efficient low-voltage ac/dc conversion for bicycle battery charger

Jim Thompson wrote in news: snipped-for-privacy@4ax.com:

Then you don't get any power generation when coasting.

Roger,

Thanks, that's what I will do.

The intended load is not a light, it is a GPS navigation system. A battery that would store enough charge to provide the required 0.2A for say 10 days x 7 hours/day would need to be 14Ah, quite heavy and bulky.

Also, adding a sufficiently large solar panel is inconvenient, considering a lot of luggage (tent, cooking equipment, clothes, ...) has to be transported as well.

greetings, Tom

Fred,

Unfortunately when the GPS is powered off it takes quite some time for it to re-acquire all the satellites, which is very inconvenient when cycling and you need to decide 'go left here or right?' preferrably without needing to stop. That's why I prefer to keep it on all the time.

(by the way the GPS would be complementary to the traditional paper maps, which offer better overview but once you get lost it's often difficult to re-establish your position)

The original plan is indeed to just use a few sets of rechargeable NiMH cells. The problem is that in a lot of cases I will not have convenient power outlets available to recharge them at night, so the dynamo solution is intended to take care of those situations.

I got a lot of useful suggestions here, which I will now try to turn into a working system (starting with simulations).

greetings, Tom

Another possibility is to use a small(5W?) autotransformer with different tabs to beef up/down the voltage of the generator to exactly the charge voltage needed. I would also put a button to disengage all load when going uphill. In that time use some LEDs to give a much lower wattage illumination.

"Tom (at tomsweb.net)" a écrit dans le message de news: snipped-for-privacy@z14g2000cwz.googlegroups.com...

change

only

Well, except that I was a bit fast and forgot the high voltage (high rpm) case.

When the peak voltage is over the battery one, then the MOSFETs intrinsic diodes make a simple bridge with the schottkies and there's no mean to control the current, except by shorting the dynamo through the 2 MOSFETs. Depending on your dynamo that _might_ be acceptable though. You could also use an intermediate "high" voltage and add a buck switcher after this. One advantage is that you'll easily find an IC that'll manage the NIMH charge... but all this is begining to be a bit too complex for the purpose, I guess.

Anyway, just for boost mode:

This depends on the A-B voltage polarity (you have to detect this). Say your dynamo position is such that A is positive wrt B.

Then Q2 is on for all this half cycle. During the same time you switch on and off Q1 so that Q1, D1, and the dynamo internal inductance and FEM make the primary side of a boost converter. You'll have to work out the equations for this.

On the other sine half cycle, Q1 is permanently on and Q2 is switched on and off.

The full bridge is getting overly complex to my taste. Gate drivers will be mandatory (level shifting) and, unlike diodes, MOSFETs switches are not autocommutating so you'll have to detect the output current polarity reversing (or equivalent) in order to switch the upper MOS off.

have

payoff.

in

is

Correction)

LT1249

the

accurate

simple.

for

extend

better

Ouch :) So many questions - serves me right for opening my mouth.

---- SWITCHIN CONVERTER ------

PFC correction is a way to prevent peaky charge current, and makes the input current look like a sinewave. It does that by running a switched mode converter at a much higher frequrency than the supply frequency and varying the switching pulse width and hence adjusting supply current throughout the low frequency cycle. The magic of the PFC circuit is that you still end up with a capacitor charged up to a DC voltage, just as you do with a simple rectifier. You were concerned about inefficiency due to peaky current, and the only way to work around that is either a choke input filter (you don't want the weight of the choke) or a PFC type circuit, or perhaps some kind of resonant filter system (you don't have a fixed frequency). So PFC techniques are of interest to you.

Let me go straight to the point. A small, modern microprocessor will do it all for you, turning on and off the MOSFET in your your PFC type switching regulator, and varying current as the instantaneous voltage varies. For all I know, it could probably turn on and off the MOSFETs which form your rectifier. As well, it can handle stuff like cutting the pedal effort at low speed and battery charge management. The micro will have an A/D converter and onboard voltage reference.

The magic of the micro controlling a switchmode is that you can tune power drawn from the source - ie you can move up and down the current vs voltage characteristic of the generator. Big generators run at high efficiency, with generator losses much less that load power. However, you may want to try loading your small generator down to the maximum output power, where generator losses equal output power, the impedance *matched* condition where you increase the load until the output voltage drops to half. Or anywhere in between.

For example, at 20Km/h, your alternator drops 9.9 - 4 volts at 0.5 amp - ie it is a source resistance of 11.8 ohms. With a 11.8 ohm load, you will get equal power delivered to generator and load. Max power in load at 20km/h will be half of ( 9.9V ^2 ) / ( 2 * 11.8 ) = 2.1 watts, which is close to your calculated 2 Watts !! So PFC lets you get the current just right for maximum power !! Of course, *you* have to pedal harder.

About your buck vs boost issue. Once you turn your inductor into a transformer with two windings, you can manage any different voltages on primary and secondary. Probably use a flyback converter. So you combine that with PFC and a single switching circuit with the one transformer does it all.

For me, all this is too much trouble for a one-off. Enough R&D for a commercial product. I would check out the simple rectifier approach. As discussed next.

---- SIMPLE RECTIFIER ------

Batteries do not like charge-discharge from cycle to cycle. While you can get away with putting the juice in as pulses, you must not put in on one part of the cycle and take out on the other part of the cycle, thus using the battery as a kind of capacitor. You end up wearing out the battery without even running it flat ! Some big wet NiCD types handle this - in relephone exchanges, but not any small types that I know of - maybe others can suggest.

About those current peaks. The source won't "go to its knees" trying to provide the peaks- if by that you mean "give up". Rather, the conduction angle will lengthen. It may seem so messy, but that is how electronic devices got their power from the mains, until about 15 years ago, when PFC came along. Inside your radio, the transformer copper resistance pretty well determines how much current flows on peaks, and in conjunction with the capacitor, how wide are the peaks. Much simpler than PFC, and it may be as far as you need to go.

If you want to check out the possibilities with a simple rectifier, I would suggest using SPICE to model possible circuits. I would model the alternator as an AC source in series with 5.9 ohms resistance. You can calculate efficiency, losses etc. This would be my first job, because if it is acceptable, I avoid a big R&D job on the fancy version.

Personally, I would put rechargable in my bike light and charge up at home.

Roger

"Tom (at tomsweb.net)" a écrit dans le message de news: snipped-for-privacy@z14g2000cwz.googlegroups.com...

would

if it

home.

Do you really need it to be on 7 hours a day?

A 10% on time will get you back to normal cells.

A GPS should not take this long, if it has been on 'recently', and has not been moved a large distance (less than perhaps 1000 miles). A typical modern 12 channel system will re-acquire in a couple of seconds. A couple of Garmin units here, and a Haicom CF module, both acquire before the display has woken up, when doing a 'warm start' like this.

Best Wishes

This actually works. At least for all the generators that I know about, the inductance of the generator is pretty high; and since the highest voltages occur at the highest rpms (and thus frequencies), the current is mostly limited by the reactance. This of course means that the rider doesn't lose the full V*I power. The OP just needs to check the state of the batteries every so often to avoid overcharging.

{I once made a simple bridge rectifier system for charging batteries via a bike, so that I could study at night for upcoming Prelims while on a bike tour. Fairly impractical in most situations, though it worked for that}.

-frank

How about:

2) Resonating the inductance of the dynamo coil with a capacitor so as to get higher voltage at lower frequencies?

Could one use a circuit that used a comparator to switch alternate halves of a MOSFET bridge (like a diode bridge, only with FETs instead)?

Just an idea to avoid the voltage drops.

Draw yourself a schematic for a full-wave voltage doubler. Now replace each of the caps with two cells. Now replace each of the diodes with a fet or scr and turn off the active deivice when the voltage or the current or the battery temp is too high or any of the above. And yes, if you use fets, you'll have to modulate them to do the synchronous rectification. Pic processors are cheap. mike mike

hmm, well avoiding mentioning ac/dc, there is a posibility to achieve more efficiency that will deal with most of your points, and maybe quite simple, that is to avoid the ineficiencies of rectification by not doing it until the voltage is steped up a bit wich is probably what you mean by (2)

basicaly the idea would be to switch the step up transformer directly acros the ac (oops i mentioned it), mosfets have inherent diodes so it will need two in series (back to back or drain to drain) so when they are both off they block the curent.

The output of the stepup device could be rectified with lower loss, (some rectification here is inevitable anyway). it would need full wave recification wich may be done with 2 windings to avoid more than one diode drop.

if the midpoint of the two mosfets is considered ground they could be driven directly by many of the available smps ics. these could provide simple output regulation by PWM. a split primary would avoid high frequency being fed to the generator wich would need decoupling to ground.

Colin =^.^=