charging LiFePo4 with hub dynamo?

[jpk]

I recently built a bike light using a hub dynamo, one of Martin's circuits and three leds, mostly for commuting use. The leds have separate housings so that I can aim one close to me (with somewhat floody beam) and two other ones slighty further away. I prefer a uniformish and longish beam to one or two hotspots and some side spill.

I currently have a super cap (1F, 5.5V) in parallel with one of the leds to have light available even when I am not moving. It works fine, but not for very long, no matter how long I have been riding. It stays bright only for some seconds and then it becomes the brightness of a typical 9-led 3XAAA flashlight for a rather long time.

So, I thought that perhaps I could replace the super cap with a 16380 sized LiFePo4 battery, probably from DX. The maximum current from the hub is 400mA and is probably somewhat stable at speeds above walking speed. At 400mA the Vf for the LED (U-bin P4) should be lower than 3.6, so it shouldn't be overcharging the battery no matter how long I ride.

Would that be unsafe/unwise to do? I've understood that LiFePo4 should be safe in most uses, but would this fall outside the "most uses"?

 

[Al Combs]

:welcome:

I don't know about proper charging from the device you're using. What kind of regulator do you charge the cells with? Don't know how you're going to have it switch over to usage mode when you stop peddling. That's one good thing about super caps, they're impossible to overcharge. Perhaps more super caps in parallel?

But LiFePO4 is always safe. To your person at least, if not your wallet. No venting with flames using this chemistry. Have you considered using three small NiCad's or NiMH's? They can dissipate small overcharge rates (less than 0.1C) as heat with no problem. Any kind of Lithium cell, even the ones that don't explode, doesn't like to be overcharged.

 

[mdocod]

The only problem with the setup is that if your Vf is around 3.5V for the LED, then you're going to have the same problem that you have right now with your cap setup, it'll get dim pretty fast when you stop because LiFeP04 tends to runs around 3.2V or less under a load, so as soon as you stop riding, the output voltage would likely drop...

Ideally speaking, I'd say use a buck style regulator on the LEDs, and use 4 NiCD (maybe AA size or smaller) in series as the power source, to be charged by the hub. You could even get a multimode driver of sorts with different output levels, find a combination of modes on the LEDs that will keep the NiCDs charged up for when you stop. But have a higher output available in reserve when you want to take advantage of some stored energy- like say- when cruising down a hill real fast, you might want more light for a few minutes.

Eric

 

[jpk]

thanks for the info, I guess I'll order one and see how it works. $3 isn't too much to pay for this experiment...

I considered having several supercaps in parallel, but with the local prices ($8) any significant change in capacity would get a bit too expensive. I know they can be found much cheaper at various net stores in the US, but shipping to Europe is typically quite expensive...

Once I get some pics taken and complete the housing for the electronics I'll post some more info in the bicycle forum.

 

[BobVA]

Given that you're moving most of the time, would a diode-isolated or relay-switched primary battery pack do the job? I was thinking a few alkaline AA's might act as your "fill in" power for several months without worrying about charging?

Cheers,
Bob

 

[electromage]

Depending on the output voltage of the hub dynamo, I would say get/build a proper battery charger circuit. If you're going to use a single LiFePo4 cell, there are many solutions designed for cell phones and MP3 players that will charge a single Li-Poly cell from 5VDC (USB). If your hub puts out around 5V, I'd use one of those.

From there, you can worry about getting the battery voltage to where you want it for the LEDs by using either a boost or buck converter. I'd imagine that you're generating a lot more power than you're using, and wouldn't need to worry about your LEDs going dim. If you use a battery charger to keep the battery where it should be, and regulator for the LEDs, they shouldn't dim until the battery is nearly empty.

 

[Calina]

Wouldn't it be possible to use a resistor in series with the super cap to limit the LED output and extent the time it is lit?

Anyway isn't the output of a 9-led 3XAAA flashlight enough for a stand light? Just wondering...

 

[Calina]

Check n4zou solution to charge batteries from a dyno:
http://www.candlepowerforums.com/vb/showthread.php?t=202121

This thread might also be of interest :
http://www.candlepowerforums.com/vb/showthread.php?t=201525

 

[jpk]

Check n4zou solution to charge batteries from a dyno:
http://www.candlepowerforums.com/vb/showthread.php?t=202121

This thread might also be of interest :
http://www.candlepowerforums.com/vb/showthread.php?t=201525

thanks, I checked the links. They are not the same thing as I am intending to do, which is basically to have the small LiFePO4 instead of the supercap to have a longer lasting bright standing light. Maybe it will work, maybe not.

This project isn't so much about what is most useful or smart, but more about what is possible. I don't really need over 200 lm for my 1 mile commute on a cruiser-style bike on well lit roads... Especially by the time it is dark enough to actually need the lights for the commute I might need to have winter tyres on. I'll probably leave the bike at home at that time...

thanks for everybody's input, I'll post some information on the project in the bike forum in a few days.