Programmable controller or custom hardware ? / wheelchair project

mitirino

1 µW
Joined
Oct 16, 2018
Messages
4
Hi everyone,

I need a custom controller for my wheelchair related project (or at least that's what I think I need, but any advice is appreciated)

I am trying to make this add-on device that attaches to a wheelchair and makes it into electric powered.
I want to use a 16" wheel and a normal hub motor.
I got a high torque version of a Ebike motor BLDC Hub motor For Sale | MAC motor.
They usually come with 8 or 10 windings, I asked them for 16, because my application needs high torque and can live with lower max speed.
They said the 16T motor will be around 700W but they didn't give a real number. The base motor is 1000W.
Their stock controller doesn't work with the 16T variation and they don't want to engage into a project to make one.

So here is what my ideal solution will allow me to do:
  • 3 or 5 assist levels (it's 100% assist so no pedals)
  • have the throttle programmed so the first 10-20% of throttle ramp up power very gradually, then the last 20-40% of throttle ramp up the power more sharply (to reach the max for that assist level). But have this mapping different for every assist level, so that level 5 for example can have a more aggressive profile from the start of the throttle range of motion compared to level 1.
  • work with a normal/generic dashboard that is not the Cycle Analyst
  • support all the other standard things like front light and rear light
  • be extremely durable. It's a hassle for everyone to deal with faulty electronics but for people with disability it is amplified, so I need a solution that's rock solid. Even at the expense of underpowering the motor (which I also like as a general idea so that the motor itself lasts as long as possible)
  • be able to continuously start from 0 km/h without helping the motor
  • have a reverse mode

Here's a video of my prototype (with my friend Quinto riding it):


It's with another motor on this video.


Any advice on an existing controller that will do the job is great.
Any advice on what needs to be programmed or built also.
If anyone here can do projects like this I am interested to talk with you.

Thank you

Dimitre
 
I would check with Grin Tech to see if they think the Phaserunner would do the basic job. (it does not require/use a display (dashboard)).

There are also a bunch of variations of VESC (also no display), but your best source of info on whether those will work for you is here on the forum (probably in one of the various VESC project threads).

Both of them are FOC controllers so can be tuned for a specific motor, and because they are FOC they use (or can use) the throttle to control motor torque rather than motor speed, so the way they assist is probably more useful to you than any of the typical controllers that use displays (dashboards), because those are almost all speed-control, not torque. WIth torque control, rather than speed control, you may not need as many assist levels or they may not need to work the way you've determined.

You may have to create a module (Arduino, etc) that does this the way you want it to, as I don't know of anything that does things the way you want it to (even the CA, though it's certainly more fine-tunable and user-configurable than the typical display/controller combos).

I'm not a programmer so couldn't help with the code, but the basic stuff you'd need is an MCU that has at least one analog in and one analog out, for throttle in and out, and however many digital I/O pins are needed for your control inputs and your display outputs (either for an LCD/etc or for individual LED indicators, etc).



FWIW, the Cycle Analyst is really the only "generic dashboard" (in that it works with any controller for all of the CA's functions including throttle modification, but it doesn't communicate with the controller if that's a requirement, but it displays all the typically useful information for EVs) I'm aware of (other than just plain wattmeters or bicycle computers); the other "dashboards" (controller-interactive displays that do communicate with the controller) all are generally specific to their controllers / brands / etc, (meaning, there isn't one that just works generically with all controllers).

So if you require one of those types of displays/controllers, then once you pick a specific controller, you can see if it has a display that does enough of what you want it to do. (or base the controller choice on which ones have a display that will do enough of the things you want).
 
  • work with a normal/generic dashboard that is not the Cycle Analyst
  • support all the other standard things like front light and rear light
I've just done a very similar project:
- bidirectional throttle for forward/backward, regen/braking in both directions
- works with any Kunteng or Kingmeter display
- proven, cheap Lishui controller hardware
- cruise control with automatic regen braking downhill
- backward beeper


regards
stancecoke

index.php
 
Last edited:
Thank you @amberwolf and @stancecoke for the replies

The Grin site says that the Phaserunner is out of stok, and there is only the Frunkerunner.
Is this also something I can try to just plug in and try? Is this also a torque controller?

I have not tried a setup with the throttle controlling the torque and not the speed.
Sounds interesting, but am not sure how natural it feels for when you are not pedalling AND throttling, but only throttling.

The arduino is something that I thought I might need to do too. Because I'm not an expert, it sounds a bit exotic to me and not sure how reliable it will be. But otherwise feels like the most attractive solution.

What do you think of these controllers?

- acceleratedsystems
- FTEX
 
The Grin site says that the Phaserunner is out of stok, and there is only the Frunkerunner.
Is this also something I can try to just plug in and try? Is this also a torque controller?
All the *Runner versions are the same basic design and firmware, so they work the same way. I recommended the PR because it is sealed/potted and so weather/vibrationproof. I haven't looked at the FR's casing/etc, so don't know if it is.

I have not tried a setup with the throttle controlling the torque and not the speed.
Sounds interesting, but am not sure how natural it feels for when you are not pedalling AND throttling, but only throttling.
It is much easier to control a powerful motor via torque control vs speed control--so if you need a system that has high torque, controlling the amount of torque directly is much easier than controlling the speed of the motor and trying to use that to apply only the amount of torque you need for any specific situation.

A lower power (relative to the vehicle mass) system is not as hard to control via speed because it doesn't usually have enough torque to be hard to control, but usually takes large throttle movements to cause changes in torque. Doing the same thing in a higher power system can make it harder to get just the right amount of torque--plus the amount of throttle needed to give a certain amount of torque changes depending on the speed the motor is already going, with a speed throttle. So with a speed throttle, you're controlling the speed the motor is trying to go, but the actual effect it has on the pulling or pushing the motor is doing will change depending on what's actually happening at that moment.

With a torque throttle, it controls the amount of torque directly, so you're controlling the amount of pulling or pushing the motor is doing, directly.



The arduino is something that I thought I might need to do too. Because I'm not an expert, it sounds a bit exotic to me and not sure how reliable it will be. But otherwise feels like the most attractive solution.
Somewhere around here is a thread I started for a bunch of little Nano devices to do things that nothing else exists to do for my trike, though I still haven't gotten back to that to do any of the things, you might find the ideas / processes useful.

I lost the other stuff I was typing (tab crashed, never happened like that before) but basically it will be as reliable as the hardware design/build and the software itself and it's error-detection/failsafing (meaning, how many problems that can happen you've already thought of and written routines into the software to catch).

I do recommend keeping a spare of any small light parts (like the arduino, or the controller, etc) that are critical to operation, so you can swap them out and keep going. (on my trike I have redundancies enough to keep going because I can no longer pedal myself very far at all without the motor power).



What do you think of these controllers?

- acceleratedsystems
Those are what the *Runner controllers are based on...but the Grin setup software for the *Runners is much easier to use and their support is better (and you have this forum to get help on, too, if needed).

ASI as that company is better known is not known for their individual/DIY support, though there are exceptions. The real problem is that AFAICT you have to buy from a reseller (not directly from ASI), and then you are dependent on that specific reseller for support, which includes authorizations for access to the setup software (or having to send your controller to them for setup), and if they aren't around, being unable to change things. (Mobipus is another like that). If you look around the forum for ASI you can find a fair bit about them and their stuff.
 
Last edited:
Is this also something I can try to just plug in and try?
You won't find a system that is tailored to your wishes and works out of the box. Of course there are customized systems, but they are insanely expensive, like everything on the rehab market.

Just buy this controller+display bundle:

and this throttle:
https://www.alibaba.com/product-detail/Wuxing-newest-half-twist-throttle-2_62067628818.html

The controller will run without display too, if you don't want one. It has a connector for head- and taillight and one for a breaklight. The light can be switched by the display, the brakelight by the brake levers (or by regen, if you want).

Flash the firmware, run the autodetect routine once and start testing. There's no need for tinkering with additional hardware like arduino or CA, as the firmware is open source, you have full access to any parameter. Of cource you don't need to use this special throttle, you can attach anything else.

regards
stancecoke
 
Last edited:
Back
Top