Squeezing the power out of a Supercap

[SimonInd]

It's entirely possible my circuit isn't accurate, I built it with the idea in my head and then fiddled a bit with it on veroboard before it worked so the diagram is from memory.

ejoyu your circuit is nice, shifting D4 maybe what my 'fiddling' resulted in, but it certainly looks cleaner than mine.

Looking at the shunting problem you can probably increase R2 as the red LED will require less current than the XM-L and it's forward voltage should be lower (red LEDs normally are although your measurements don't back this up). The other possibility is to switch LEDs so its the rear that's less bright. I get enough comments about how bright my back light it so it can probably be reduced a bit.

I agree with your calcs, I don't see how the Caps can be overcharged as the maximum voltage accross them can only be Vf for the 2 LEDs and 1 1N5818 minus Vf for the 1N4001. That's the beauty of 5.5v caps as for this application.

To jdp's point about seperate rectifiers, I agree it's less confusing but are you getting as much power out of the dynamo? My starting point was two LEDs in series that could take > 500mA. If the dynamo can be considered a constant current source then keeping everything in series gets the most out of it unless I'm missing something?

Last question, what did you use to draw the circuit? much nicer than mine...

Cheers
Simon

[minisystem]

Simon and ejoyu, skimming the thread I didn't see any mention of capacitor charging time in your circuits. If your caps are depleted, is there much of a delay before the LEDs light up when you start riding?

[ejoyu]

ejoyu your circuit is nice, shifting D4 maybe what my 'fiddling' resulted in, but it certainly looks cleaner than mine.

Well, it's your circuit really, just with D4 changed. So, we're both as clean!

Looking at the shunting problem you can probably increase R2 as the red LED will require less current than the XM-L and it's forward voltage should be lower (red LEDs normally are although your measurements don't back this up). The other possibility is to switch LEDs so its the rear that's less bright. I get enough comments about how bright my back light it so it can probably be reduced a bit.

I've already swapped the front and rear LED's around, and I'll have a go at increasing R2 and see what happens. I'm gonna guess that I probably won't notice any dip in brightness of the rear LED, which is good

The actual rear LED is from here - it's a fairly old LED, hence the greater voltage than an XPE (I presume). But, it only cost me £1.10 delivered (I was buying something else off the same guy), so it was a bit of a bargain!

I agree with your calcs, I don't see how the Caps can be overcharged as the maximum voltage accross them can only be Vf for the 2 LEDs and 1 1N5818 minus Vf for the 1N4001. That's the beauty of 5.5v caps as for this application.

Ah, I forgot about the Vf of the Schottky diode - I'll add that into further calcs!

Last question, what did you use to draw the circuit? much nicer than mine...

Cheers! I'm using Ubuntu, and in the package manager there's something called gEDA Schematic Editor - that's what I used!

Simon and ejoyu, skimming the thread I didn't see any mention of capacitor charging time in your circuits. If your caps are depleted, is there much of a delay before the LEDs light up when you start riding?

Approx 2 seconds, if that - I haven't really noticed. Plus, when I've finished cycling I unplug the lights, so the caps tend to stay charged at about 4.5v until the next time I use them. If I left the lights connected, they remain lit for minutes (over 3 minutes at least - that's when I got bored of waiting!).

[SimonInd]

Re: charging time, I let my caps discharge as I have no on/off switch, however when I start off I get light straight away. I think this is due to the capacitors ESR preventing it from taking all of the current from the dynamo so some is available for the LEDs. I have to cycle for a few minutes to get the necessary charge for the standlight which seems to back this theory.

[SimonInd]

I'm running ejoyu's circuit now and it's working fine, with R2 doubled to 100 ohms. However to solve the shunting issue I'm thinking that replacing D2 with a transistor, emitter to ground, base and collector connected to C2/R2 might work. As the voltage accross C2 increases Vbe drops until the transistor shuts off preventing shunting. It might need D2 to remain between the emitter and ground to protect C2 from over-voltage.

Any ideas if this would work? I'm at work so can't try it until at lease the weekend but it sort of makes sense on paper...

Cheers
Simon

[Steve K]

hi Simon,

Any chance you could sketch out the proposed circuit?

Going back to Ejoyu's circuit, I'm not certain that the 50 ohm resistors would really be diverting that much current from the LEDs. If the 50 ohm resistor had 3 volts across it, the current would be 60mA. Not a lot, compared to the dynamo's 500mA, but certainly a loss. With a 100 ohm resistor, the loss would be closer to 30mA, and even less of a concern.

has this been a real problem, or will this just be a way to optimize the circuit?

regards,
Steve K.

[SimonInd]

Here's what I propose, edited with MSPaint as I'm at work:



Arguably it's just optimizing, the circuit works very nicely and my next step is to improve the optics as the LED is taped where a bulb would go in a retro chrome front light and this is too far back - new reflector and lenses on their way from China.

I think the sensible next step is to measure the current through the LEDs to verify that none/little is being wasted once the caps are charged. Depending on findings the possibility of adding the transistor can be explored.

[Steve K]

I'm having trouble seeing the schematic. Do you have a link to it?

Steve K.

[SimonInd]

http://s9.postimage.org/on7xw2zb3/circuit.jpg

[Steve K]

thanks for the link.

I've seen people use BJT's this way, but don't recall the specific reason. Maybe to match the Vbe in a BJT used elsewhere in a circuit?

In this application, I think it'll just act like a silicon diode. The voltage across it will be a function of the current through it, but will behave like a diode. How does it work when you wire it up like this? Do you see a difference, and if so, can you describe how it differs from when it uses a plain diode?

Offhand, I would assume that when the dynamo first starts and the supercaps are discharged, the supercaps would absorb most of the current until fully charged.

Steve K.

[SimonInd]

I'm not going to get to wire it up until at least the weekend, I was just trying to think of a way to kill the flow through R2 once the capacitor is charged. Thinking about it you're probably right, when C2 is charged and the transistor shuts off the capacitor will start to discharge through R2 until it opens again so in practice it will act like a diode.

Oh well, back to the drawing board.

[Steve K]

I would have to sit down with pencil and paper in order to figure out a good way to avoid that bit of current flowing through the resistors without making it too complex. There are always compromises in engineering... trying to optimize the function without impacting other qualities too much.

I threw together a little taillight with standlight recently.
http://www.flickr.com/photos/kurtsj0...7617009273346/
It uses a charge pump to increase the supercap's voltage and thereby increasing the standlight runtime. I've been thinking about adding some circuitry to shut off the standlight while the bike is moving, but didn't come up with anything that was suitably simple. No plans to do anything other than just breadboard it and play with it, so it's not a real concern, but it is something to ponder.....

Steve K.

[Savvas]

Hi Steve,
These are all 'standard' 5 mm leds - right? What headlight would you run with this charge pump arrangement?
Savvas

[Steve K]

hi Sam,

I'm using this type of taillight (but not standlight) with my Cree headlight. The headlight just rectifies the AC dynamo power and drives a few XR-E's. This means that the taillight gets the full 0.5A from the dynamo. With some of the other dynamo headlights that have switching regulators, I'm not sure what the dynamo current ends up being, so I don't know if the taillight would be running at lower current.

Steve K.

[Savvas]

[QUOTE=SimonInd; Oh well, back to the drawing board.[/QUOTE]

Hi Simon,

Anything further to report on your circuit? Could you perhaps just substitute a diode where you have the transistor?

I've managed to get hold of some of the 4F caps from the same supplier you obtained them from. Can you tell me please which is the negative terminal? Is it the one that comes off the 'top' of the button (the one with the little white arrow)?

Thanks,

Sam.

[SimonInd]

No changes on the circuit, if you read back up the thread there is a diode where I was proposing a transistor. Recently my 5cm reflector and glass arrived from dealextreme so I've been fitting that into my light. Just trying to decide between a flat thin glass or a heavily domed one which concentrates the beam at the expense of spread.

Having ridden with this circuit (with the 50 to 100 ohm resistor change) for over a month now I must admit I'm loathe to change it. On the move my front light is comparable to the CREE XM-L based torch I was previously running from a 18650 battery (admittedly on mid rather than full brightness) and my rear light is brighter than anything else I've seen. The standlight is good for over 5 minutes and comparable to the lesser bicycle lights out there. It seems to charge from nothing in a couple of minutes riding, with no noticable drop in light brightness (although there must be some).

The circuit is squirreled away inside one of these http://www.sjscycles.co.uk/dynamo-re...amp-prod23694/ so I just have a two core wire in from the hub and another along to the rear light. I'll probably add a switch for the summer but this time of year having lights always on is fine.

From memory the -ve terminal is the one off the top with the white arrow, good luck with construction it's a great little circuit. Use Schottky diodes for the rectifier to minimise any losses there.

Cheers
Simon

[Bobblehat]

Hi Simon

When you say no change ..... are you still using the red led near C2 and 100 ohm and the white led near D3 & D4 (in ejoyu's schematic)? I read somewhere above that there were some experimental swapping round of the two leds.

How easy would this circuit be to adapt it to a double white led front light and still have a rear light? The reason I ask is that I have a nice 2 x XPG (in series)front light running off a "pilom" circuit #7, and a paralleled low current rear led light, neither of which has any standlight. I use an oldish Shimano hub dynamo that seems to do the job nicely.

I'd like to use the double led front light if possible (by altering its present circuit), but would happily ditch the cheap multi 5mm led rear for something with a bit more output, if needed. The 2 x XPG's in series have a max Vf of about 6.3V at the sort of current I'm running at. Would the Vf be lower with your circuit?

BTW, great topic everyone!

[eh4]

How about using super capacitors as a buffer for charging NiMh batteries?

here's a thought experiment by an ebike maker about the power generation potentials with regenerative brakes on a bicycle: http://www.ecospeed.com/regenbraking.pdf

[eh4]

duplicate.

[Savvas]

Bobblehat - good question although it's perhaps not quite clear if you are simply asking about running your 2 X front XPGs with separate standlights or if you are in fact proposing a 2 + 1 front and rear standlite arrangement. Can you clarify please - this is (as you say) an interesting thread.

I'm also interested to know if you used any sort of series balancing resistor with your paralleled rear 5mm led tail light. I have tried this myself and found that - having a lower vf - the parallel rear light seemed to divert energy from the headlight. Steve Kurt pointed out to me that the vf of front and rear lights in such an arrangement may need to be balanced (by some additional rear loading) to avoid this.

Savvas

[Bobblehat]

Hi Savvas

Yep! It's a 2 + 1, 2 front + 1 rear, I was interested in, both with standlight.

I can see how elegant the existing circuit is with a single led front and single led rear, both having independent standlight functions, but it would be even better for me if I could adapt the circuit/concept to use my double XPG front light in place of the single XML. I may need to ditch the hacked multiled rear in favour of a single high power rear led …. but that would not be a big problem.

I know I’d also have to ditch the “tuning” and “smoothing” components of my Pilom #7 circuit, but the standlight part of the circuit probably does all the smoothing necessary anyway (with the bonus of the standlight), and I’m not convinced that the “tuning” part is a massive gain at the modest pace I ride at during my part busy road, part canal path commute. I’m quite ready to give up the tuning element for a good standlight function.

The rear red tail light is a bit of a DIY botch job, having been used on a 7.2V battery system in a previous life. It does suck a little from the dynamo headlight but not enough that worries me. It’s very bright, using 3P in series with 4P (don’t ask why!) ultrabright red 5mm leds.

I did originally use just a series resistor tacked on to the DC output side of the headlight’s bridge rectifier at first, about 150 ohm, but I found it had odd effects on the 2 XPG headlight … like poorer slow speed output and noticeably lower output of the headlight. I then experimented with a separate bridge and smoothing cap for the rear and the same 150 ohm series resistor and found that most of the slow speed and output problems were sorted. It’s a very suck-it-and-see approach, and probably has very little electronic merit, unless I just fluked it …. but it seems to work!

[Bobblehat]

OK ... I've had a first go, in the style of ejoyu, at a possible solution to using my twin XP-G dynamo light in the twin standlight circuit. It seems too easy but ......



........ is it that easy?

Only one XP-G is on the front standlight, and the diode strings may need tweeking to keep the 4F caps below 5.5V. I've assumed that the XP-G's Vf at a max of 600mA is 3.2V and I'm using 2.5V for a modern red power LED.

What do you reckon? Have I forgotten something?

[jdp298]

It can be that easy, but have you built this yet? I'd guess that all the lights will stay on once you stop the input, with probably the second XP-G brightest because it's got a capacitor all to itself, whereas C2 will discharge across all 3 LEDs and through R2 to ground.

If I'm honest, you're as well to make separate circuits for front and rear. If there isn't the space in the rear enclosure, then have the circuit up front and run the wires to the Luxeon Red just as you have to do now. This way you'll know what each capacitor is powering. I run a pair of 1F caps for each of 2 1W front LEDs in old dynamo light housings and a just a single 1F cap for the light at the rear. Each LED has its own rectifier, caps and LDOs, which also means any vibration or pot-hole induced damage won't take everything out.

Still, yours will work, I'm quite sure about that.

[znomit]

I would drop C2/D2 and parallel the rear led with one of the XPGs (put a limiting resistor on the rear, 500mA is too bright for a tail!).
It would make the circuit a lot simpler.

[Bobblehat]

Thanks guys.

jdp298 ..... no build yet, just testing out the theory before ordering any parts. The current set-up of 2 x XP-G works well ... but no standlight on it, nor on the rear light I hacked together. I can't figure out the route that C2 would discharge through the white LEDs? How does any current that might pass through the white LED's return to the supercap -ve?

jdp298 and znomit .... I hear the advice ....... keep it simple, eh! I was trying to get my head around the idea that putting the LEDs in series maximizes the power from the dynamo.

My current tail light robs about 80mA at peak power and is plenty bright when on the move ....... I guess I'm already living with that sort of drop on the XP-G's current and am happy with what they put out. Znomit's suggestion sure sounds easy to understand and would just mean me adding only one standlight circuit to the current set up (plus a bit of rewiring!) Would it be better to use a power red LED in parallel to the XP-G or a few 5mm ultrabrights? In asking this, I'm ignoring factors to do with side visibility etc, of the rear light at this point of the discussion, just thinking about lumens generated vs current allocated to the rear. If there is little in the choice, I'd try the ultrabrights from my current rear light, rewired to suit.

Any one got a 4F supercap to sell? I'm in the UK

[znomit]

Would it be better to use a power red LED in parallel to the XP-G or a few 5mm ultrabrights?

Personally I would go with an XPE at around 100mA. Still very bright and you're not robbing much current from the XPG. Also if the XPG loses connection the XPE can still handle the full 500mA.

[Savvas]

Any one got a 4F supercap to sell? I'm in the UK

Any one got a 4F supercap to sell? I'm in the UK

I bought some of these - they turned up OK though haven't been used yet.
See: http://www.ebay.com/itm/4-Farad-Supe...-/260922443794
Savvas.

[MikeAusC]

I bought some of these - they turned up OK though haven't been used yet.
See: http://www.ebay.com/itm/4-Farad-Supe...-/260922443794
Savvas.

The MAXIMUM rated current for these capacitors is 4 milliamps - that's not going to produce a lot of light.

[Savvas]

The MAXIMUM rated current for these capacitors is 4 milliamps - that's not going to produce a lot of light.

Mmm - hadn't noticed that! Would it be better to go back to the 'standard' Elna (or similar) 1F, 5.5v cap? This seems to work well and is (I think) what most of the commercial standlights seem to use.

Savvas.

[MikeAusC]

The only high-current supercaps I've found are all rated at less than 2.8 volts, so you need to stack them to get higher voltages, with voltage balancing circuits added.