I mentioned previously that I was working on a project for my side hustle, but didn’t say what it was.
This is it. It’s a 32×32 px LED matrix, battery, and supporting electronics.
The first purpose will be a time trial start clock. (This shows the countdown for each person leaving the start house) CrossMgr, the software I use for timing, publishes information for the start clock over a WebSocket connection to drive a web version of the start clock. I wrote a python script that slurps that data and drives the LED matrix, counting down the time, as well as the bib number who’s next to roll out.
The project combined a bunch of stuff. Software, obviously. Electronics, as I integrated a bunch of random stuff including the LED matrix, a Raspberry Pi, voltage regulator, battery, buzzer, cooling… I 3D Printed the mounts for the boards. I built the box (from scrap wood!) that holds it all.
Considering a commercial version of this is a grand or more, I like this version.
I’ve been soldering more things, and with soldering comes fumes that you don’t want to breathe in.
So, after a couple of hours in CAD, sixty hours of printing, 3.75kWh of electricity, and a few minutes of assembly, I have this beast.
Inside is a 140mm fan, so it’s nice and quiet, and 130mm replacement charcoal filters for a commercially available fume extractor. The filters friction fit in front of the fan, so it’s easily changeable. I have a 12V wall wart power supply and a switch on the back.
I’ve got a project in the works for my side hustle, with the parts coming this weekend, so this is just in time.
Previously I posted the corner clip. After that post, I went at it with a knife. Unfortunately, it ended up splitting along a layer line. But actually, it was fortunate. I was able to clean up the spinning insert, and using some CA glue, fix where I split the layers. I used it in building the frame for the enclosure and it worked perfectly.
There are downsides of having a 3D Printer on the workbench in a “woodshop.” One of them is losing a workbench. The other one is dust. Dust, everywhere.
Deciding to solve these problems at the same time, I built an enclosure for the printer. It is super simple, cobbled together with pocket holes. The walls are acrylic I found at Lowe’s, with some cheap hardware for the door. Currently, all of the electronics are still inside, something I plan on changing. I want to move the control panel outside of the cabinet for convenience, as well as create an interface panel for OctoPrint to control manual XYZ movements. And a big red button for stopping jobs. Everyone needs a big red button.
Speaking of OctoPrint, not being on the workbench means I need a new place for the camera, as well as some lighting. To provide the lighting, I bought some cheap LED strips on Amazon which I’ll affix to the inside of the cabinet. I found an old computer fan. after fitting with an air filter I want to use it to create some positive pressure to keep the dust at bay.
With the lights and OctoPrint, I obviously need a camera too. I had a Pi camera from another project which I decided to use instead of the Logitech camera I was using previously. But, with a Pi camera, you need a way to hold it up.
This is the best part of the project for me I think. I found a model of the camera PCB on GrabCAD and it was perfect. I was able to import it into Fusion and design a mount around it perfectly. On the first print, I got a perfect articulating joint and bracket and a perfect fit around the camera. Maybe I’m getting the hang of this. My only regret is that I built it at 0°, but it is one part to reprint to adjust the angle.
So this was the plan. I saw in a catalog a widget that would let you clip your tape measure to a corner so you can easily check square. “I could make that.” And so I tried.
It kind of worked, but not as well as I wanted.
This was my first attempt at a print in place assembly, where the swivel clip sat inside the main body, but was supposed to have enough margin that they would print as two disconnected bodies. If you look closely, it didn’t work. I tried loosening it up with a knife, and if it still doesn’t work I’ll have to do an autopsy on it.
Back in January, I posted a rendering of my plans for a bench. By early March, I had “finished” it!
Since the picture to the right, I completed the dust box and insert plate for the router, added T-track, installed electric and dust collection, and ran the jigsaw into it while cutting a piece (oops).
The table saw stands just proud of the table, making an excellent outfeed table. Dust collection for the saw routes to a blast gate just to the right of the saw, and there’s electric just behind it.
The dust box for the router is connected directly to the 2″ pipe for dust collection. With the Shop Vac, it catches a huge amount of dust off the router (it misses the larger chips but that’s fine). I mounted a safety switch on the center leg, which turned out to be perfect height to hit with my knee to stop the router.
I found myself working primarily on the “back” side of the workbench between the miter saw and the bench, so there’s also a dust port over there, as well as electric.
I still need to build a fence for the router, and I need a better place for my 3-D printer.
The other day, I got frustrated trying to connect my vacuum hose to my Kreg Jig. I looked at the spring clamp I was using and sighed, then grabbed my calipers.
Based on this Reddit thread, I purchased an Ender 3 from MatterHackers. So far my most successful prints have been vacuum adapters and a Raspberry Pi case.
Speaking of Raspberry Pi, before my printer had even shown up I had discovered OctoPrint, a program to manage your printer from a web interface. Being able to see my printer (through the webcam interface) from anywhere, abort a print if it is going haywire, and fine-tune nozzle or bed temperatures has really made it easy to work with the printer. I’d highly recommend it. Oh and of course you can print your own case too.
Being able to print whatever I want (reasonably), whenever I want (reasonably), is amazing. The future is here.
I’m a nerd, a fast I’m sure no one is surprised by. One of the side effects of this is overcomplicating simple things.
Because I’ve always had a cloud server for hosting things like the ADT Pulse/Pushover integration and this site (among others), I decided to complicate that by running a Kubernetes cluster at Digital Ocean for the ultimate in ridiculous personal infrastructure costs.
That said, it’s been super easy to work with. I host 8 WordPress sites, the backend for Holeshot Events, as well as some personal apps for home stuff, on 2 $20/month nodes. Could I do this more efficiently? Probably. Would it be as ridiculous? Nope.
It’s been a great learning experience, and with the release of k3s I’ve been thinking about installing it on the home server too.