I do in-house backups for my data, but the system that eats the data now is an all-in-one Intel Atom PC with a dying fan, and I don’t really trust it. Luckily we’re in the future, so fanless ARM-based computers are everywhere, so I built a little backup machine out of one. Details after the break.
Sorry I haven’t posted more. I have done some cool projects, but finding time to post has been hard. To help fill that gap, here’s a report I wrote up detailed a small chapter in my involvement with Team Blue Devil Ocean Engineering, which is Duke’s entry into the Ocean Discovery XPRIZE, a world-wide contest to develop the technology to map 500 km^2 of ocean floor in 24 hours. It presupposes some knowledge about the project, which you can find in this brochure or even these slides, or you can just dive in and have fun gawking at this crazy thing we built, sunk 2km deep in the ocean, retrieved, then debugged.
Details of an intense 48-hour effort to build an deep-ocean-survivable Arduino control circuit for underwater rockets is after the break. It was written as an after-action report for the project, so the language is a bit drier than usual, but I think it’s still a fun read.
It’s basically a really weak leaf blower, but it looks like a giant bazooka, and it was $5 at the time.
It runs on obsolete “VersaPak” batteries, which were part of a proprietary tool battery system from at least 10 or 15 years ago. I had an old electric screwdriver that ran on one of these around then. Each battery was a 3.6V NiCd pack, and this unit uses two of them at once to run at 7.2V.
I wanted to revive it, but I wasn’t going to be hunting down decade-old batteries on ebay to do so. I decided to use 18650 rechargable Li-ion cells (the same kind used in laptop batteries) because they’re 3.7V each (close to the original’s 3.6V) and I had a bunch lying around.
I popped the housing apart, drilled a hole in the side, and added a 2.1×5.5mm panel-mount female DC barrel jack that hooks to where the VersaPaks would connect in. Now I could power it from my bench power supply to verify it worked, and it did.
My first attempt at a battery solution was a single pair of 18650 cells in a cheap ebay case. This worked, except the wires that came pre-soldered to the battery case were a very small gauge, and I actually felt these wires heating up when I used it. This was no good – I was a bunch of my power to wire resistance! Also, the thing ran the pair of 18650s down pretty quick.
I don’t have any pictures of any of that because I did that part a long time ago, then lost interest. Recently, I felt like doing a little project, and the thing was sitting here, so I finally finished it.
I got two new two-cell battery cases, and snipped their tiny pre-soldered wires. I soldered on good 22 AWG wires direct to the outputs, and hot-glued the wires for mechanical strength. I ran both sets of battery pack wires to a single male DC barrel plug, so as to run two pairs in parallel (2S2P in battery-people speak). I screwed the two cases to the housing (there’s plenty of room to screw into where the VersaPaks used to plug in) and plugged it in, and presto. No more warm wires and much better battery life.
I like this thing because it can quickly blow sawdust and stuff in the shop without sending screws and heavier stuff flying around. Also, Reginald is terrified of it.
I’ve come to realize that I can’t run the MPCNC in my office — it just kicks out too much dust. I could add a vacuum, but I bet there’d still be a bunch kicked off. Therefore, I’ve modified the machine to be portable and wifi-enabled, so I can take it to the shop out back.
To do this, I’ve done three main things:
- Attached a Raspberry Pi running OctoPrint, with configuration made so I can upload gcode via a Windows share (samba).
- Added handles to either side, and eyelets with rope for a shoulder sling, allowing it to be tipped over, collapsed, and carried out by myself.
- Protected the electronics with a removable cover made of hardboard.
Details after the break if you’re interested.
The pen holder was actually somewhat complex. The pen I wanted to use was a Bic 4-color, and its barrel has a very slight taper to it, so I had to model it fairly precisely and it took two prints to get the dimensions right. Now that it’s done, though, I can get really nice, repeatable drawings. I’m thinking of developing an algorithm to convert images into combinations of the pen’s four colors and emit gcode to approximate color printing. It will be ugly due to color theory (there’s a reason printers use cyan/magenta/yellow/black), but it might look neat.
The MPCNC with my rudimentary pen holder was able to draw this. I need to add some kind of spring downforce on the pen to deal with small differences in depth (which is why the lower left is faded). That should also help with pen accuracy, because on the high parts it’s grinding the pen too hard against the page and its getting stuck. I also need to turn up my overlap to get it to solid fill better.
I tried a test run of the Optimal Fabrication Test Model (it’s dickbutt…we talked about this).
Results were…mixed. It started out strong by making key outline portions, so I left the room. I came back when I heard the spindle inexplicably struggling from downstairs. I come back to find the spindle has sunk all the way into the foam, and the nut that mounts the endmill to the tool has itself ground a sizable trench through the foam. Bits of foam are everywhere.
It turns out two separate failures happened. First, the Z coupler came loose…that’s my fault, as the plans called for nylon locknuts, but I couldn’t find any locally, so I used plain hex nuts, so the screws tightening it vibrated loose. Locknuts: ordered.
Second, there was a flaky connection on the X-axis, which is why it’s a vertical trench instead of a vaguely dickbutt-shaped trench. Apparently the wire I used doesn’t like to crimp well in Dupont connectors, so I crunched them all harder and added a bit of solder to be sure.
I 3D printed a pen holder to attach to the tool, so I can do ink-based tests while I work out the kinks. I did a dickbutt print on the same crappy foam, offsetting it a bit upward. The result is surprisingly good, given that the foam is nowhere near level, and I used the same program that assumes a 1/8″ endmill. Look at the solid ink on those eyes…nice!
I was building a little robot, and I wanted belt-driven four-wheel drive using GT2 belt (an inexpensive timing belt common in 3D printers). The problem: I only had bulk GT2 belt, not closed loops (see extremely advanced diagram on right).
For my first attempt, I used end-slips to cut out the teeth from part of the belt, then super-glued it to the back of the other side. This works okay, but the joint was very inflexible owing to the near doubling of the thickness, and it eventually came apart after a few days use.
I’ve since found a better solution. What we need is a backing: a material I could adhere to the back of both sides of the belt, and which would be strong, flexible, and compatible with superglue.
I found such a substance in my trash can! When you buy 3D printing filament, it usually comes vacuum-sealed in clear plastic to keep it dry. The parts of this plastic that have been heat-pressed together are strong and flexible without stretching — it’s the perfect belt backing!
So I cut a narrow strip and super-glued it to the back of one end of the belt. Superglue dries hard, so I put down the glue in stripes so that the belt would still be flexible. After it dried a little, I glued the other end of the belt on, making sure to provide glue at the joint itself as well.
I let it dry under a weight for a few minutes, and the joint came out almost as flexible as the belt itself. The resulting bond has no give when stretched, goes over pulleys well, and survived the maximum tension I could put on the pulley by hand.
Pictures of belt on the end result (a little yellow robot):
Here’s the short story in picture form: