Any Analogue Computing experts got any thoughts on an analogue ebike?

Tuesday afternoon musings:

I got interested in Analogue Computing a while ago and started thinking if it would be possible to design an analogue control system for an ebike?

I’m guessing a BLDC motor would be out, but what about a VFD motor (like a washing machine motor)?

I’m thinking the problem would be to get the battery to generate a 3 phase output.  I’m not sure if a sine wave inverter could be done with analogue control.

It seems like the inputs for this control system could simply be small motors to measure wheel and crank speeds as a voltage (there is probably a better solution).  You could probably have some sort of switch/variable resistor to determine what gear you are in (if that was needed).  From there it would be a question of using these values to provide the correct frequency to the motor.

As you can probably tell I’m definitely not an expert on this but it would be cool to see if anyone on STW is.

Basically what I’m wondering is if it’s possible to make an ebike with no microcontrollers or micro-electronic components.

I’m sure it would be possible, but why bother?

Microprocessors are so cheap now that you have to have a very very good reason to not use one in more complex electronics.

I’m sure it would be possible, but why bother?

Just for stopping giving bikes stupid prefixes would be a good enough reason.

Next let’s make e-bikes that play a little tune.

I’m sure it would be possible, but why bother?

Intellectual curiosity.
Also, it would mean no one could ever refer to a bike as an analogue bike again unless it was an ebike with an analogue control system.

Depends how much control you want over it?

Simplest – reed switch for cadence, simple transistor timer circuit that cuts the power to a DC motor in the hub via a relay if it’s not reset by the switch within a second. Power manually set by the rider.

Still simple(ish) – strain gauge in the crank, amplify that signal, convert it to a PWM and via an SSR to the motor.

Less simple but still achievable – As above but generate AC to the motor by feeding a PWM into a transformer.

It’s been a while since I was at uni and did analogue electronics so don’t know the exact methods but intuition says thats all plausible.

From memory, there’s multiple ways to create an oscillator from DC voltage. There should be some fairly simple circuits using an op-amp or something using a crystal oscillator or 555 timer.

Common sensors from detecting rotational movement are usually voltage or current based which would work nicely with some op-amps to create some function used to control the motor.

But all of that requires a lot of theory, a lot of maths and a lot of data sheets etc …

All much easier and quicker and better quality done using microprocessor or Arduino or similar and using commonly available parts for those eco systems.

I used to work in control systems at UMIST. I have some experience with analogue computers but it’s at least 40 years ago and I have probably forgotten most of everything I knew. Could you start small and build a simple speed limiter for a model electric car?

Early fuel injection systems were mechanical – basically a simple mechanical analogue computer. Later they used analogue electronics, then switched to digital electronics in the 1980s. Yes, it’s possible to make analogue systems that control things like this, but everyone switched to digital systems because they have huge benefits.

It would almost certainly be tricky (but not impossible) to come up with a system that uses purely linear analog amplifiers/circuit elements but should be fairly straight forward to come up with a system that uses these in conjunction with some hardware controlled switching stages. This is probably how the first variable frequency drives would have been built as they were around long before the availability of commodity processors suitable for the task.

Early fuel injection systems were mechanical – basically a simple mechanical analogue computer.

That a lot of words to describe a carburetor.

I think Concorde was all controlled by analog computers, such as the air intakes, so controlling a motor should be easy. (I am not an electronic engineer).

That a lot of words to describe a carburetor.

A carburetor is a type of mechanical computer, but mechanical fuel injection isn’t a carburetor.

I might be missing something, but if you wanted to control mid-drive VFD induction motor, would it be as simple as taking a BLDC motor and attaching it to the crank somehow so that it gave you 3 signal outputs as you turned the pedals.
You then connect the signal wires to 6 separate transistors (with diodes for when the voltage is negative) and then connect the battery and induction motor to the transistors.
That way the speed of your pedaling and the speed of the induction field of the motor would always be matched.
Am I missing something or would that be all you need (nevermind speed cut off and controlling the level of boost for now)?

Next let’s make e-bikes that play a little tune.

Siemens make trains that play a tune when accelerating, so shouldn’t be difficult to make e-bikes do they same

Google YouTube for Siemens Taurus for the most tuneful ones

I might be missing something, but if you wanted to control mid-drive VFD induction motor, would it be as simple as taking a BLDC motor and attaching it to the crank somehow so that it gave you 3 signal outputs as you turned the pedals.

you’re effectively describing block commutation of a BLDC motor using a hall sensor reference there. Would indeed work but won’t be as smooth or efficient as a “proper” controller.

https://www.ti.com/lit/ml/slyp711/slyp711.pdf

Next let’s make e-bikes that play a little tune.

if it was analogue electronic would that tune sound warmer?