alexlockhart

I posted this over at candlepowerforums the other day, and realized I should post it here as well. I've been reading this forum occasionally for several months, and profiting from the helpful, knowledgeable folks here. If you're interested in this, and want to reply, I suggest reading the thread at cpf to get updated, since I'm keeping up with that one regularly, but I'll keep up here as well, so you can post here if you don't want to register at cpf.

A frequent frustration for me when on long bicycle tours is the difficulty of keeping my batteries charged for my bike lights, cellphone, camera, laptop, and GPS. Now that I've found that hub dynamos can produce far more than the 6V/3W they're rated for, and now that I have money to invest in an electrics system for my touring bike, I've decided to build a system that does what a car electrics system does: generate electricity from motion, store that electricity in batteries, and use it for lights and other things. I drew up a quick-n-dirty block diagram in MS Paint to help me understand how it will all work, and thought I should include it here. It's at the end, but you may want to skip down and reference it as you read my long and complex explanation of the system.

The power in the system will be generated by a Schmidt SON hub dynamo. All of its power will go to charge a Li-Ion battery pack. The pack will have twelve 18650 cells (wired in 6S2P, output voltage range of 18-25.2V), since that's as many as I can fit in the water bottle they'll be housed in. I plan to charge them by sending the varying AC output of the dynamo to an automatic switching voltage doubler/fullwave rectifier circuit to give me at least 15V at low and high speed, and feeding that into a switch-mode buck regulator which will output 12V. That regulated 12V will power two battery chargers designed to charge a 3S Li-Ion pack from 12V, with their inputs (from the regulator) wired in parallel, and their outputs (to the battery pack) wired in series, so that their total output voltage to the 6S pack will be the 25.2V it needs.

My front light will be 4 Cree P4s controlled by a bFlex, hooked straight to the battery pack. I'll also use a switch-mode buck regulator to output 12V from the batteries, and use that to run 12V car battery chargers for my cellphone, camera, laptop, and GPS batteries.

I've already bought the parts for my front light, but they're not all delivered yet, so I'll assemble that when they come. I also have 4 adjustable switch-mode buck regulators rated for 60V input, which I mentioned above, and my Schmidt SON is on order. I haven't yet bought the Li-Ion batteries or chargers, and my design for that system is not yet finalized, so I'm looking for input.

I have some design constraints. Using 4 Crees for the front light requires a minimum 17V input to the bFlex, so I have to wire my batteries in at least 6S so that I can deliver that voltage when they are almost drained, at 3V/cell. That high battery voltage also means my buck regulator can always deliver 12V to my various other battery chargers. But I can't find Li-Ion battery chargers that can take 12V and charge a 6S pack, so that's why I'm planning to use two chargers designed to charge 3S packs from 12V, and wire their outputs in series.

Almost as an aside, my laptop takes 19.5V, so I will probably set one of my buck regulators to output 19.5V from my Li-Ion battery pack and plug that straight into the laptop, instead of buying a step-up converter designed to power my laptop in a car, since that would be unnecessarily stepping down to 12V and then back up to 19.5V. Also, the laptop power pack is rated to deliver 3.34A at 19.5V when plugged into AC wall power, so I could use one of my switch-mode buck regulators to step that down to the 12V that my Li-Ion battery chargers need, thus enabling me to charge my Li-Ion pack from AC wall power.

One potential problem I see is wiring the output of the Li-Ion chargers in series to charge a 6S pack with two 3S chargers. Does anyone have experience with this? Will the charge regulation be affected?

As I've said, I'm asking for your input, both problems you see and ideas you have. I don't have any experience working with dynamos or building lights, and have only limited experience designing and building electric circuits, but I've been reading and understanding a lot (mostly from the excellent posts here), and I think I have a working design. The details of the light design and building deserve their own thread, and I'll probably need to start a new thread for adapting Martin's voltage doubler/fullwave rectifier circuit to power battery chargers through a regulator, too. Right now I'm just trying to get this electrical system design out of my head and into the real world, and I need your help.

Alex

DCCommuter

First, I applaud your ingenuity.

However, I think you are going to have problems with the charging part of the circuit. The reason that Li-Ion batteries need "smart" chargers is that it is difficult to tell if they are fully charged. Old fashioned "dumb" batteries like lead-acid and NiCad have very simple behavior. As you charge them, their voltage increases. When they hit the maximum they are fully charged and you can stop. Li-Ion and NiMH have a different behavior. If you apply a constant current, the voltage will climb, then peak and decline. The fully charged voltage is below the peak voltage, and the battery will hit the fully charged voltage twice during the charge cycle. In order for a charger to charge a battery fully, it has to track not just the voltage, but also the voltage history, so it knows when the battery hits the target voltage for the second time. That's what makes a smart charger smart. Running from a generator, your charger won't be able to track the voltage history, because there isn't a reliable source of input power.

Similarly, I think you will have problems running chargers in serial unless you can isolate them from each other. I'm also concerned that the simple smart chargers you are using won't be able to handle a battery that is being used while it is being charged.

I think you might be better off using "dumb" batteries, because it is so much easier to measure depth of charge. Then you cut also implement something where the generator is switched off when bike speed drops below perhaps 10 mph (as measured by generator voltage) unless the batteries are at 50% or less. It would save your legs on the uphills.

Of course I never understood why they don't just make voltage-regulated generators for bikes. That would simplify things greatly.

StephenH

It seems to me like you might be expecting to get more power out of that system than what you're willing to put in. Have you actually checked total output energy you can reasonably generate in a day, versus what the different gadgets will consume?

For example, that 3.34A @ 19.5V is 65 watts- versus a 3 watt generator.

cyccommute 

That's not the reason that Li chemistry and NiMH batteries have smart chargers. NiMH batteries are sensitive to heat. Too much heat can kill the battery. That's why NIMH chargers need to measure the change in voltage and the change in temperature. If you generate enough heat you can even vent the cell.

Li chemistry limits voltage...and temperature...for different reasons. First they have a very low tolerance of overcharge. If you do happen to overcharge them, you can form lithium metal at the negative electrode which can react violently with other materials in the cell..."resulting in a violent reaction called ‘venting with flame’" The material used for the negative electrode can be changed to a saver material at the cost of battery lifetime. I suspect that most of the batteries made now use the safer electrode, however you still have to control for voltage to keep from killing the battery. Each cell also has a protection circuit to further protect the battery from overcharge, overdischarge (which can lead to more plating), temperature and limits the rate of discharge.

Additionally, charging an Li battery at low temperature (0 C) can lead to plating of the Li on the anode...again not a good thing! For summer this isn't an issue but for winter riding, you could end up in such a situation, especially on an uninsulated bike.

alexlockhart

The battery behavior you're describing matches everything I've read about NiMHs, but is not true for Li-Ions. Ni-MH is a similar chemistry to Ni-Cd, and Li-Ion is very different from either of them. Also, you state that depth of charge (I've heard it called state of charge) is much easier to measure with "old dumb" batteries like Ni-Cd and lead-acid, but that also is contrary to everything I've read. Li-Ions charge and discharge with a somewhat linear voltage curve, so their state of charge is easily measured by their voltage, whereas nickel-based and lead-based chemistries have a much flatter voltage curve through charging and discharging, making them much harder to measure. I've seen this stated in numerous places, notably Battery University, which has excellent explanations of battery behavior.

I've had others explain why running chargers in serial will not work, so I've abandoned that idea, and am working on some other ideas for how to make it all work. I'll post an updated design later, with notes from others in other forums. And I've experimented with running LEDs from the outputs of a battery charger while it's charging Li-Ions, and it seemed to work fine, but those were not long or exhaustive experiments.

I think if I wanted to save my legs on the uphills, I wouldn't be planning to tour fully loaded through the mountains of Mexico this winter. I'm not nearly as concerned about weight and resistance as most cyclists, partly because I'm young and strong, partly because I have really low gears and don't mind going slow. Still, a manual switch to disengage the generator might be useful for saving my legs on long steep uphills.

Yes, I do have a very good idea of how many watt-hours I'll be able to generate in a day with the Schmidt SON, and how much power will be consumed by my gadgets. The dynamo is rated for 3W output at 8km/h, but since I'll be averaging about 24km/h, I'll be able to produce an average of 9W, assuming I load it properly (which is a big part of my design). So that's 9Wh of power for every hour I'm on the bike, and I'll be riding for 6-8 hours a day, so assuming 20% loss for all the electronics, that's 43-58Wh of useful power in a day. Yes, my laptop power brick can output 65W, but my laptop actually consumes about 15W when running on batteries, and can draw up to 20W to charge the batteries, so if I use the laptop for an hour a day (to process my photos and update my journal every night) that's about 20 Wh per day used by my laptop. My cellphone battery takes 4Wh, my camera battery takes 8Wh, my GPS battery takes 6Wh, and those only need charging every 4-7 days. In addition, I'll have about 100Wh of "reserve" in my big Li-Ion pack of twelve 18650 cells, which can be charged from AC power when it's available, so I think I'll have plenty of power.

Alex

icedmocha 

Why not use a motorcycle battery? Everything you are doing is interesting, though I don't know much about batteries. For your cell phone, camera, and laptop, would a Solio charger work?

alexlockhart

A motorcycle battery is a lead-acid chemistry. These are cheap, but have terrible energy density in comparison to Ni-MH or Li-Ion. Lead acid batteries have 30-40Wh/Kg, Ni-MH have 30-80Wh/Kg, and Li-Ion have about 160Wh/Kg. So for storing electrical energy on a bicycle where weight is a concern, Li-Ion is by far the best choice. It's significantly more expensive, but I have money designated for this project.

Solar panels are ohmic devices like dynamos, meaning they require some dedicated circuitry to get maximum power from them to do "normal" things like charging batteries. For the same power generation capability, they are heavier than a hub dynamo and more fragile, and their power generation is at the mercy of the weather, whereas a hub dynamo is always on. I did consider using solar panels, and this one is what I'd pick, but I think a hub dynamo is a better power source.

Alex

ModoVincere

Seems like an overly complex solution. You could run an Inoled 20 light or two Schmidt E6 lights off the SON, probably install a bridge rectifier and some NiCad or LI AA batteries in the system so the lights don't go out when you are at a stop sign, all for 1/2 the weight and a lot less complexity.

alexlockhart

Yes, but:
The complexity is not for the lighting system, as there are far simpler solutions if that's all I want to do. Besides, the Inoled and E6 lights are less efficient (incandescent), less bright, and offer a less useful beam pattern than my Cree LED light system. The complexity of the system is necessary to be able to charge my various batteries while out on a long tour, which is the primary reason for its existence. I do have some better ideas from some discussions in the other 2 forums where this is posted which I think will result in a less-complex design, but it will still be necessarily more complex than a bridge rectifier and battery inline with the lights.

Alex

alexlockhart

OK, I've had a lot of good responses from people on the 3 forums where I posted this, and have changed my design ideas somewhat, so I'll try to synthesize the good suggestions and update my design here for everyone.

The idea of wiring my Crees in 2S2P had occurred to me, but I had dismissed it since then they would each see half of the current the driver gives, meaning my maximum drive current would be 500mA per LED. However, I knew I wasn't going to need or want them to be run at 1000mA each, and figured that I'd experiment with setting the max level at 500mA or 750mA and see how I liked it. 500mA per Cree would mean about 120 lumens each (before optics losses). As a comparison, I currently have a DiNotte Ultra5, which produces 120 bulb lumens with a 20 degree collimator, and a Seoul flashlight as a helmet light whose maximum drive level is about 500mA, meaning it's also producing around 120 bulb lumens, but it's fairly tightly focussed. Imagining the amount of light from 4 Crees producing 120 lumens each, with the beam pattern I'll have, I think that will be more than enough light for a screaming descent. So now I plan to run the Crees in 2S2P, as suggested, and power them with a battery pack of 3 cells in series (3S4P for 12 cells). Several people said that wiring the two 3-cell chargers in series for the 6-series setup would not work due to the chargers being unable to sense the voltage regulation, so that idea was abandoned, which is just as well, since I've also been pointed to the safety concerns of overcharging a cell in a 6-series pack due to imbalances. So using a 3S4P configuration turns out to have many benefits. Another advantage of that setup is that my nominal battery pack voltage will be 11.1V, so I won't need any regulation between the batteries and my various battery chargers designed to run on 12V.

I was so excited to find this that I spent the next several hours reading the webpage and many of the posts on the Yahoo discussion group for it, repeatedly exclaiming "This is SO COOL!!"

The older versions could only charge lead acid and nickel batteries (using a pulse-charging method, very efficient), but the new one (just now in production) has an end of charge sensor which can measure the instantaneous pack voltage in between charge pulses, so it can charge Li-Ions safely with a user-settable cutoff voltage. It also can charge batteries which are being used simultaneously, which is a necessary part of my design that I was unsure about with the other cheap charger I was going to use. I need to read more about this and get info from Jean-Michel (the inventor) before planning for exactly how to use this, but it looks like exactly what it says: the missing link.

One of its best features for making use of dynamo power is its maximum power point tracking (MPPT), which adjusts the load given to an ohmic device to continuously find the sweet spot for power production. This will ensure that the dynamo produces at least 3W at 8km/h (its design spec) and at least 9W at 24km/h (my average touring speed), meaning I'll have 43-58Wh of useful power after 6-8 hours on the bike (assuming 20% loss in the electronics).

Yes, it does mention 20V max, and 8W max power processed before it starts limiting the supply (from the dynamo). I'll have to find out if that means I have to protect it from higher voltages (with a shunt regulator, as you suggest) and if I may want to wire 2 in parallel (which is possible) to make use of the available power all the way up to 26mph.

So, assuming the TuneCharger is what I think, here's my updated, simplified, safer, design: Dynamo connected to input of TuneCharger, which is connected to 3S4P battery pack (9V-12.6V range), which is connected to the bFlex, which drives my 4 Cree headlight in 2S2P, and also acts as a 12VDC source for connecting chargers for various batteries, including my laptop. A bonus is that when I have AC power, I expect I can plug my laptop power pack into the TuneCharger, and it will charge the batteries using that input just like it would with the dynamo input, only much faster since it has 65W to work with.

My next step, and perhaps the only remaining questions, is to find out from Jean-Michel if the TuneCharger will do what I want, and what I'll need to consider to make it work. Anyone see any problems, or have any concerns? I won't be able to start building and experimenting until everything arrives, so I want to be ready to go with a solid design, to minimize my time and expense in experimentation and making mistakes. So far I've had excellent responses and suggestions, thanks to all!

Alex

Frankinbiker

I too have been searching for a solution like this. The tune charger seems to be a perfect off the shelf solution. I am hoping to power two 5LED front headlights and a 10-15 LED taillight while having excess power go to batteries in an efficient manner for times when the bicycle is not moving. This seems better that other systems that try to switch between power from the dynamo and the batteries.

I am thinking about using the internal electronics and batteries of an emergency light hand crank dynamo but do not know if it will work.