Fyberpunk

It's oversized! It doesn't have good wire routing yet! It can put an industrial sewing machine motor's worth of horsepower into a domestic! I even managed to break the textolite gear! It's, uh, not going to be this way when I get around to making kits eventually, but this is an interesting stage of development to say the least.

How we got here

I finally got a precise measurement of the motor shaft with one of my dad's micrometers, which I used to finalize the CAD model I showed in the last post. Then I ordered it from a shop in China, with pretty precise tolerances. No problem, 10 shafts for about $40 a pop. I could've ordered a hundred for surprisingly little more, but I decided that sitting on that many would be silly.

I ordered all the motor-related parts I'd need (possibly sans one cable) from the ODrive store a while back, and they arrived before the shafts, which is all I really needed. The shafts themselves arrived the day after I left on vacation, so the project had to wait until I got back from New York. It was a lovely time, by the way.

So, shafts, motor, and motor driver in hand, I still needed a few more parts to put it all together. A coupling to adapt the motor's 10 mm keyed shaft to the 8 mm stub I left as a drive input on the machine, importantly, but also a few brackets to hold the motor to the machine and the ODrive S1 board to the motor. I ordered a jaw coupling for the former, and 3D printed some pieces for the latter. Finally, after assembling the whole thing, I also got a 600 W, 48 V power supply to run the thing, which means I should be able to get up to 0.8 HP, times the efficiency of the whole system. That's an industrial sewing machine motor's worth of power!

The whole thing ended up quite a bit longer than I would've liked because I didn't have a good way to cut down the motor's shafts. I now feel like that's a good thing, because I know the current setup is absurdly overpowered for the machine and I need to tone it down for any version I try to share with other people. I don't want to be responsible for breaking more of these nice old machines than I do personally, hah.

What I'm alluding to is that I managed to break the original textolite worm wheel. This was only by running the machine unloaded with high acceleration and jerk. I managed to crack a piece out, probably not coincidentally right next to the metal pin driven through a hole in the textolite. So after ordering a replacement from eBay for $15, I got back on my merry way, now with half the acceleration limit I had set when I broke the gear. No problems yet, fingers crossed.

All this time, I was running it without any position feedback for the needle, just for the motor itself. Since the machine has a gear ratio of 10:51, I could still command it to go exact numbers of stitches forward, but stopping the needle at a particular place wasn't really possible without a homing procedure. I foresaw this problem, so I also ordered a magnetic angle sensor from the ODrive store, which I'm using to sense the handwheel's angle on the machine. I printed the bracket and mounted it a couple days ago, attached to the boss where the original mains plug was. I got everything tuned yesterday, and it's running great.

Where it stands now

The machine runs smoothly and with unbelievably high torque, from about 30 stitches per minute up to 1100, where I've set the max speed in the ODrive's configuration. It runs slower than that, down to certainly 6 stitches per minute, but loses its torque at such low speeds. With my current, hopefully safe, acceleration limit of 40 stitches/s², it can come to a complete stop from half speed in about one stitch, or from full speed in about four stitches (square laws strike again). I've decided that's just fine, and I don't need to push my luck, since usually I'll slow down before I stop at the end of a seam anyway.

I'm not happy with the synchronizer bracket yet, as it's much too thin and flimsy. It's easy to bump too close or far from the magnet and make the ODrive panic from losing position sensing. So I've got to stiffen it up a little more. It's in a good location on the machine, too, so I'll be adding some switches and buttons to it probably, as well as using it to hold whatever microcontroller board I end up using to run the show.

I don't have any physical control for the machine's speed yet. I temporarily hooked up the original rheostat pedal as an analog input to the ODrive for velocity input, but it doesn't work very well, and I want to find a better pedal to use. Something meant for precise and fast readings, rather than speed control of a universal motor. Expression pedals for music come to mind, which might be the route I go since I can also use it for music. I've also thought of vehicle accelerator pedals, though floor-mounted ones of a shape I like tend to be expensive. I've also thought of building something bespoke, like an encoder, pulley, and belt attached to the end of a pedal. But using off-the-shelf parts means when it eventually breaks, it's easy to fix or replace, so I'll probably just buy an expression pedal soon.

Yes, that means I can't really sew with it yet. But I'm not sure whether I'll try to make the foot pedal work first, or jump straight in to making the embroidery unit I envisioned for it when I bought the machine. Sure, I can do both, and I probably will, but I think I'll try to finish the foot pedal before I really get working on the embroidery unit, which is a whole little CNC project that I'm looking forward to.

For the future, and what I think I'll actually document enough for other people to reproduce it, I'd like to use an ODrive Micro and a much smaller BLDC motor, so the thing is limited to no more than 100 W. By my reckoning, I should be able to fit such a setup inside the original motor's housing, which would be fantastic. It would look more original, and maybe even let the machine still fold into the cabinet, but upgrade it with a bit more power than it originally had, and servo control. I've ordered five ODrive Micros to build some stuff with, but they're not supposed to arrive until January because it's a crowdfunding run. So this improved, public version will have to wait until early next year, I'm sorry to say. I guess if someone really wants to overbuild one with an ODrive S1 and this huge motor that can break the machine without much difficulty, I could release the designs for what I have now once it's a complete sewing machine again. But really, I don't want to be the reason for breaking someone's family heirloom, so proceed at your own risk if I decide to release it.

I'll leave you with a demo video.

⁓Clara