Is there a reason you couldn’t add it to your board… outta pins or somthing?
[edit: If more complexity is OK, you could add another rotary sensor directly to the e-motor. It wouldn’t catch missed steps, but the motor sensor could be used to start/stop the timer. Better yet, a non-invasive circuit that senses pulses (LPF for the PWM freq) on the e motor leads.]
I definitely could add it to the firmware. I am running Klipper on an skr not pro.
I just don’t want to stop the print if I run out of filament (and I don’t want any chance of a false alarm on a print).
It is more appealing to me to have a separate project that won’t affect my printer, but will still notify me.
I was thinking I would put it on the line before the ptfe tube heading to the direct drive extruder. So I would have 400 mm or so before I needed to change it (if it ran out). That should let me filter out anything too serious.
I also don’t want it to ever cause a jam. That would get it thrown in the trash.
I would connect it to an ESP32, so I have a lot of room to do any complicated logic. But simpler is better.
It is very possible I am making it harder on myself by inflicting my own weird requirements.
It occurred to me you could have an external circuit that monitors the encoder for filament movement, and also taps into the STEP line to the E motor driver, to know when the firmware is attempting to move the E stepper. Then if the E stepper is moving and time goes by without the encoder moving, it means something is wrong, independent of any firmware.
Every time I unwrap a new filament spool I weigh it and mark the empty spool weight on the spool, along with the date that I unwrapped it.
Before any long print when I’m not sure if there’s going to be enough filament I weigh the spool again and subtract that empty spool weight. I haven’t run out of filament during a print in years. I have found the slicer estimates of filament required for a print to be accurate enough, though I usually enter the filament density myself.
I occasionally print some rather large objects and keep some 5 lb spools on hand for those sorts of things.
Can you set the filament sensor input for NO? You could use two bearings as the pinch rollers. Connect your sensor leads one to each bearing mounting screw. When the filament runs out, the bearings touch and complete the circuit, and tell the controller the filament is out.
I just throw it out. Life is too short.
Yes…but that wont tell you if the filament just jams.
I wouldn’t call your requirements weird as long as they are there from experience. What is the biggest problem with any printer depends mostly on what you do with the printer. For me nylon clogging in the nozzle was a somewhat common mode of failure (ends in grinding at the extruder), which can only be caught with motion. It would suck if any sensor resulted in a jam.
For sure a switch will add less friction than rotary, but that doesn’t seem like much compared to the force available from a DD extruder. What kind of extruder are you using? I upgraded my mk8 to one of those cheap dual gear ones, and holy cr*p it’s like the vulcan death grip now. It would take a serious jam to stop it… my e can actually skips steps with nylon.
I have a hemera direct drive. In the past 6 months, I have probably had 10x more filament run-outs than jams. I can’t remember ever getting this hemera jammed, actually.
Generally the only time I jam my extruder is when the end of the filament is really stuck to the spool. If they just tape it on, it’ll just pull it free, but sometimes they secure it through a couple holes in the center, then it jams once the spool hits the top of my dry box.
I found a new way to jam an extruder that’s just about guaranteed lol… 3dxtech PLA-CF using an all metal heatbreak. With my crusty old teflon heatbreak it has no problem, but a genuine all metal e3d is a no go. Go figure… sucks having to change back and forth just for one filament though.
In other less exciting extruder news… the filament sensor I posted above got jammed badly last weekend. Took it apart, and found pink pla shredded into what looked like cotton candy, wrapped all around the unshielded bearings, lol. Yep, had to learn the hard way about unshielded bearings. I found a good deal on some rubber sealed abec3 on amazon, and holy c*** was I missing out before. The bearings are so smooth you can’t even tell they turn… made a huge diff in how the sensor feels/operates. Wish they fit on my primo, lol.
I have to say, having a sensor is not a bad idea even if you have a reliable rig. With some of the more expensive cf filaments, a failed primo core for example might waste $60!
I missed this earlier… what I do is use advanced pause and nozzle park when the sensor runs out. So the print can be continued whenever I get back to it for inspection. For some filaments and models you simply can’t open the enclosure door to check the extruder without the model popping off the bed as cool air hits it. Otherwise a false alarm just means more time to complete the print. A positive alarm also just means there’s a couple inches of extruded line on the model missing… if that’s OK (depends on model) then the print can still be finished. That was what I was after designing that sensor… making that couple of inches as short as possible.
