivity 25.00mm/rev, allow 50% to 200%, check printing moves every 10.0mm, version 4, mag 131 agc 96, measured sensitivity 24.62mm/rev, min 97% max 104% over 42855.6mm

alankilian

@petrkroupa

It's great that adding capacitors to your sensor PCB get you working sensors.

Adding a large cap to the PCB helps when your 3.3 Volt supply is not continuous or low-impedance.

It's possible (I'm not saying it is, but for others...) that you've got marginal crimps on your wiring to the sensor. This can cause very short outages in power which would be compensated by the large cap on the sensor. This is a good thing to have if you have marginal connections in power or ground. Also good to have "Just in case."

Do you use wiring that is crimped from the factory or do you make your own?
If you make your own, do you crimp? Do you have an expensive crimp tool or an inexpensive one?
Do you crimp AND solder?

Anyway, this is great that you've got it working and should help others if they also have wiring that somehow is not able to power the sensor continuously.

PetrKroupa

@alankilian - I crimp and then solder it for sure. I don't have the original tool. I changed the whole line several times. Subsequently, I shortened it to the shortest possible distance. Twisted together and used shielding. Even so, sometimes the sensor made "sensor error". I read the datasheet for the sensor used, looked at the overall wiring on the github, and found nothing wrong. When I had no other idea and tested 5 sensors, I tried capacitors. Well, finally it worked.

I'm not sure why. And since the sensor is directly on the extruder and powered by 1LC, I can't even check what's going on there with the oscilloscope.

It is true that adding a capacitor was probably the last attempt. Then I would give it up. It cost me a lot of time, but the idea of such a precise sensor is simply tempting. It may not always help and if someone has another problem with the sensor BUT as I say, it helped to me.

Surgikill

@alankilian I used a good crimp tool as well, EDM jaws. I do not solder seeing as that is usually not recommended for crimp connections.

@PetrKroupa I'll try adding some different capacitors. I just ordered shielded wire for it, so maybe that helps.

It also looks like there are some filament monitor fixes in 3.4 https://github.com/Duet3D/RepRapFirmware/commit/2ee6b1a26dcf7fd0b6037c519ac3e056f1e15c62

0

dc42 committed to Duet3D/RepRapFirmware

Bug fixes to filament monitors

Surgikill

Well I found 1 issue. Part of the shaft was rubbing on the filament monitor. I resin printed the case and all the dimensions came out perfect, but apparently there is tolerance in there made for FDM, so the sensor was way too close. I'm going to try some test prints and see what happens.

alankilian

@surgikill You can just add a paper shim in the case and increase the spacing to 0.25mm as recommended.

PetrKroupa

@surgikill -
If the magnet is too close, it can easily happen. Then you will often have "too litle mowement" and if the scratches are quite deep, then probably "sensor error". I did it too . If you can solder well, only the sensor can be replaced. It is relatively cheap.

I recommend assembling the MFM really carefully. Take your time. Do not be afraid to use sandpaper and, if possible, define the axial clearance of the shaft minimum as possible.

as @alankilian writes, feel free to put under electronics a piece of paper to get the ideal AGC value. smaller, better but too small meant potential contact between the shaft and the sensor.

Surgikill

@petrkroupa Yea the scratches are pretty deep. At this point I'll probably just use a microswitch. The magnet has caused more issues than it has fixed. I should have just done that from the beginning. The laser one didn't work worth a damn, and now the magnet one, which I thought would fix all my issues, doesn't work worth a damn. $100 gone.

Do you have a PN for the sensor? I can solder a new one on.

PetrKroupa

@surgikill - AS5601. I2C magnet sensor. Find it on some inet shop.

Do not give up. When MFM works, it's the best filament sensor I've found so far. He just wants a little kind of assembly.

Surgikill

@petrkroupa It's not that. I have product to ship and I can't be dealing with a sensor that does not work and ends up causing issues on prints.

Surgikill

@petrkroupa Well I bought some of those magnet sensors on amazon and they came in today. They are AS5600, not AS5601. Are they still usable, or are they too different?

Also, this is the circle on mine.

PetrKroupa

@surgikill I'm not sure but according to the internet it is 5600 for electronic potentimeter. Angle range 18-360. Which is not appropriate. You need a complete turn.

Nicely scratched wheel. If no metal is visible, the sensor may still work. But it's not certain.

Surgikill

@petrkroupa it doesn't look like there is metal visible, but it's hard to tell with the brass that is embedded in there. I still get random "tooLittleMovement" notifications, so it's not working right. It will hold 94-107% solidly, then a random "tooLittleMovement" notification shows up.