You guys see the Mechaduino Kickstarter? It’s a very tidy and fairly cheap little servo board that you bolt onto a standard stepper to make it run closed-loop. Lots of robotics hacking potential, or it can accept a step/dir signal for 3D printer control.
Two things I really like about this:
- The encoder approach is dead simple and works on single-shaft steppers, it’s just a chip on the PCB and a special magnet you attach to that small exposed bit of the rotor shaft on the rear of the motor.
- Unlike previous open-source closed loop stepper attempts I’ve seen, they’re doing proper field-oriented control. Without getting into the technical details, that’s a fancy way of saying it can apply 100% torque at any time with near-zero angle error, and it will send less current to the motor when it needs less torque. Basically treating the stepper like a BLDC motor with 50 poles.
I still have some technical questions (which you can see in the KS comments page) but I’m backing it.
https://www.kickstarter.com/projects/tropicallabs/mechaduino-powerful-open-source-industrial-servo-m?ref=discovery
They do have stuff on GitHub and you might find more information on their hackaday project page if you haven’t seen that.
Yeah, glanced through that, didn’t immediately see answers so I just asked in the KS page. They’ve been answering everything posted there so far.
Do you think the motion would be any smoother/quieter with their driver system or is that asking too much?
Ummm. It COULD be very quiet. Really don’t know, it would take a lot of digging into the guts of the torque/velocity/position loops and such to figure out from first principles. Probably should just ask them. The servo loop is very much audible frequency, but if the motor current waveform is interpolated fairly smoothly between servo loop updates, it should perform similarly to an interpolating driver like the TMC2100 – which is very quiet at constant speed and somewhat louder while accelerating.
There is potential for buzzing/hunting noise with any PID servo loop if the gains are too high, but I’d be a little surprised if that’s an issue, since they’re auto-tuning PID, and there’s no gearbox between the motor and encoder to add backlash/deadband.
It could be good depending on the accuracy, and resistance to EMI. Curious why the use of a stepper when is should work on regular DC motors.
How is this different than the ustepper? http://www.ustepper.com/index/
Also a Kickstarter. GitHub here: https://github.com/uStepper/uStepper. I have some. We started a similar design before I received my rewards. Then got busy.
Ustepper is 50-some bucks, btw
@Brook_Drumm the Mechaduino is cheaper (at KS prices anyway), and much higher spec (48mhz 32bit ARM vs 16mhz 8bit AVR, 14bit vs 12bit encoder – these both make a meaningful difference in capabilities). And about ten times more importantly, the ustepper doesn’t do field-oriented control as far as I can tell. It’s just a 4988 and a microcontroller that catches lost steps. That’s not taking full advantage of commuting the coil energization and varying current for true torque control. I don’t really want to spend the next half hour explaining why in a G+ comment, but all serious industrial stepper-servo systems use field oriented control… Just tracking and repeating lost steps is only scratching the surface of what you can do with a stepper.
As for why it’s better than DC servos, a true hybrid stepper servo system achieves the highest power density of any small motor package. Hybrid PM stepper motors are extremely efficient and powerful when you don’t have to keep large safety factors and run them at full current regardless of load in open-loop mode.
Plus… All our printers already have NEMA 17 motors, and bolting one of these onto the back of your existing motors is a pretty appealing retrofit compared to adapting everything to new motor form factors.
To be clear, I don’t know all the details yet, but it looks very appealing based on the info available.
Some of this makes sense but ustepper does use a magnet (as?/with) encoder. I’m not an electronics guy. And it does bolt up to Nema 17. Our plan was to make a Nema 23 version. If this is better, I’m all for it. I was curious if the latest Kickstarter project is based on and improves the prior design… If do, I assume it is open source. Seems likely. Since we make our own electronics, we could offer them cheaper. I’ll want to see how this one plays out before I pick up our version of the project.
@Brook_Drumm “I’m not an electronics guy” ha, don’t let him fool you
@Brook_Drumm I think you’ve got it right, it’s definitely the same form factor and encoder concept as the ustepper. Same open source license too. Probably reasonable to assume it’s a derivative in some sense or other. (The similarities I have seen so far are in the functional hardware implementation, which is not part of the CC license, so attribution isn’t strictly required unless they reused code.)
From their project log on http://hackaday.io it seems like they may have started by experimenting with the ustepper design with the intention of improving on it (since they started with the Atmega328 as an MCU) but then moved on to an Arm M0+ MCU and succession of magnetic encoders before ending up with their final design.
I guess it’s a bit of a cousin design to the ustepper, then. It does seem a bit more capable just from spec sheets and components, but it’ll be interesting to see what real world performance is like and just how beneficial it is.
I love that banana. Random.