Architecture for modular, expandable EV platform?

brookst

1 µW
Joined
Nov 18, 2021
Messages
2
Hi all. New to this hobby and this forum -- thank you all for the great conversations I've lurked and read.

I'm looking to develop a small EV experimentation platform, with the goal of being able to support anywhere from one to four motors, in various go kart / buggy / moped / other configurations.

I've got big picture questions:

  • Are there best practices for multi-motor configurations? Should I be looking for single batteries that have the amperage to drive all motors (perhaps one battery for 1/2 motors and another for 3/4, just to optimize for size and weight on smaller experiments)? Or is it cleaner to do one battery per motor? If one-per, is it wise or dumb to connect in parallel to share load/recharge?
  • Is there such a thing as a multi-channel speed controller that can drive multiple motors at independent speeds (e.g. turn outside wheels faster on a corner)? I'm not turning them up in searches, but may not know the right terminology
  • Are there off-the-shelf ECU type microcontrollers that provide Arduino/etc programmability with the hardware for canbus, hall type outputs, etc? Or is everyone doing homebrew for the microcontrollers?
 
brookst said:
[*]Are there best practices for multi-motor configurations? Should I be looking for single batteries that have the amperage to drive all motors (perhaps one battery for 1/2 motors and another for 3/4, just to optimize for size and weight on smaller experiments)? Or is it cleaner to do one battery per motor? If one-per, is it wise or dumb to connect in parallel to share load/recharge?

If you parallel the batteries, you've effectively only got one battery for the whole system.

If you need "critical redundancy" for a specific application, you could use separate batteries for each motor/controller.

If you just need independence as a capability in case of failure, you could do that but with the ability to parallel the batteries, or parallel them normally but be able to disconnect them, etc.

Any switchable-power system will cost more and be more complex, and have more points of failure. But the ability to do certain things with them may be worth any or all of those.


[*]Is there such a thing as a multi-channel speed controller that can drive multiple motors at independent speeds (e.g. turn outside wheels faster on a corner)? I'm not turning them up in searches, but may not know the right terminology
"traction control" might cover that, terminology-wise.

I don't know of a single controller that can directly do multiple motors in EV sized applications. (there have been some for ebike-sized stuff).

Sevcon probably has a controller that can be interlinked directly with others of the same type and setup to be operated like this, but Sevcon is not really a DIY-friendly type controller. They are notoriously difficult to setup for non-experienced users, and the company isn't prone to support a non-large-OEM user (and not end-users at all). They're also not cheap, nor small, so arent' particularly applicable to non-large-EV projects.

The Lebowski open-source controller brain here on ES is capable of some interccontroller communication, but this may be limited to the throttle control (I can't remember). It is just a brain, so you have to build a power stage yourself, or adapt it to an existing one. (there are numerous threads about these).

However, if you are intending to use a central MCU system anyway, then you simply use that to read your sensors and output appropriate control signals to the controllers for whatever response pattern your EV will need. If you create the menu system and settings with "any" usage type in mind, then a system like this should be adaptable to any EV of any size, from kid's cars thru ebikes thru cars and trucks and beyond. I would recommend looking at systems like the Cycle Analyst v3 for some idea of some of the control and limiting types of features you might be able to do, and then define various use-cases of EVs you'd like to use the system for to come up with other possibilities. Then define options and a menu system that would allow use of this control system with any set of controllers, with whatever control inputs those might have available.



I have a thread about Arduino Nano "tidbits" that discusses certain functions, like traction control based on steering input, etc., but I haven't gotten beyond the concepts yet, no hardwre or software yet exists. The discussion of various features may still be useful.
 
Thanks so much @amberwolf -- exactly the kinds of thoughts, advice, and references to existing projects I was looking for.
 
Separate controllers will run their respective motors at independent speeds. If all you want is the function of an all-wheel differential gear, simply use separate controllers with a common throttle input. (Connect the +5V wire from only one controller, but signal and ground wires from all of them.)

If you want more active directional control (commonly called "torque vectoring"), then you will need a more sophisticated control unit that synthesizes a different throttle signal for each controller. That's a lot more complicated, but the motors and controllers are still simple.

There's a small chance that connecting several controllers to the same battery could introduce enough electrical noise to cause a problem (especially if the battery current capacity is marginal), so using a separate battery per motor would eliminate that possibility.
 
Back
Top