Switching dynamo circuit

[nicknoxx]

Has anyone made a circuit that uses one or two LEDs when going slowly but automatically switches in a couple more for going faster?

It seems to me the difficulty would be that it would have to switch in extra LEDs when the current rises to a certain level but switch them out again when the voltage drops.

[nicknoxx]

Conversations on the standlight thread have answered this question today . There's a coincidence!

[cycle3]

Receintly purchased a set of LED driving lights from Ebay seller 24x7 diy. 1 was faulty out of the box (Did not work) and good luck when you contact them. Expect 24x7 diy to blame the wiring (yours). after all, It can't be 24x7 diy's issue or a failed unit you see, as they test each unit before sending - ho ho ho. Electronics that dont fail (ever). Did I tell you about Santa Clause, the three pigs and the easter bunny - they are all real you know.
Ebay 24x7 diy - No, No, No, Don't, Don't Don't. I never will again. Once bitten twice shy.

[Steve K]

Conversations on the standlight thread have answered this question today . There's a coincidence!

I'll assume that you found my project that I've semi-documented on Flickr:

http://www.flickr.com/photos/kurtsj0...7621965148305/

For the sake of the general audience, I'll just briefly say that it senses the speed of the dynamo and controls whether just two or four leds are powered.

regards,
Steve K.

[mrradlos]

Has anyone made a circuit that uses one or two LEDs when going slowly but automatically switches in a couple more for going faster?

It seems to me the difficulty would be that it would have to switch in extra LEDs when the current rises to a certain level but switch them out again when the voltage drops.

I got mine from kadikater, works really well. I think he still has some left. I'm sure he is prepeared to explain how they work, because I can't. 1 LED from 1 to 7km/h, 2 LEDs 7-15km/h, 4 LEDs (or 3 LEDs) faster than 15 km/h.

[jdp298]

This problem has been rumbling away in my head for a while, and even though Steve's ccts are excellent and worthy, I think I may have a simpler answer, and one that also prolongs your standlight, should you wish to run one.

Essentially, take the rectifier (whatever flavour) output and then run in parallel:
- zener diode
- Supercap(s) (protected by zener)
- Low Dropout Regulator (LDO) with what ever LEDs
- LM317 regulator with extra LED(s)

Having tried them in the past, LM317s won't pass anything unless the input voltage is higher than the output, whereas the LDO will regulate and then gradually decline as the input voltage decays. This means that when you're going well, everything is on, but when you're slow or stopped, the LM317 LEDs go out and the supercap only powers the LDO and its LEDs (which I would have in the next build as 'be seen by' lights more than anything else).

Perhaps not used as the maker intended, but it's the behaviour I think you're after.

[Steve K]

Just to be sure I understand the proposed idea, are all the LEDs essentially wired in parallel, but with different regulators in series with them?

If so, then you've only got 500mA to share among them, and you'll only get the light that you would get with a single LED.

[jdp298]

Completely right, and yes you only have 500mA to go round, but by having different regulators, it means the supercap doesn't have to drive all of them once you stop.

I also didn't notice (it may be, but I can't tell) that your excellent and complex cct can draw any more that 5

[Steve K]

oh.. I assumed that the proposed circuit was a response to the original question about switching between more and fewer LEDs in order to get more power at higher speeds.

Regarding the standlight, I'm not sure that having one LED wired across the supercap will draw noticeably less current than having two LEDs wired across the supercap. The Vf is still the same, but the LED's resistance will be slightly less. The LDO will have minimal voltage drop, so it doesn't regulate the current at all.

Thinking about your proposed circuit... I'm not sure that the second LED (powered by the LM317) would ever turn on. I suppose it depends on what the LDO is set to regulate at. If it is set at the voltage that is the LED's Vf at 500mA, then the dynamo voltage would never exceed that Vf plus the LDO's dropout voltage. If it was set at the LED's Vf at 250mA, then the dynamo output voltage would indeed rise enough to forward bias the second LED and the LM317. This disregards the change in Vf over temperature, of course.

My circuit will take whatever current the dynamo can supply, which will vary with speed and the forward voltage of the LEDs that are currently wired in series. Nominally, that'll be 500mA or slightly less.

[jdp298]

oh.. I assumed that the proposed circuit was a response to the original question about switching between more and fewer LEDs in order to get more power at higher speeds.

Well, it kind of is that too; or at least it's meant to work that way. If I have 2 LEDs in parallel, then I wouldn't plan to run either at more than 250mA, and that would drive the resistor values and regulator voltage levels. I'm not so sure the Vf of one regulator plus LED would clamp the supplyable voltage. I always thought a regulator would tolerate a higher input than output. Certainly it won't clamp a battery's voltage?

The LDO will have minimal voltage drop, so it doesn't regulate the current at all.

Well, yes. That's what we have a series resistor for. And the decaying voltage on the supercap, which is before both regulators, will see the LM317 cut out completely while the LDO continues to pass whatever it can. Perhaps I wasn't clear about that before...

On reflection, I'm not sure this would work brilliantly for the original question, but here's why I think it. At slow speed, the lights flicker, rather than stay dim. Without a cap, it's even worse. I get a voltage spike every time a pole crosses within the hub. Smoothing those widely spaced spikes is the job of a seriously big capacitor, certainly more than the measley 2F I operate. I reckon the voltage spikes at low speed would make both lights flicker and then by the time you're doing 6 mph or so, both would be on reliably anyway. The standlight function would work to prioritise one over the other though, I'm sure of that.