I could use some recommendations for controllers with pedal assist.

[Jamming]

I have an ebike I put together years ago that has a failing controller.

The motor is a 1,000 watt Mac motor with 10T winding and hall sensor.
The battery is a 48v 20/amp LiFePO4.
They are both still functional and have barely 100 miles of wear and tear on them.

What i'd like to do is upgrade the controller on it with a compatible torque sensor for pedal assist.

 

[amberwolf]

There are not many available controllers not already part of a complete system that will actually read a torque sensor directly, which reduces your choices for features and functionality. IIRC some of the open-source firmwares for certain controller models and brands support some of the torque sensors, but you'd probably have to locate the files for the FW and flash them onto the controller and it's display yourself.

There are a few controllers from aliexpress that have been posted about that can use certain versions of the Erider torque sensor.

But there is an option to allow you to use whatever controller you wish, as long as it can operate from a throttle input alone: The Cycle Analyst from ebikes.ca can read many torque sensors, and be used to convert this to the throttle signal to command the controller to do whatever you have setup the CA to do based on your input.

So if you already like the system you have now, and just want to add torque sensing to it, the CA would be a "simple" way to do that, especially if you can just get another of the controllers you have now (or one functionally the same). (simple in that it just goes between your control inputs and the controller, adds a speed sensor to a wheel, and a current sensor between battery and controller, and then you set up all the menu options in the CA to cause it to respond the way you wish the bike to operate).


As an aside, what is failing on the controller?

 

[Jamming]

There are not many available controllers not already part of a complete system that will actually read a torque sensor directly, which reduces your choices for features and functionality. IIRC some of the open-source firmwares for certain controller models and brands support some of the torque sensors, but you'd probably have to locate the files for the FW and flash them onto the controller and it's display yourself.

There are a few controllers from aliexpress that have been posted about that can use certain versions of the Erider torque sensor.

But there is an option to allow you to use whatever controller you wish, as long as it can operate from a throttle input alone: The Cycle Analyst from ebikes.ca can read many torque sensors, and be used to convert this to the throttle signal to command the controller to do whatever you have setup the CA to do based on your input.

So if you already like the system you have now, and just want to add torque sensing to it, the CA would be a "simple" way to do that, especially if you can just get another of the controllers you have now (or one functionally the same). (simple in that it just goes between your control inputs and the controller, adds a speed sensor to a wheel, and a current sensor between battery and controller, and then you set up all the menu options in the CA to cause it to respond the way you wish the bike to operate).


As an aside, what is failing on the controller?

I'm not so sure I can get another controller like mine, it was custom made by Edward Lyen out of San Fransisco many years ago using an infineon controller, and I don't have contact info for him.

 

[Jamming]

As far as controllers go, since it's the cycle analyst that would bring the pedal assist function, is there any controller that would work with the Mac motor?

 

[amberwolf]

What is the problem with the present controller? (it might be fixable)



Best I can recommend for "easy" setup is to pick a displayless controller from a seller you think you can trust**** that operates completely from it's throttle input and does not require any other control input to operate, and can supply the same current at the same voltage range as your present controller does.

I expect most controllers of the correct capabilities would probably work, but whether any particular controller will work "for sure" depends partly on the speeds desired / the RPM of the motor, as the MAC has been known to sometimes be difficult to drive at faster RPMs because of the high ERPM (high pole count, gearing, etc). (The Ezee has a similar issue).

As long as the controller can work at a fast enough speed, it will run the motor fine. Unfortunately almost no controller lists it's ERPM, so there's no certain way to tell if it will work.

It seems probable that the really cheap ones won't, but I can't know for sure. Part of it depends on their MCU's speed and part of it on their powerstage design (so the gates can switch the FETs fast enough to commutate the motor fast enough).

I expect that many of the ones using Infineon / Xie-Chang MCUs, or STM32 MCUs, would work, as controllers that have worked with the MAC (including the Lyen series) have used those. IIRC LiShui and KT both use the STM32 or clones thereof, many generic controllers use Infineon or XC's.

LS and KT controllers are generally used with displays to change settings, assist levels, etc., though you can run them headless by themselves after you set them up via the display. (or you can use open-source firmware uploaded to them that gives you more direct options to use them with alternate control inputs like torque sensors, etc for some of them).

Generic controllers often are headless, not even capable of using a display. Greentime, Ecrazyman (if still around), and osme others I don't recall the company names for have been documented around the forum as decent controllers.



****for me that starts with a sales page with complete info / specs / wiring info on their stuff that you can actually read, not full of mistakes, obvious lies, copies of someone else's ad that doesnt' apply to the item for sale there, or incomplete info, like many have. If I have to ask a seller for info that should be right there, then that's a first strike, and if they don't provide that info in a complete and correct fashion, that's a last strike. Reports from others indicating specification or wiring mismatches are also a strike, and also of problems with devices that are unquestioningly not an end-user issue (which are not that uncommonly blamed on the seller or product, unfortunately).