Most efficient end-to-end pedal generator?

bigchief

100 mW
Joined
Jun 20, 2009
Messages
45
I'm trying to power a PC from a bike pedal generator, and have some parts at my disposal, but I wanted to figure out the most efficient setup possible.

The PC uses a "Pico power supply", so it is powered from 12 volts DC. It consumes between 40-160 watts depending on if I am gaming with the GPU or not. The idea is I want to hook it up to a 12v nominal lithium ion battery pack and charge that pack with pedaling. If I don't exercise at my PC it'll shut off. Sounds motivating! But it needs to be efficient enough that I'm not drenched in sweat all night either; I'm shooting for cardio heart rates here :D

Stuff I have:

- Bike frame to sit on and pedal
- 32v 500w ebike hub motor
- 32v 350w "hoverboard" self-balancing scooter wheel (brushless DC motor like the ebike)
- Cogs of varying tooth count that I can weld to any of the motors above
- tons of 18650 cells... like 1kwh worth of them

Stuff I could buy:

- I scoured the internet for an MPPT, buck-boost wind turbine charge controller. I found exactly one that can take an AC (i.e. PMDC) motor, and supports custom charging profiles for Li-Ion: https://www.amazon.com/dp/B01IX6OYEA/
- If it would be more efficient than the bike or hoverboard motor, I could buy a Turnigy RC motor, since those are powerful and cheap, but I think the bike hub wheel or hoverboard wheel might be a better match for the volts/RPM range that would be practical with gearing

Challenges:

- Rectifiers top out at 81% efficiency from what I read which seems pretty poor, but MPPT charge controllers claim up to 98% (I assume when input/output are within a few volts of each other), so I'm thinking I'm going to need to go straight from motor to MPPT to squeeze the most energy into the batteries
- I need to keep the thing as simple as possible, since all conversions sap more power. Same with hooking up the chain drive... preferably it would be one chain from the crank to the motor I'm turning, with weight added to the crank to act as a flywheel
- I think the output to the PC power supply should ideally be as close to 12v as possible; I'm wondering if that would require a voltage regulator between the batteries and the power supply, or if a large enough pool of batteries, (for example 3 18650 cells at 4.0v nominal in series, 30 in parallel) would be able to stay within +/- half a volt from 12vdc if I'm keeping up with pedaling.
 
I can't really help with the hook up, but for sure your bike hub motor will make a decent, efficient, generator.

Others have produced the same thing commercially using a basic direct drive rear hub motor.

Use the power to keep a cheap lead 12v battery full.
 
bigchief said:
- I think the output to the PC power supply should ideally be as close to 12v as possible;
Is the Pico PSU a true 12VDC input, or is it an Automotive 12VDC input?

The former may mean it needs tight regulation, while the latter often means 10VDC-14/15VDC range. You'd have to check the specs for your model. They may be printed on it's label; if not the manufacturer should be able to tell you.
 
Would think besides a highly efficient motor and gear train at the rpm you will be able to drive, the controller that is able to dial in the exact load you can deal with being a major factor. Not being able to dial in (both cadence and motor) rpm and load will leave you at a compromise on the bio side as well as the motor generator side.
 
dogman dan said:
I can't really help with the hook up, but for sure your bike hub motor will make a decent, efficient, generator.

Others have produced the same thing commercially using a basic direct drive rear hub motor.

Use the power to keep a cheap lead 12v battery full.

I wouldn't want to use a lead-acid as they themselves charge and discharge inefficiently. Well-matched 18650 can charge/discharge at over 99% efficiency. The pack I pulled out of my hoverboard seems to be in perfect shape, I'd just need to rewire it down to 3-4s.

I removed one of the two hoverboard wheels as well. It feels very smooth (less drag than my ebike wheel), and hooking a multimeter to any two of the three wires then spinning it with my hand returns about 7 volts ac. Is that accurate if I'm only using the multimeter on two of the three wires?

amberwolf said:
Is the Pico PSU a true 12VDC input, or is it an Automotive 12VDC input?

The former may mean it needs tight regulation, while the latter often means 10VDC-14/15VDC range. You'd have to check the specs for your model. They may be printed on it's label; if not the manufacturer should be able to tell you.

It is the former, so I do need to pick up some tight 12v regulation. If you know of any particularly efficient modules that can do this at 30+ amps I'm all ears.

speedmd said:
Would think besides a highly efficient motor and gear train at the rpm you will be able to drive, the controller that is able to dial in the exact load you can deal with being a major factor. Not being able to dial in (both cadence and motor) rpm and load will leave you at a compromise on the bio side as well as the motor generator side.

That's a good point. I went ahead and ordered the Wind MPPT that I linked, because among other features, the manual suggests (in broken English) that it has a programmable RPM limiting feature, via electronic braking (i.e. regen braking to control the torque needed). If their support team's English is as good as their manual, I'd be better off just getting the thing off Amazon to give an honest try, and returning it if it's garbage, than trying to ask them questions.
 
By all means, use the battery you got. I just meant use something, so it can buffer out all the variations in voltage and or amps. Lead cheap if you had nothing. You can input to the battery variably, but pull smooth power from the battery as needed, up to your ability to pedal it back in.
 
One thing that most smart trainers are using that may be of interest here is the ERG mode they can run in. The wahoo kickr is one I am a bit familiar with that uses what looks to be a eddy current brake to dial in accurate power loads. You or your program can tell it to load the drum to say 150 watts and you can spin at 60 or 100 rpm and still only push 150 watts. Takes a second or two to stabilize the load if your not holding a relatively steady cadence but settle in nicely once you smooth it out. Would be interesting to use this type of loading for generating as you can keep a relatively constant output that would be somewhat independent from the rpm. You could pedal slow and grind out some watts, or spin lightly depending on how your feeling or how long you want to generate or just vary it a bit as you warm up or fade out.
 
Back
Top