Filament profile sensor, milestone one finished. Still a ways to go before having something useful, but the initial testing went as expected.
Auto bed leveling; filament sensors…
Someone’s gonna put together a commercial grade reprap soon. =D
Very cool, even some $7200 printers don’t have this yet. Har har gibe gibe.
Nice! What kind of sensor are you using?
URL for someone that does not understand what this means?
@Guy_Sheffer This sensor measures the diameter (and shape?). You have to know how much filament (plastik, ammo for 3D printer 
) will be pushed through the hot end (the hot nozzle that prints) for exact prints.
@Thomas_Sanladerer It’s a tsl1401 line sensor, the same others are using.
@Guy_Sheffer I have filament that is 2.84mm, some that’s 3.01, and one roll varies measurably along its length. Each time I switch filaments I have to recalibrate the extruder. This is an attempt to give the printer enough intelligence that it will adapt on the fly during printing.
It’s also a tool that people can use to perform quality assurance testing on filament suppliers. Doing it by hand with calipers is tedious, but this sensor should be able to take a thousand measurements a second. Take a power drill, attach an empty reel, put the filament through the sensor then to the empty reel, and check a whole reel of filament in a matter of minutes.
Should be a fun project. But this milestone was simply a software test. It’s going to get a whole lot better over the next few milestones.
@Adam_Davis i was asking because it look so very different from my TSL1401 
What kind of accuracy are you getting with that lensless setup?
And you’ll definitely need to use a dedicated microcontroller for the sensor readout - RepRap firmwares use interrupts all over the place which would either mess with the readout sequence or, if you disable other interrupts while doing the readout, you’d interfere with the bot’s movement.
Next up is positioning sensors.
Yep, it’ll have its own processor. Probably an uno processor for the simpler version, and a due processor for the performance version. The sensor pictured is the larger smt package. I have a couple of the smaller smt sensors too, but they are slightly more expensive so I’ll probably stick with this package. That’s a ground down 5mm LED on the right for size comparison.
As far as accuracy, at this point, I’m getting about 0.1mm accuracy, which is nearly useless. But I wasn’t expecting it to be this good anyway, and accuracy wasn’t the point of this milestone.
The next milestone will be using two sensors, and at that point I can bring a lot more math together and get a good measurement. At minimum without stretching I should be able to get at least the resolution of the sensor, 62.5um. But with some interpolation I should be able to increase that. It’ll all be open source though, so once I release it I expect others will improve it further.
@Adam_Davis you should checkout our community around universal filament identification. So far @Thomas_Sanladerer has shared his version of a similar device and we’ve got people working on firmware/software around this problem and collaborating.
https://plus.google.com/communities/107859862288161234107
We should be doing a hangout next week to keep the discussion going.
@Ross_Hendrickson Thanks for the pointer, I’ve joined.
I’ve wondered if you could use laser mouse components to do something similar.
Yes, but the arrays are smaller (so they can fit more per wafer) so you’d have to use optics, and still might not get enough resolution. But they are very useful for sensing filament movement and closing that loop…
Filament sensing is optional but inevitability.
@Josh_Ajima I went for the movement part of it with the ADNS9800 over here: @eof_cr_lf The project has been postponed due to waiting for my foldarap (but still believe it’s worth the patience @Emmanuel_Gilloz ).
Another approach I like very much by @Mark_Heywood : Airtripper Extruder Filament Force Sensor.
Impatiently waiting for turning this theory into practice.
Interesting to see this development going