OpenSource DIY, simple, motor controller

[casainho]

I want to have advanced control of the motors and full system of different EBikes motors I have and electric scooters. Over the time I saw the manufacturer changing the motor controller on the same motor, for a totally different microcontroller and then we had to start again wall the work to develop our OpenSource firmware. And it is very difficult, because sometimes the manufacturer only sells at high prices and / or use a very specific, unknown microcontroller and we need to learn everything again.

My idea is to develop a DIY, simple, motor controller that can be reused between different devices like EBikes and electric scooters, that have similar range of power on the motors. This motor controller must have space for custom hardware, to handle the differences between devices like torque sensor reading circuit of EBikes and that are not needed on the scooters.
O motor controller like this will be easier to repair and adapt to different devices, and will be future prof.

Some similar ideas:

• http://makeatronics.blogspot.com/2014/05/bldc-motor-control.html
• http://makeatronics.blogspot.com/2014/08/smart-bldc-commutator-hardware.html

Youtube video: https://www.youtube.com/watch?v=eMslHyssvhs

[Crow-Control]

This is actually a great, workable and detailed proposal :)

I also agree that this might be the best long-term solution, instead of building custom firmwares for 1001 different motor(controllers)

Biggest challenge for m500 and m600 would be also correctly supporting the HAL sensor and Torque sensor.

[casainho]

This is the same logic for the display. On SW102 display, they recently changed the hardware and is almost impossible to open the display and flash our firmware.

I am working on the display first as I have motor controllers ready for my EBikes, but then I will start with the motor controller. There are other developers developing the firmware for the M365 motor controller, that I am familiar with and I will reuse that firmware and popular STM32 motor controller for this custom motor controller.

[Crow-Control]

I think thats a reasonable way forward indeed.
It would also remove the need of supporting CANBUS right away (although it's the beter protocol of the two)

[casainho]

The question is that CANBUS although good, adds the need to have at least one or two extra chips. That adds cost and more complexity. Is CANBUS really needeed??
On EBikes and scooters as M365, only UART is used, even between the motor controller and the battery BMS. For the display, they even use 1 wire UART only!!

[Crow-Control]

The question is that CANBUS although good, adds the need to have at least one or two extra chips. That adds cost and more complexity. Is CANBUS really needeed?? On EBikes and scooters as M365, only UART is used, even between the motor controller and the battery BMS. For the display, they even use 1 wire UART only!!

If you don;t want to support the stock Bafang controllers, there is no reason to support CANBUS either.
I think the thing with CANBUS is mostly the fact it's much easier to daisy-chain and lock-down.

[nick2204]

Agree that an open hardware solution is a good way to go, but think you're much better off adopting a tried and tested design such as VESC6, on a suitably sized/shaped board. There are many examples of customised solutions to learn from and will have the benefits of a starting framework. To achieve decent FOC motor control you can't go too "cheap".

[Crow-Control]

VESC6

No one is going to pay 200 bucks for a constroller for a 600 buck motor.

Also: Just because something is DIY designed, doesn't make it beter or worse. Nor does more expensive means beter.

[casainho]

I am not a big fan of VESC because it used to handle only up to 12S battery pack while on EBikes is usually to have 13 and 14S.

I am more fan of simple FOC boards, like the one used on M365 scooter or other similar EBike and electric unicycle motor controllers.

Also VESC firmware is a bit complex for me, I plan to use a more simple one that I understand, this one: https://github.com/Koxx3/SmartESC_STM32_v3

My target is not the best FOC or the best motor controller. For me the best is the one I can easily reuse between the different devices and I can easily DIY and repair.

[nick2204]

That's cool, only a suggestion to make a quicker start. The VESC architecture is pretty flexible to different voltage/power, and I wasn't suggesting using standard VESC firmware.
I had assumed that others were keen to replace the horrible bafang controller/firmware for something with better performance and control as I am.
A controller capable of FOC is essential to achieve better performance than stock.

[Crow-Control]

That's cool, only a suggestion to make a quicker start. The VESC architecture is pretty flexible to different voltage/power, and I wasn't suggesting using standard VESC firmware. I had assumed that others were keen to replace the horrible bafang controller/firmware for something with better performance and control as I am. A controller capable of FOC is essential to achieve better performance than stock.

Afaik the M500 and m600 already use FOC

[nick2204]

Afaik the M500 and m600 already use FOC
Yeah they do, so if a replacement controller is going to improve on it then a simple design as pictured earlier in the thread isn't going to cut it.

[Crow-Control]

Afaik the M500 and m600 already use FOC
Yeah they do, so if a replacement controller is going to improve on it then a simple design as pictured earlier in the thread isn't going to cut it.

It's an example, but another example (the stock board) is also relatively simple and does support FOC.
Actually the stock board is not much harder, it uses some quite nifty tricks to support FOC as cheap as possible :)

Even with the STM32 it's perfectly possible to do quite decent FOC. Assuming the powertrain is and sensing is done solidly.

[Crow-Control]

However, what I don't really get is why @casainho is still calling the stock board expensive. It's 60 bucks excl shipping.
It's hard to do it cheaper.

[nick2204]

However, what I don't really get is why @casainho is still calling the stock board expensive. It's 60 bucks excl shipping. It's hard to do it cheaper.

Yeah I was just thinking the same thing, given the specs.
Any idea what changed from the 3.1 to 3.2 controller? Seems FLX only have stock of the 3.1 at the moment.

[casainho]

I have the idea the parts for this motors are expensive and harder to buy, maybe I need to update myself.

The TSDZ2 and M365 motor controllers are 30 bucks each, both implement FOC and similar power.