Squeezing the power out of a Supercap

[steveo_mcg]

Mine flicker terribly at low speeds. How ever at filtering speed i'm not convinced this is a bad thing.

[Bandgap]

Personal opinion :-)

For me, it is the thought that the driver will see me, but be so distracted by my flashing light that they hit some other poor soul.

A focussed XP-G running at 500mA is as bright as a camera flash, and that flashing at 5Hz is almst impossible to ignore.

Steve

[ianfield]

5Hz puts you in the realm of the so called "brain wave" slow rhythm frequencies -fortunately 5Hz is somewhat removed from the alleged "brown frequency".

A search on TENS pain relief should turn up at least some info on this topic.

Personal opinion :-)

For me, it is the thought that the driver will see me, but be so distracted by my flashing light that they hit some other poor soul.

A focussed XP-G running at 500mA is as bright as a camera flash, and that flashing at 5Hz is almst impossible to ignore.

Steve

[Bobblehat]

So ..... the super caps perform no smoothing in this circuit? Or is that just the non-standlight LED has no smoothing?

[ianfield]

Wouldn't know - I only commented on what someone else said.

So ..... the super caps perform no smoothing in this circuit? Or is that just the non-standlight LED has no smoothing?

[jdp298]

C2 will provide some smoothing, C1 hardly any.

It would not be hard to put a smoothing cap across +Vcc and Gnd, that would ameliorate flicker but won't cut it out.

For what it's worth, I still counsel against having all 3 LEDs in series, this is what will eat up the energy for the Red. When stopped, C2 effectively tries to power the whole circuit, whereas C1 can only power the last white XP-G.

[Bandgap]

C2 will provide some smoothing, C1 hardly any.

It would not be hard to put a smoothing cap across +Vcc and Gnd, that would ameliorate flicker but won't cut it out.

I agree. You would need a big cap across the whole thing. I don't think it is worth it.
As it is, the LEDS will still flash badly because even the caps are only supplying a few dozen mA between the rectifier pulses.
De-flashing and standlights are two related, but different things - a bit outside the scope of this thread I feel. - my fault for mentioning it. My option is to use a push button to force the system into standlight mode briefly when cycling through pedestrians or close up behind cars.

For what it's worth, I still counsel against having all 3 LEDs in series, this is what will eat up the energy for the Red. When stopped, C2 effectively tries to power the whole circuit, whereas C1 can only power the last white XP-G.

I do have a circuit where one capacitor powers the red led and the other cap powers one of the white leds.

I PMed it to BobbleHat for his thoughts rather than wastre every ones time.

If anyone else is interested, I will attempt to post it.

Steve

[Savvas]

As it is, the LEDS will still flash badly because even the caps are only supplying a few dozen mA between the rectifier pulses.
..................
I do have a circuit where one capacitor powers the red led and the other cap powers one of the white leds.
................If anyone else is interested, I will attempt to post it.

Steve

I have had sucess with a 4700uf smoothing cap across the rectifier output using 2 X leds.

Circuit? Yes please!

Savvas..

[ianfield]

I have had sucess with a 4700uf smoothing cap across the rectifier output using 2 X leds.

Circuit? Yes please!

Savvas..

A small SLA battery gives much better smoothing, it also provides a loose but significant regulation of the generator and keeps the lights going a lot longer than a supercap if you get caught up in gridlocked traffic.

With a hub generator you can leave an SLA charging whenever you're moving so the lights always have plenty of running time.

[Bobblehat]

Hi Folks

This is the circuit that Bandgap mentioned above, redrawn by me from his sketch (any errors will be mine ). I've spun the diagram around to give a more conventional +Vcc at the top, instead of sidey-ways-on





The usual rider applies of adjusting D1's and D2's to keep the supercap volts under the max rating (5.5V for these caps).

Bandgap also notes that ...... "this arragement has to be used with a bridge rectifier that blocks reverse flow. - because if Vcc is shorted to ground, up to minus 11V could be imposed across the middle white LED, which can usually only stand minus 5V. "

[Bandgap]

Thanks for drawing that so neatly and posting it BobbleHat.

I have had sucess with a 4700uf smoothing cap across the rectifier output using 2 X leds.
Savvas..

Interesting, was it a hub dynamo?

and was the flashing negligible?

Steve

[Savvas]

Interesting, was it a hub dynamo? and was the flashing negligible?Steve

Hi Steve,

Yes to both questions. Beyond dawdling pace, no flashing is noticeable. I used a largish (63V I think) cap.

Savvas

[Savvas]

The usual rider applies of adjusting D1's and D2's to keep the supercap volts under the max rating (5.5V for these caps).

So once the circuit is built up, what's the best way to measure potential across each of these individual leds (to get the diode-drop string right)? Is is simply a matter of bridging them with a voltmeter and spinning the test rig?

Savvas

[Savvas]

A small SLA battery gives much better smoothing, it also provides a loose but significant regulation of the generator and keeps the lights going a lot longer than a supercap if you get caught up in gridlocked traffic.

With a hub generator you can leave an SLA charging whenever you're moving so the lights always have plenty of running time.

Thanks Ianfield,

Have you tried this? Where would one get an SLA small enough to use for this purpose? The smallest I've seen are 6V ones about the size of a pack of cigarettes. Is that the size you are thinking of?

Savvas

[Bobblehat]

So once the circuit is built up, what's the best way to measure potential across each of these individual leds (to get the diode-drop string right)? Is is simply a matter of bridging them with a voltmeter and spinning the test rig?

Savvas

That's what I was planning ....... then err on the side of caution with the diodes, using combinations of 1n400x and schottky to keep the cap volts below 5.5V.

Been thinking about where to scavenge a motor for a test rig ....... maybe one that can be easily adjusted in speed with a simple triac/diac box I have gathering dust somewhere. Not that keen on sacrificing my best mains hand drills!

[Bandgap]

So once the circuit is built up, what's the best way to measure potential across each of these individual leds (to get the diode-drop string right)? Is is simply a matter of bridging them with a voltmeter and spinning the test rig?
Savvas

Put the DVM across the supercap and cycle - if it gets to above 5.5V, add another diode in its charging string (eg, D1s for C1).
Repeat for the other supercap.

Diodes add about 0.6V drop. For 'half a diode drop' use a Shottky diode instead (around 0.4V).

Then cycle hard and long to make sure there is no time 5.5V is exceeded.
Use 1A diodes as they are easy to find - even though 0.5A would be enough.

For reliability - DON'T go above 5.5V.
If you only get 5.0V, then so be it if taking a diode out gives over 5.5V

EDIT - sorry BobbleHat, just read your entry and realised you covered most of this

Steve

[ejoyu]

Hi guys!

Some more good stuff going on in this thread. I just want to echo SimonInd's comments: my supercaps are still going strong. The things to note:

• My commute is much shorter than SimonInd's - around 4 miles one-way, and I don't use the lights in the morning.
• I would guess they've only been in service for around 30 hours in total - much less than SimonInd's
• I don't let my caps discharge - I unplug the lights when I get home.

This is all still using my original circuit with a bottle-top dynamo. Quite a loud little bugger, but I've found this to be an advantage as pedestrians now hear me when they decide to appear in-between parked cars...!

I have a few of the 4F caps left over - if anyone wants some, just let me know (I'm in the UK).

Cheers!

[steveo_mcg]

Ejoyu, if you've got two going spare I can paypal you some cash. I've got caught out by chineese new year and I've got a conversion and a build for a budy going.

[Bobblehat]

ejoyu ....... you got PM

[Bobblehat]

Hey ...... ejoyu ......... have you still got a couple of 4F's going spare? Did you ever get my PM?

[ejoyu]

Hey ...... ejoyu ......... have you still got a couple of 4F's going spare? Did you ever get my PM?

Yes and yes - I thought I replied but perhaps I didn't! In any case, I've pm'd back now!

[Bobblehat]

Yes and yes - I thought I replied but perhaps I didn't! In any case, I've pm'd back now!

Thanks ejoyu ..... just got this last PM Ok ........ thanks for reserving a couple ...... you're a gent!


Reply on the way soon.

[Savvas]

Hi Folks This is the circuit that Bandgap mentioned above, redrawn by me from his sketch (any errors will be mine ).

Is anyone actually using this circuit? I'm curious as to how it is performing?

Specifically:
- what's are the F and R standlights like?
- does the headlight suffer at all?
- how many diodes were needed to get the cap voltages sorted out?

Thanks,

Savvas.

[Bobblehat]

Is anyone actually using this circuit? I'm curious as to how it is performing?

............... Savvas.

I'm gathering parts now and will give a test report when built! It may take a little while to get them all. I decided on a complete new build of that circuit rather than taking my twin XP-G apart.

I've an updated diagram with Bandgap's rear light safety diodes, some smoothing caps and the bridge rectifier drawn in, but I'm hesitating on posting it until tested ........ this thread might get a bit diagram heavy ..... but if anyone insists!

[Savvas]

..... but if anyone insists!

I'm an insistent type (so I'm told) - I'm looking to use this circuit (if it works) with a new light for winter in mind!

Incidentally, over at mtbr Rob Dean has been discussing what appears to be a new version of Kerry's triple dynamo standlight headlight with some new features - 2 of the leds act as standlight and they also provide a low speed trail light as well. I am flummoxed as to how he has fit it all into the tiny Easy2LED housing he has used!

Also, I thought I'd mention as a reminder that there have been a few dynamo lighting and standlight threads here at CPF. IMHO they all contain valuable info. I wonder of there's any way of grouping them all together in some sort of index for those new to these conversations?

Savvas

[Bobblehat]

I'm an insistent type (so I'm told) - Savvas

Untested cicuit .....



Remember that the series diode strings D1's, D2's, D3's do not necessarily represent the actual number of diodes needed, but show that you need to adjust the number, and type, of diodes to achieve the desired Vf (as per the notes in the diagram).

[steveo_mcg]

ejoyu your package arrived the other day, thank you muchly.

I'm using it with a single emitter with the 4f supercap so the Vmax is never higher than the LED Vf, does any one use a resistor in this configuration? I'm guessing the fairly high internal resistance of the cap is enough to stop it dumping its energy straight into the emitter.

[Bobblehat]

I'm using it with a single emitter with the 4f supercap so the Vmax is never higher than the LED Vf, does any one use a resistor in this configuration? I'm guessing the fairly high internal resistance of the cap is enough to stop it dumping its energy straight into the emitter.

steveo, you can put the supercap in parallel with the LED, but this is not using the supercap to its full potential ..... it's best explained by the man who showed us the way in this thread

http://www.candlepowerforums.com/vb/...t-work-)/page2 msg #52

The resistor is used to give a longer "on" time (at the expense of brightness) for the standlight. Two emitters and a 5.5V supercap powering just one emitter as the standlight seems a very good match.

[steveo_mcg]

Problem is i've only got one emitter so the circuit Vf is only ever ~3.2v. Its a conversion of an old bulb lamp for a mate so getting clever with vregs etc isn't worth it and its only got room for one LED as i'm keeping the old reflector. It gives virtually no hot spot but as a get you seen lamp is should be pretty good.

[steveo_mcg]

I'm getting slightly unexpected results... I'm running the lamp + supercap from a battery and regulator to test things and while connected to the batter the circuit sits about 3.2v once the power is disconnected the voltage drops to 2.7v, is this normal or is the supercap just taking a very long time to charge?