Oh, one thing I guess I should mention: The sensor has a FAULT output (same as the LED), which is programmed to turn on when it detects errors like the filament being outside of the expected diameter range. So with that output, it can be directly used as a filament runout sensor as well, even without firmware support for the InFiDEL's measuring output!
Oh good timing, I have issues with my filament sensors and was looking at a replacement. Pity I only have some of the parts on hand, no dowels - order time. Btw your youimagine design link doesnt work!
This is great! Thanks for putting this out in the wild, Tom. Have you/are there plans to cover integration with the firmware on the printer controllers?
Ha! I was about to ask if there was a way to use it for that. Guess I'll put that thing in my DIY queue now. Might be a nice testing ground for my PineCil as well. Thanks for making this design available, very much appreciated
I second that :-) even if it didn't work that well it is interesting. I would like to make my own filament extruder one day and hearing about potential issues with one is very valuable.
11:50 THIS is the reason i love you. Seriously dude, thank you for everything. Best 3d printing channel together with cnc kitchen as being the more analytic half. Great job as always!
@@alf3071 Yes, this is a really good application example. You could use very cheap filament with this sensor and the sensor tells the machine when the filament is too thick and stops automatically. Furthermore, this sensor would serve as a filament runout sensor. Thus it would be a multifunction filament monitoring sensor.
what i like about you is that you care about us printers. but all work deserve money, i hope if some 3d printer manufacturer takes you idea and design that you'll get paid!
I'm sitting here _thinking_ of CADding up stuff like furniture I need to revamp my living space (because my time is mostly taken up by my day job that I actually need to pay them bills and sleeping/gaming to recover from said day job), and Tom here is actually making his ideas happen. Good times.
Thanks for the design. Could the design be changed tp read the movement of the filament as well as the diameter. I was looking around for something that could be used to notice the nozzle is clogged and filament is skipping in the extruder. I could imagine using a mouse centre wheel quadrature encoder attached to one of the bearings to do that job.
Well, here's another data piece that Stefan can start including in his filament strength tests. "Was the filament over/under sized during the printing of each test part?" I love the tinkering/constant improvement in this hobby. Thanks for giving us another tool in our toolbox!
In order for the extruder to compensate the thickness difference of the filament it must know the distance between the sensor and the extruder or? I assume this feeds live data from the point where the bearings touches the filament, but it will take some time before that point of the filament is consumed.
I'd love for you to consider redoing the filament maker in a updated 2021 video, using the knowledge you've gained since building it, the things you'd change, the tricks to make it cheaper. Kinda like you went over with the Infidel sensor.
This is a great idea and project! Some (hopefully helpful) critisism: If I'm not missing something, your sensor only provides the intended information under the assumption that the filament is perfectly round. It measures the diameter at a specific point, but not if the diameter is consistent. So actually you can not transform this into information about adaptation of the flow factor. You mentioned that better quality filaments are usually not oval - but imo they also should have a consistent diameter. I think this is idea has potentially higher impact for cheaper filament and to be able to compensate tolerances better.
I hope to see both firmware support and preassembled modules become available, it would be nice to have and make inconsistent filaments more useable, thanks for the design Tom!
Need to see filament extruder working, it looks like it took more filament to make it than it ever produced 😁. Fantastic work on sensor, will be pottering around with the idea this weekend, thanks for all your hard work.
You would be better with a second magnet attached to one of the bearings (disc magnet with a central bore), and a second hall sensor. Lower cost than an encoder.
@@danvalnicek What about a series of quick short retracts that cause the filament to move back and forth along only a short distance. Its not unreasonable that a 5mm span of filament might have variations in diameter that are smaller than the measuring capabilities of this, thus unless you used a large distance or time window for detecting motion, you might get false positives within that same 5mm length of filament during the repeated retractions. If you used larger distance or time windows to account for this, then you'd end up in a situation where a legitimate jam or break would take too long to be detected, making recovering from it difficult, if not impossible. Either a modification to this to specifically detect motion, or a secondary external sensor designed for this sole purpose would be a better option.
Cool! So, I can imagine that when measuring the diameter you'd want some sort of filament speed correlation as well. It could make sense to integrate a speed measurement in this thing too, perhaps also a magnet that rotates with the bearing, measuring its rotation and thereby the filament velocity ? Could then also be used to close the loop between extruder motor signals and actual filament extrusion. Imo a closed loop extruder motor isn't enough because the filament can still skip. Well, here the bearing would have to not skip over the filament too. Edit: thinking about it, it seems more likely to me that underextrusion due to the way the extruder drives filament, is a more significant factor than the filament diameter.
@@MadeWithLayers I understand. My point was (see edit) that if you're going to use this to prevent over/underextrusion, I think the filament speed is more significant and should be looked into first. But that's just my idea though.
Hi, thank a lot your description and ideas.. but i need little hepl with the SMT control board (populated)--- where can i buy it? or i must make it by myself?
totally solved the problem of at home filament making so if its used in the extrusion and the print process pulling through an oversize nozzle seems to be the most effective but only see the laser sensors on expensive extruders
I am trying to diy a filament extruder called the petpull2. I would like to advance the design a bit with this diameter sensor. Thank you sooooo much for this contribution!!!!
This is great, I would not expect HAL to be that precise but your prev video on the subject cleared the suspicions. Now think about 1Wire protocol, 1Wire can run on long cables nicely, it is pretty stable and is for e.g. used for hvac, fire and similar sensor installations... I2C is terrible on cables especially around noise generators like the extruder motor... look just how many issues with prusa laser sensor board that's on i2c and how for e.g. same sensor from duet that grabs the i2c from sensor with attiny on board next to sensor and sends data to mb via 1wire type signal works without any issues (ignoring issues that sensor can't work with some filaments)
If I havent overlooked anything then Prusa could implement this in their MK3 chimney design. They only need another magnet and replace the optical sensor with the hall effect sensor. Then the extruder gears become the sensor itself. Their tension should be enough to compensate for debris messing up the reading.
the best way for an end user to tell is the appearance of your model. Once you already have the roll, you're going to use it anyway. It's better that reviewers have these sensors and review lots of brands to let us know which is best. I've been using tecbears and SIMAX3d which is cheap and good tolerance on amazon. SIMAX3d being the best I've gotten on amazon and bought recently for $16 per roll to my door next day. I may actually pay for some prusament now that you've provided real scientific proof that its truly the best of the best in the previous video.
i love it but the only thing that may need to be done is a timing offset for the filament signal. When you feed filament in the tube someone could feed all the way through... put a line on filament with a sharpie and then back up the filament.. if that measured length in mm can be inserted the offset for accurate timing vs multiplier could be calculated! :)
Shouldn't be. The only potential problem which comes to mind is field strength of the magnet, and that's already accounted for by the calibration step.
Moin, 5:52 ist es nicht Wartungsfreier zwei Magnete auf der gegenüberliegenden Seite des Hebels zu nutzen, die sich abstoßen, damit man auf die Feder verzichten kann? LG juck.
Awesome! You made it really compact. Looking forward to have it supported in Marlin. Would be nice to have measurements on 2 axes even though the filament is usually pretty round. I tried designing something similar but your idea and implementation are 1000X better than mine haha. Congrats!
You should put a rotary encoder on the idler bearing like the btt smart filament sensor does so that it can detect filament jams aswell as measure diameter. Would be the best all in one filament sensor and would cost very little.
The ESP8266 has the noisiest analog input I know. In addition, you can only measure up to 1 volt. To start the ESP8266, the pins must be pulled to plus or minus in the correct order so that it starts at all. For the ESP I have to recommend an external AD converter. I would recommend an Atmega328 (Uno, pro mini, pro micro), Attiny85 or similar. The 8 bit computers have a better ADC than the ESP series.
Great job ...thanks can you tell us more about your experiance in making your own filement maker ... Why dont you give the new sensor another test with filement maker..please
Just a few notes from the peanut gallery: 1) Any variance for PLA carrying a static charge? lol 2) @6:30 - How about another set of spring tensioned roller bearings at the end (left side of the object) instead of a full length of bowden tube? Add a small threaded insert (that matches thread size on the bowden coupler) with enough threads to screw into your hotend radiator. End result - I'm betting this will put enough tension on the filament to act as a direct drive, perhaps?
Should do a video of Klippers filamenr hall effect support. It is even cheaper and because of the differential setup, it reads the magnet more accurate. Worth a try! Love your videos!
I think this is a great design and has a lot of potential. I am tempted to create a pull request for your code to introduce a feature which not only interpolates the calibration values based on the known diameters of the tools, but also through using a statistical test with repeated measures to ensure the calibrations are providing both accurate and also repeatable measurements. I believe the Prusa firmware does this with the PINDA probe, which uses precisely the same theory of operation as this sensor. You might notice that it takes several repeated measures on each calibration point on the bed. Averaging these measures can help improve the accuracy of calibration values during a calibration cycle, but if it also incorporates a statistical test to ensure that the values are within the expected range for this sensor then anomalous values can be prevented from affecting the recalibration (and can also be used as a self-diagnostic in case the sensor is no longer within acceptable calibration). If the LED is programmable based on the PCB design, that would be very useful for giving feedback to the user about what is happening if things go wrong during this procedure.
@@MadeWithLayers I created a new public repository on GitHub which is tracking everything you released. This can also version control EAGLE files since they're in a plaintext file format. I updated some of the filenames and folder structures to adhere to more common naming conventions, hopefully it will make the repository easier for people to navigate. Since this is public people will be able to fork it, make changes, and propose changes that can be reviewed and incorporated back into the original with pull requests. You can find a link here: github.com/drspangle/infidel-sensor/ Edit: I realized that you specified CC0 for the license on this, so I've updated the repo accordingly. I'd been trying to get in touch with you, but there's no emergency about it.
This is awesome, as always thanks for your contributions to the 3d printing community. I hope now the duet guys will show a little interest in this feature, in the past they weren't interested when I requested it.
How about having the magnet and sensor on two opposing lever arms, instead of only one? Then both can swivel together to follow kinks in the filament without compromising the distance reading.
What if the filament is oval in shape, with the minor and major diameter fairly different, but the cross sectional area is still equivalent to a 1.75 mm diameter filament? If your sensor rides on the major or minor diameter, its "Effective" filament diameter will change significantly.
Excellent engineering! This answered all of the questions I had about the sensor from the diameter measurement video. I would like to see your take on an extruder.
My main reason for wanting this is that I make my own filament from PET bottle. It's not 100% good consistency and sometimes a short length will become oval as it does not shrink down properly, thereby causing jams. If I can detect that early enough I should have very little failed prints from then on.
Great work! How do you make sure that the drills for calibration don't interfere with the Hall-effect sensor? Couldn't the metal lead to magnetic flux?
Hi Great product but on major flaw that nobody had though about it cause event filament is breaking inside filament loader it will no longer to be able to detect it! Cause it primary function is only detect missing filament! It's need filament movement detector like consumption it will be done by adding a roller that will detect if roller is stopping! With that feature you could say 100% bullet proof on both case
wondering if you could take a raspberry pi zero 2, and wire a filament sensor straight to the GPIO pins, then have a tiny raspberry pi zero filament sensor and octoprint server AIO? :P
Hallo Thomas, I understand the filament is assumed to be round and not oval as you mentioned at the beginning of the video, so the diameter of the filament is only “measured” on one “side”, let’s say the x-achse (z being through the length of the filament), so the y-axis is not measured, which could actually then carry the off-diameter... Wouldn’t a 60° arrangement help to improve eccentricity problems? Of course that would require more work than already needed, yet only checking on one side basically ignores the other sides (60 or 90° offset) just my thought Grüsse aus USA 😉
Future iterations probably would be to measure 2 or 3 diameter measurements, and if that works out people can make a very own average quality extruder and filament and stil get good quality prints making filament cost and selection phenomenally cheaper. Great work, thanks for sharing!!
Hi Tom, nice design. So I made one, unfortunately for me, the hall sensor is too sensitive and was picking up magnetic field from my extruder motor. Whenever the motor was running the measurement was all over the place. Do you think this could be solved using two hall sensors measuring in opposite directions to cancel external signals?
Thermistors are linear, but hall effect is polynomial due to square of the distance stuff. I'm working on a clone that uses a pi pico; I was just porting the code but uh…I ended up reimplementing everything. I'm also experimenting with using a segment of a sphere to friction-fit the bearing, since you can totally lock up the center of the bearing and the outside rotates because that's the whole point of the bearing. That changes your $3.20 custom board to a $4 (€3.36) pi pico board plus like 5¢ of headers (or some solder) from AliExpress. Not free, but not terrible. With a little luck, I might be able to knock out some of the fasteners, as I think I have a design using three 623zz with no dowel pins or screws, plus the magnet and hall sensor. Maybe €0.35 more expensive than yours.
This is a good design and your ideal of using a V-slot bearing is a good idea as well. Just to build on the idea a bit, using the V-Slot bearing on the moving arm of the sensor and then using a wheel with a bit more grip on the other wheel with a slotted disk and a photo interrupter would allow for the speed of movement or distance moved through to be measured. This combined with the filament diameter could then not only detect the diameter of the filament but also filament out (can already do this :-) ) but also if the filament jambs of breaks past the sensor. Just a thought. :-) Keep up the good work Ray
I like this, any way you could do a v2 that also includes an encoder? I have a printer (doing 3mm filament) that often jams and "prints" nothing. Adding an encoder to this, you could have it not only feedback the width of the filament but also either a pulse to show filament is moving (marlin supports this already! BTT smart filament sensor) or even recorde the rate of motion for closed-loop control
Instead of using the separate board, I wonder if it could just be added as an input to a Raspberry Pi running Octoprint? Then if there was a plug-in you could use the information that way? Thanks for presenting this Thomas.
Oh, one thing I guess I should mention: The sensor has a FAULT output (same as the LED), which is programmed to turn on when it detects errors like the filament being outside of the expected diameter range. So with that output, it can be directly used as a filament runout sensor as well, even without firmware support for the InFiDEL's measuring output!
And yes, the pins and grub screw I used in the build were temporary stand-ins as the proper parts hadn't arrived yet 😅
Oh good timing, I have issues with my filament sensors and was looking at a replacement. Pity I only have some of the parts on hand, no dowels - order time.
Btw your youimagine design link doesnt work!
Or potentially a sensor that detects unexpected sudden bulges in filament diameter, as well. Prevent clogs before they happen...
This is great! Thanks for putting this out in the wild, Tom. Have you/are there plans to cover integration with the firmware on the printer controllers?
Ha! I was about to ask if there was a way to use it for that. Guess I'll put that thing in my DIY queue now. Might be a nice testing ground for my PineCil as well. Thanks for making this design available, very much appreciated
I would love to see the filament maker in action!
Yes, seems odd.
I am surprised he still has it! LOL.
Would be fun
Very interesting, please show it to us.
I second that :-) even if it didn't work that well it is interesting. I would like to make my own filament extruder one day and hearing about potential issues with one is very valuable.
Thanks so much for your effort Tom 🙌 as always, it’s highly appreciate it!
It’s a lot of work behind all iterations and I love how it came out :)
11:50 THIS is the reason i love you. Seriously dude, thank you for everything. Best 3d printing channel together with cnc kitchen as being the more analytic half.
Great job as always!
i wish you actually tested this on a printer with cheap filament. I wanna know if it makes a different
if the filament is too thick it will get stuck in your extruder tube leading to failed print or maybe worse
@@alf3071 Yes, this is a really good application example. You could use very cheap filament with this sensor and the sensor tells the machine when the filament is too thick and stops automatically. Furthermore, this sensor would serve as a filament runout sensor. Thus it would be a multifunction filament monitoring sensor.
what i like about you is that you care about us printers. but all work deserve money, i hope if some 3d printer manufacturer takes you idea and design that you'll get paid!
great project thomas! and yes, ofc i want a video on your filament extruder!
THANK YOU! This is a nice step forward. I'm excited to see the open source community run with it!
This type of community driven innovation is exactly what makes the 3D printing community so great.
I'm sitting here _thinking_ of CADding up stuff like furniture I need to revamp my living space (because my time is mostly taken up by my day job that I actually need to pay them bills and sleeping/gaming to recover from said day job), and Tom here is actually making his ideas happen. Good times.
you get a like for that pained abbreviation, love it!
Thanks for the design. Could the design be changed tp read the movement of the filament as well as the diameter. I was looking around for something that could be used to notice the nozzle is clogged and filament is skipping in the extruder. I could imagine using a mouse centre wheel quadrature encoder attached to one of the bearings to do that job.
So... When you're pushing old filament out of the module, are you _purging the infidel?_
Klipper already has support for a filament width sensor that changes the extrusion multiplier.
and you need no electronics at all, just two hall sensors (to cancel noise)
Well, here's another data piece that Stefan can start including in his filament strength tests. "Was the filament over/under sized during the printing of each test part?" I love the tinkering/constant improvement in this hobby. Thanks for giving us another tool in our toolbox!
DIY filament (recycling)? Yes please! Sign me up!
I am impressed. Can you tell us something about the accuracy of the measurement?
In order for the extruder to compensate the thickness difference of the filament it must know the distance between the sensor and the extruder or? I assume this feeds live data from the point where the bearings touches the filament, but it will take some time before that point of the filament is consumed.
I'd love for you to consider redoing the filament maker in a updated 2021 video, using the knowledge you've gained since building it, the things you'd change, the tricks to make it cheaper. Kinda like you went over with the Infidel sensor.
Thanks Thomas, for such a wonderful gift.
This is a great idea and project! Some (hopefully helpful) critisism: If I'm not missing something, your sensor only provides the intended information under the assumption that the filament is perfectly round. It measures the diameter at a specific point, but not if the diameter is consistent. So actually you can not transform this into information about adaptation of the flow factor. You mentioned that better quality filaments are usually not oval - but imo they also should have a consistent diameter. I think this is idea has potentially higher impact for cheaper filament and to be able to compensate tolerances better.
You always keep it classy Tom.
Be proud you are one of the best channels out here.
I learned a lot from you.
can you please talk about that filament maker? I'm super curious
I hope to see both firmware support and preassembled modules become available, it would be nice to have and make inconsistent filaments more useable, thanks for the design Tom!
I love that you are just sharing this! Thanks and great job Thomas!
Need to see filament extruder working, it looks like it took more filament to make it than it ever produced 😁.
Fantastic work on sensor, will be pottering around with the idea this weekend, thanks for all your hard work.
Only Thing left is to Put an encoder in one of the bearings to detect Filament movement / jams
You would be better with a second magnet attached to one of the bearings (disc magnet with a central bore), and a second hall sensor. Lower cost than an encoder.
Just a really small encoder would be easy to integrate and basically plug and play
Why we can't just assume that the filament is not perfect and that if the filament moves the diameter changes?
@@danvalnicek , that is a really good point, assuming the precision is good enough (which it probably is).
@@danvalnicek What about a series of quick short retracts that cause the filament to move back and forth along only a short distance. Its not unreasonable that a 5mm span of filament might have variations in diameter that are smaller than the measuring capabilities of this, thus unless you used a large distance or time window for detecting motion, you might get false positives within that same 5mm length of filament during the repeated retractions. If you used larger distance or time windows to account for this, then you'd end up in a situation where a legitimate jam or break would take too long to be detected, making recovering from it difficult, if not impossible. Either a modification to this to specifically detect motion, or a secondary external sensor designed for this sole purpose would be a better option.
Awesome work as always Tom, a very interesting project. Also very apreciative that you open source your work.
I wonder if this can't also be used to detect jams through a lack of changing values
Could you show some prints with and without the sensor on a printer? To see what difference it makes. Thanks.
what about adding a filament cleaner ahead of the InFIDEL to prevent any false readings? Could be a simple wiper made out of a sponge.
Yes please cover the filament extruder maker thing I've been dying to soon if there's a doable homebrew way to recycle my pla
Thanks for iterating it useful and releasing the design. It's one of those things that I've never gotten around to building from scratch.
Cool! So, I can imagine that when measuring the diameter you'd want some sort of filament speed correlation as well. It could make sense to integrate a speed measurement in this thing too, perhaps also a magnet that rotates with the bearing, measuring its rotation and thereby the filament velocity ? Could then also be used to close the loop between extruder motor signals and actual filament extrusion. Imo a closed loop extruder motor isn't enough because the filament can still skip. Well, here the bearing would have to not skip over the filament too.
Edit: thinking about it, it seems more likely to me that underextrusion due to the way the extruder drives filament, is a more significant factor than the filament diameter.
Of course there are tons of features that could be added - for now, I just wanted something that does this one function well.
@@MadeWithLayers I understand. My point was (see edit) that if you're going to use this to prevent over/underextrusion, I think the filament speed is more significant and should be looked into first. But that's just my idea though.
Doesn't the drill bit calibration affect the magnetic field?
It's on the other end of the lever, so I doubt any detectable effect.
@@francistaylor1822 Ah you are right, I forgot about that.
Hi Tom! Can the tiny85 code be used directly on an arduino if I substitute the analog pins of the tiny85 with the analog pins on the ardino uno?
Hi, thank a lot your description and ideas.. but i need little hepl with the SMT control board (populated)--- where can i buy it? or i must make it by myself?
It seems that we have ran into the same problem
I would Love to see a Video for Setup on the Prusa I3 mk3s+
totally solved the problem of at home filament making so if its used in the extrusion and the print process pulling through an oversize nozzle seems to be the most effective but only see the laser sensors on expensive extruders
I've seen with some extruders that extrude oversized and stretched to the correct diameter. I'm guessing the analog voltage could be used for that?
Thank you for this, Amazing work toward the betterment of printing and the open source community. Hats off to you!
This is a great contribution. Thank you Tom 💚🙏
I love the debugging and the programming that you kinda skimmed over. Great work
I am trying to diy a filament extruder called the petpull2. I would like to advance the design a bit with this diameter sensor. Thank you sooooo much for this contribution!!!!
This is great, I would not expect HAL to be that precise but your prev video on the subject cleared the suspicions. Now think about 1Wire protocol, 1Wire can run on long cables nicely, it is pretty stable and is for e.g. used for hvac, fire and similar sensor installations... I2C is terrible on cables especially around noise generators like the extruder motor... look just how many issues with prusa laser sensor board that's on i2c and how for e.g. same sensor from duet that grabs the i2c from sensor with attiny on board next to sensor and sends data to mb via 1wire type signal works without any issues (ignoring issues that sensor can't work with some filaments)
That printed cutaway is sick!
Nice video Will this also work with the filament extruder to controller the speed of the winding roller to control the diameter thank you
Would this function with the Duetwifi and ReprapFirmware?
If I havent overlooked anything then Prusa could implement this in their MK3 chimney design. They only need another magnet and replace the optical sensor with the hall effect sensor. Then the extruder gears become the sensor itself. Their tension should be enough to compensate for debris messing up the reading.
the best way for an end user to tell is the appearance of your model. Once you already have the roll, you're going to use it anyway. It's better that reviewers have these sensors and review lots of brands to let us know which is best. I've been using tecbears and SIMAX3d which is cheap and good tolerance on amazon. SIMAX3d being the best I've gotten on amazon and bought recently for $16 per roll to my door next day. I may actually pay for some prusament now that you've provided real scientific proof that its truly the best of the best in the previous video.
So, how well does it work with metal infused filaments?
Where is the print test?!!
I wanna see if it does anything!
i love it but the only thing that may need to be done is a timing offset for the filament signal.
When you feed filament in the tube someone could feed all the way through... put a line on filament with a sharpie and then back up the filament.. if that measured length in mm can be inserted the offset for accurate timing vs multiplier could be calculated! :)
If I use 6x3mm magnet, should I make software or calibration changes, or just making sure it fits is enough for it to work properly?
please show the filament maker as i want to make one too, to recycle my filament for prototyping !
now if you can measure the rotation of the wheel x thickness you can have a precise volume as well as a filament jam/ run out sensor
Is there something new on this really cool project? How to get this work with Marlin?
If I modify the housing to fit it, would there be any issue with using a 6x3mm magnet, cuz I just bought 200 of them for another project?
Shouldn't be. The only potential problem which comes to mind is field strength of the magnet, and that's already accounted for by the calibration step.
Moin,
5:52
ist es nicht Wartungsfreier zwei Magnete auf der gegenüberliegenden Seite des Hebels zu nutzen, die sich abstoßen, damit man auf die Feder verzichten kann?
LG
juck.
Awesome! You made it really compact. Looking forward to have it supported in Marlin. Would be nice to have measurements on 2 axes even though the filament is usually pretty round. I tried designing something similar but your idea and implementation are 1000X better than mine haha. Congrats!
Measure it from one axis, and check for a jam with the other axis. Then the wheels make an "X" and its still compact
Is the idea that you could vary the extrusion speed based on filament diameter?
You should put a rotary encoder on the idler bearing like the btt smart filament sensor does so that it can detect filament jams aswell as measure diameter. Would be the best all in one filament sensor and would cost very little.
Looks pretty similar to the idler on an extruder, why not integrate this directly into the extruder mechanism?
The ESP8266 has the noisiest analog input I know. In addition, you can only measure up to 1 volt. To start the ESP8266, the pins must be pulled to plus or minus in the correct order so that it starts at all. For the ESP I have to recommend an external AD converter.
I would recommend an Atmega328 (Uno, pro mini, pro micro), Attiny85 or similar. The 8 bit computers have a better ADC than the ESP series.
Combining this and vibration correction has the potential to produce some seriously clean prints.
Nice sensor How precise are the like 0.01 mm or 0.001mm thank you
But, does it produce prettier prints?
Great job ...thanks
can you tell us more about your experiance in making your own filement maker ...
Why dont you give the new sensor another test with filement maker..please
Great video and love that you're making this all free. Definite thumbs up for the cut away too!
Just a few notes from the peanut gallery:
1) Any variance for PLA carrying a static charge? lol
2) @6:30 - How about another set of spring tensioned roller bearings at the end (left side of the object) instead of a full length of bowden tube? Add a small threaded insert (that matches thread size on the bowden coupler) with enough threads to screw into your hotend radiator. End result - I'm betting this will put enough tension on the filament to act as a direct drive, perhaps?
YES! PLEASE do more videos about your extrusion system. We need more leaders to follow for our own extruders! :)
With such a great invention and video I couldn't stay unsubscribed any longer. Thanks !
Should do a video of Klippers filamenr hall effect support.
It is even cheaper and because of the differential setup, it reads the magnet more accurate.
Worth a try!
Love your videos!
Could you please make a video about the differences between the Ender 3 pro and the Ender 3 v2
I think this is a great design and has a lot of potential. I am tempted to create a pull request for your code to introduce a feature which not only interpolates the calibration values based on the known diameters of the tools, but also through using a statistical test with repeated measures to ensure the calibrations are providing both accurate and also repeatable measurements. I believe the Prusa firmware does this with the PINDA probe, which uses precisely the same theory of operation as this sensor. You might notice that it takes several repeated measures on each calibration point on the bed. Averaging these measures can help improve the accuracy of calibration values during a calibration cycle, but if it also incorporates a statistical test to ensure that the values are within the expected range for this sensor then anomalous values can be prevented from affecting the recalibration (and can also be used as a self-diagnostic in case the sensor is no longer within acceptable calibration). If the LED is programmable based on the PCB design, that would be very useful for giving feedback to the user about what is happening if things go wrong during this procedure.
Tom, you didn't put your firmware or PCB design files in any kind of version control. Would you be able to put them on github perhaps?
Never used github before - but feel free to upload the firmware and I'll be happy to link to it as the recommend firmware choice!
@@MadeWithLayers I created a new public repository on GitHub which is tracking everything you released. This can also version control EAGLE files since they're in a plaintext file format. I updated some of the filenames and folder structures to adhere to more common naming conventions, hopefully it will make the repository easier for people to navigate. Since this is public people will be able to fork it, make changes, and propose changes that can be reviewed and incorporated back into the original with pull requests. You can find a link here: github.com/drspangle/infidel-sensor/
Edit: I realized that you specified CC0 for the license on this, so I've updated the repo accordingly. I'd been trying to get in touch with you, but there's no emergency about it.
Aaaad.. done! Thanks for uploading!
Is it suitable for rigorous ovality and tolerance readings for homemade filament extrusion lines?
Agreed, evolving your design with a 3d printer is awesome. I just went through 5 revisions on a new CF/PEKK extruder design and I'm finally happy
This is awesome, as always thanks for your contributions to the 3d printing community.
I hope now the duet guys will show a little interest in this feature, in the past they weren't interested when I requested it.
How about having the magnet and sensor on two opposing lever arms, instead of only one? Then both can swivel together to follow kinks in the filament without compromising the distance reading.
What if the filament is oval in shape, with the minor and major diameter fairly different, but the cross sectional area is still equivalent to a 1.75 mm diameter filament?
If your sensor rides on the major or minor diameter, its "Effective" filament diameter will change significantly.
It would be super to see the filament maker working!
Excellent engineering! This answered all of the questions I had about the sensor from the diameter measurement video.
I would like to see your take on an extruder.
My main reason for wanting this is that I make my own filament from PET bottle. It's not 100% good consistency and sometimes a short length will become oval as it does not shrink down properly, thereby causing jams. If I can detect that early enough I should have very little failed prints from then on.
Great work! How do you make sure that the drills for calibration don't interfere with the Hall-effect sensor? Couldn't the metal lead to magnetic flux?
Hi Great product but on major flaw that nobody had though about it cause event filament is breaking inside filament loader it will no longer to be able to detect it! Cause it primary function is only detect missing filament! It's need filament movement detector like consumption it will be done by adding a roller that will detect if roller is stopping! With that feature you could say 100% bullet proof on both case
wondering if you could take a raspberry pi zero 2, and wire a filament sensor straight to the GPIO pins, then have a tiny raspberry pi zero filament sensor and octoprint server AIO? :P
did you add a cut off for if it runs out of filament?
Hallo Thomas, I understand the filament is assumed to be round and not oval as you mentioned at the beginning of the video, so the diameter of the filament is only “measured” on one “side”, let’s say the x-achse (z being through the length of the filament), so the y-axis is not measured, which could actually then carry the off-diameter...
Wouldn’t a 60° arrangement help to improve eccentricity problems? Of course that would require more work than already needed, yet only checking on one side basically ignores the other sides (60 or 90° offset) just my thought
Grüsse aus USA 😉
Future iterations probably would be to measure 2 or 3 diameter measurements, and if that works out people can make a very own average quality extruder and filament and stil get good quality prints making filament cost and selection phenomenally cheaper.
Great work, thanks for sharing!!
Hi Tom, nice design. So I made one, unfortunately for me, the hall sensor is too sensitive and was picking up magnetic field from my extruder motor. Whenever the motor was running the measurement was all over the place. Do you think this could be solved using two hall sensors measuring in opposite directions to cancel external signals?
I was just called that on site the other day. How conveniently personal :p
Oops! Never again ☺️☺️☺️☺️☺️
Thermistors are linear, but hall effect is polynomial due to square of the distance stuff. I'm working on a clone that uses a pi pico; I was just porting the code but uh…I ended up reimplementing everything. I'm also experimenting with using a segment of a sphere to friction-fit the bearing, since you can totally lock up the center of the bearing and the outside rotates because that's the whole point of the bearing.
That changes your $3.20 custom board to a $4 (€3.36) pi pico board plus like 5¢ of headers (or some solder) from AliExpress. Not free, but not terrible. With a little luck, I might be able to knock out some of the fasteners, as I think I have a design using three 623zz with no dowel pins or screws, plus the magnet and hall sensor. Maybe €0.35 more expensive than yours.
This is a good design and your ideal of using a V-slot bearing is a good idea as well.
Just to build on the idea a bit, using the V-Slot bearing on the moving arm of the sensor and then using a wheel with a bit more grip on the other wheel with a slotted disk and a photo interrupter would allow for the speed of movement or distance moved through to be measured.
This combined with the filament diameter could then not only detect the diameter of the filament but also filament out (can already do this :-) ) but also if the filament jambs of breaks past the sensor.
Just a thought. :-)
Keep up the good work
Ray
May be add some sort of encoder on the bearing on the input and that could be used as a filament motion detector/sensoe
If you add an encoder off the idler bearing, then it could also be used as a filament jam/runout sensor.
Thank you Tom, great topic. Nice work.
I like this, any way you could do a v2 that also includes an encoder? I have a printer (doing 3mm filament) that often jams and "prints" nothing. Adding an encoder to this, you could have it not only feedback the width of the filament but also either a pulse to show filament is moving (marlin supports this already! BTT smart filament sensor) or even recorde the rate of motion for closed-loop control
Instead of using the separate board, I wonder if it could just be added as an input to a Raspberry Pi running Octoprint? Then if there was a plug-in you could use the information that way? Thanks for presenting this Thomas.
The RasPi does not have analog input, but I could absolutely see the I2C interact wired up directly to the RasPi instead of the printer's mainboard.
This is awesome! Let's see it in action!
Also, I'd love you to show us your old filament extruder.
Thank you! Note my problems are order of magnitude higher, that it a tenstion measurement device for binding or missing filament.