This one might be an oldie, but it’s certainly a goodie.
Way back in 2005, [David] and [Charles] needed a project for one of their engineering courses. With so many loose resistors scattered over the lab, they decided to build an automated resistor sorter (PDF warning) to separate these resistors and put resistors of the same value together in the same bin.
The electrical and programming portion of this build is relatively simple – just a PIC microcontroller reading the value of a resistor. The mechanical portion of this build is where it really shines. Resistors are sorted when they pass through small plastic tubes mounted to a wooden frame.
There are several levels of these tubes in [David] and [Charles]’ sorter that move back and forth. The process of actually sorting these resistors is a lot like going down a binary tree; at each level, the tube can go right or left with the help of a solenoid moving that level of the frame back or forth.
[David] and [Charles]’ project wasn’t entirely complete by the end of the class; to do so would require 8 levels and 128 different tubes on the bottom layer. Still, it worked as a proof of concept. We just wish there was a video of this machine in action.
Tip ‘o the hat to [Alexander] for finding this one and sending it in.
To sort more values simply have the first 7 values go in the correct tube and everything higher in the 8th tube. Then run the sort process again on the resistors that go into the 8th tube.
I am impressed that they were able to get a mechanical design that works.
Seriously, did they use word and paint to make the document and included figures? This is one of the worst reports I ever seen! Not only the figures get all pixelated, they didn’t even cared about using a fixed width font for the source code!
Other than that it is an interesting project but I can still think of a few designs that are simpler to implement.
Seems fine to me all the pix are there and all the figs were clear.
yeah, concept seems quite fun, but program looks strange… why calculating it in excel and then hardcoding values? the second pic soft looks even more magical… and hmm.. ‘pcb board’ ? ‘printed circuit board board’ ? ‘software on the pcb’ ?
And they keep referring to the micro-controller as a PCB.
sourse-codez haz fonts?
I’d rather use some kind of revolver rotated using some servo (or stepper if we’ll be talking about tens of kilograms of resistors). This would make much more compact build even for higher number of containers when compared to this “binary tree”.
Why did each level even need to have that many tubes? Seems like a lot of energy was spent making sure each level had (n-1)^2 tubes and could move far enough to adjust for the next level down, when each level could have just had 2 tubes and moved a lot further. Skimmed the PDF to see if there was a reason they treated this as a whole tree instead of treating each stage as just the branch of a tree, but I couldn’t find one.
They don’t say how they measure the resistance and how they feed the resistors one by one into the system, especially as the aim was to sort resistors “scattered across the workbench”, meaning potentially bent. Seems to me that would be the hard part.
Exactly, that’s the first thing that popped in my mind. I was hoping for a diagram, even a youtube of the resistor aligning and measuring mechanism.
The sorting mechanism could be a simple rail through which the resistors roll and a solenoid and trapdoor for each resistance range. When the resistor rolls over the appropiate trapdoor, it opens and the resistor falls on the correct bin.
Back in high school, we made a marble sorter with a similar design mechanically. By far the hardest part of the project was feeding the marbles in one by one, which seems like it would be infinitely harder with resistors. I got all pumped when I saw this article because I thought they would have made some awesome feeder, but alas, I must say I was a bit disappointed. Also, the board leaves a little to be desired, but if it was someone’s first board design, not bad at all!
Oh, and it’s good to see someone else that uses Ultiboard! I have since switched to Eagle, but for a while Ultiboard was my program of choice.
@Hardware Problems: Use liner bearings, not (LOL) Teflon tape.
@Software Problems: Disable unused peripherals, other pins will magically start to work.
Link seems dead. I happened to mirror it: http://aricciodotcom.files.wordpress.com/2013/01/dg_cs_finalreport.pdf
this is where is it located on the website
http://www.swarthmore.edu/Documents/academics/engineering/E90/dg_cs_FinalReport.pdf
Yay!
God, I __hate__ link-rot.