Squeezing the power out of a Supercap

Z

z_jdp298

Guest
This would be my first post to CPF.



I have a hub-dyno rear standlight of my own design, which without drawing the diagram, is a bridge rectifier, 5.5V 1F supercap, then an LM317 regulated to 2.1volts, fed across 1 red LED.



Its not the only rear standlight, I have a BM tophat or whatever they call it. The circuit works just fine and the supercap from Maplin seems to cope with whatever comes its way. On test I used a Hornby controller and it didnt even get warm at 7v. Admittedly the load was minimal at that point. the Bridge Rectifier holds the dyno voltage down, as does the ohmic resistance of the cable from fwd to aft.



Given the LED needs 2.1 volts, the supercap will drive it until it shows about 2.8v, at which point it can't squeeze through the LM317 and it fades. I had reckoned to use a PNP transistor, opening up when the output of the LM317 went low, thus connecting the supercap straight through for a bit more power, but that's not effective, nor can I make it work. Is there some voltage regulator IC or circuit design which can drag some more power out of the supercap for me?



Jon, near Swindon
 

jdp298

Newly Enlightened
Joined
Mar 22, 2011
Messages
115
Location
With your sister, somewhere you wouldn't like
How did that happen then? There were some nice replies to this before but I suppose things have been shaken up a bit by the rebuild.

No matter, I have improved this design enormously and it stays on for ages, brighter than the Busch & Muller standlight above it and longer.

A Low drop-out regulator, 3mm oval red LED with a bridge rectifier, smoothing cap and supercap (1F).

http://www.maplin.co.uk/150ma-ultra-low-dropout-positive-voltage-regulator-46323?ordercode=N68CA

http://www.maplin.co.uk/superbright-oval-leds-45920

I think I might be over-driving it but it's been good up to 25mph with the hub dynamo and hasn't given up yet. If it goes too quickly I might just put 2 of the LEDs in series to protect each other.

Before the replies talked about boost/buck regulators etc. Is there is a reason no one likes these?

Jon near Swindon
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
I think I remember some discussion about this prior to the Great Crash. Wasn't I suggesting the Zetex ZXSC310 as a way to efficiently use the charge in the cap? It's a handy little gadget that runs down to 0.8v, IIRC.

I like LDO's, but not as a led driver. There really needs to be some way to limit the current. Maybe if you put some resistance between the regulator's output and the led, just to limit the current?? I'd rather see some form of current regulator, preferably one that had a very low headroom requirement. I tried looking up a simple current regulator, and didn't see exactly what I wanted. National Semiconductor does make a simple current source that can deliver up to 10mA and only needs 1v across it to work.

This would be an excellent excuse to learn about energy storage in caps! The fundamental equation is E = 0.5 x C x V^2. E is joules, and you'll want to calculate the energy when you start with a fully charged cap and the energy that's left when you can no longer make use of the remaining voltage.
I'll throw out the other basic equations too: Q = C x V and I = C x dv/dt

The basic theme is that you'll get more energy out of the cap if you can use the voltage down to a very low level. i.e. draining the cap down to 3v is using less of the available energy than if you can drain it down to 0.8v.

I think the thread about standlights was recovered... it'll have a lot of the relevant discussions and ideas in it.

regards,

Steve K.
 

jdp298

Newly Enlightened
Joined
Mar 22, 2011
Messages
115
Location
With your sister, somewhere you wouldn't like
I thought that was kind of the point of an LDO, it was a current regulator? I didn't get too much from its data sheet. The one I'm using is meant to be 150mA at 3.3v which to be fair is over-driving the LED but it's not burning out. I could add a resistor I suppose, that'd be reasonable.

We got round to mentioning the functionalities before but I don't think you mentioned the ZXSC310 by name. 2 rather simplistic reasons I might not go with it. (i) This works nicely (ii) It'll be a bigger board. I'm a bit strapped for space in the old mudguard reflector/dynamo filament light housing, only just get the current setup inside it.

I agree with the capacitor equations completely, it also shows me that I get a squared law increase in energy the higher voltage cap I can find. Not too many 10v 1F caps about though.
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
LDO's are normally voltage regulators. The 150mA is probably the amount of current it can pass without damage. The datasheet should be detailed enough to tell you.

did we discuss current regulators built from two transistors? It does eat up about a volt, which is a lot in a design like this. It sort of depends on what the forward voltage of the led is. Do you have any idea what it is? i.e. what the Vf of your led is?

In a space constrained design, part of the math is to figure out how big of a cap you can fit, and then trade the cap's voltage against the capacity. Going to a higher voltage rating will mean the cap has to be larger. In some respects, the physical volume of the cap is directly related to how much energy it can store. The trick will be optimizing how much energy of the stored energy you can utilize. This sort of optimization is one of the essential skills of a good engineer. :)

Steve K.
 

jdp298

Newly Enlightened
Joined
Mar 22, 2011
Messages
115
Location
With your sister, somewhere you wouldn't like
Both the bits ofm information you asked for are at the links. I can't make head nor tail of the LDO datasheet and what Vout is meant to be. Vf for the red LED is meant to be 2.2V but if I'm at 3; it's not in excess of the stated maximum reverse; inside the other typical Vfs for the range. But there's no data for max Vf.

Ultimately much better than a LM317 which droped out at about 3V no matter what it was regulating to. I could use the 8-pin version of this LDO to really set the voltage correctly but that'd be space I don't have to populate the rest of the circuit.

As I've said, the circuit is in and working in the available space. This cap at 5.5V 1F. http://www.maplin.co.uk/double-layer-capacitors-98185
 

jdp298

Newly Enlightened
Joined
Mar 22, 2011
Messages
115
Location
With your sister, somewhere you wouldn't like
Hmmm, perhaps I was overdoing it. Had to replace the LED. Found another spare 'normal' superbright red LED and threw a 18ohm resistor in there. It survived nicely after a couple of steep hills saw me at 34mph. I also got a spare oval LED and the right resistor for 3.3v should this one give up on me. I have to say, though, that I still don't really understand the problem of a voltage regulated power supply for an LED. Will Ohm's law not take over and whatever internal resistance the LED has limit the current in the appropriate way?

Spearately, does anyone know what LEDs Busch and Muller use in their headlights? It's time I made a new front light to complement the halogen bulb with an equally bright LED standlight. And I'm thinking LDOs are the way ahead again.
 

jdp298

Newly Enlightened
Joined
Mar 22, 2011
Messages
115
Location
With your sister, somewhere you wouldn't like
We got there, front light circuit made, tested and waiting an enclosure! Some bike shop ought to have something with a reflector I can bodge.
5628099317_cb38b63ab4_m.jpg

5627869621_a3c0022882_m.jpg

Will this standlight work? Yes it will. Stayed on for 10 mins after 2 mins running with test source (Hornby railway controller) at 3 quarters power. After about 5 minutes it wasn't probably enough to even be seen by. But the one at the back of the bike with the same circuit lasts longer at full brightness than the Busch & Muller above it. (Admittedly the B&M turns itself off after 4 mins).
 
Last edited:

lumière de bicyclette

Newly Enlightened
Joined
Oct 10, 2011
Messages
6
Hi There,
I know this thread is a bit old, but I've just come across it.

I've been experimenting over the last weeks with different configurations for bike lights using a shimano hub dynamo.

It looks from your photograph that you're using an Osram LED, which is the one that I've found to be one of the better ones. However, I ended up using a 3W Dragon X plus. rather than a 1W, I just used a full wave bridge, with a !F 5.5v supercap (maplins) connected directly to the LED,

As the dynamo only gives out a maximum current of 500mA, as soon as the LED draws this amount of current, the voltage is self limited to about 3.6V. There's a small amount of strobing at very low speed with such a low value capacitor, but over about 7mph it's not noticeable, and on coming to a halt, the standlight function works for several minutes, all be it as a light to be seen by rather than to see with!

Also used the same configuration using two similar LEDs in parallel one with a 5 degree collimator, and one with a 30 degree. but this time used 2 x 10F 2.5v caps in series (to give 5F 5V) This set up works well, and there's no noticeable loss of brightness to one LED when the second one is connected.

Both lights have been tested on the big hill where I live at up to 38mph, no problems.

Just bought a couple of Samsung sunnix 6 180 lumen LEDs which are currently at the giveaway price of 1.72 euro post free from wwwLED1.de, these have a max current of 250mA, so 2 in parallel take the 500mA from the dynamo. the voltage required to drive these is 8.9V which the dynamo easily gives out, again self limiting when the max current is drawn. I've use 4 x 20F 2.5V caps in series ( to give me 5F 10V) and with the bike on a stand indoors the results look very encouraging. Certainly more suitable for high speed riding, so may have the finished unit switchable with the originally described unit.

My only problem really, is the beam patterns, which I don't think are as good as using a reflector such as the B & M units. I did buy an excellent LED bike light from LIDL with a single LED mounted above the reflector shining downwards which gives a good beam pattern, but it's ultrasonically welded together, and I'm loathed to destroy it, however, when /if they come on offer again, (they were only about £8 I think ) I intend to get a couple and butcher them for the reflectors.

Hope this may have been of interest.....pete
 

MikeAusC

Enlightened
Joined
Jul 8, 2010
Messages
995
Location
Sydney, Australia
. . . . I still don't really understand the problem of a voltage regulated power supply for an LED. Will Ohm's law not take over and whatever internal resistance the LED has limit the current in the appropriate way? . . .

At the same voltage, a change in temperature can cause the LED current to vary greatly.

If you increase the LED voltage by 10%, the LED current increase TENFOLD.
 

MikeAusC

Enlightened
Joined
Jul 8, 2010
Messages
995
Location
Sydney, Australia
Why not use the one-transistor current regulator.

LED connects to +ve and collector of NPN transistor. Current sensing resistor is selected to drop 0.6 volt at the the desired current - connects to Emitter and -ve rail.

Two silicon diodes in series connect cathode(+) to -ve and anode to Base and to a resistor which also connects to +ve. The resistor should feed 5ma at (V+ve)-(1.5).

You need to make sure the Supercap NEVER has more than 5.5 v across it.
 

jdp298

Newly Enlightened
Joined
Mar 22, 2011
Messages
115
Location
With your sister, somewhere you wouldn't like
The 1F supercaps from Maplin are rated 5.5V normally, safe to 6.3v surge, so says their website. I've been wrestling with the front light, conceptually, for a little time now. Once I stop, the brightness immediately drops but then stays roughly constant for a couple of minutes before fading away again. I've never been too worried about thermal runaway as the safe range of the LED exceeds that which the LDOs regulate to. What I want is to stay at max brightness for as long as possible. I think the next step might be to take a similar risk to the rear light and do away with the zener, or at least push it a little closer to 6v. (Imagines heart attacks among the careful). And replace the 3.3v LDOs with 5V ones and a resistor in series with the same LED. It works out to only be 4 ohms.

Separate question, my B&M DT toplight plus standlight is staying on for about 20 seconds max, even after a 13 mile commute. Is it getting old and has anyone ever successfully got inside one and fixed it?
 

lumière de bicyclette

Newly Enlightened
Joined
Oct 10, 2011
Messages
6
Apologies if this post is out of sequence, I posted one yesterday, but it's not yet materialised. <br>
<br><font color="#0000cd">?. See post 9.<br>
<br>
Bill<br>
<br>
</font><br>
You say you're wanting the front light to stay bright for longer, I see from your circuit diagram, that you're using the two 1F caps in parallel giving you 2F in total, with the LED drawing a current of say 350mA, the voltage in the cap will fall pretty rapidly, so the led will take less current, and stay dimmer longer if that makes sense. There won't be enough energy in the cap to support full brightness for very long.<br>
<br>
I tried using two 20F caps ( in series for higher voltage as they were only 2.5V) that kept the led lit brightly when stationary for a lot longer, trouble was, that at the start of the ride, it took ages for the caps to charge up to the LED's operating voltage.<br>
<br>
Perhaps it would be better if we used (and a lot cheaper) some re-chargeable batteries?<br>
<br>
I've considered using protection circuitry, but really don't see the point, the LEDs I use safely take all the current that the dynamo will produce, so self limit the voltage. <br>
<br>
Think I could well have bought a commercial one for all I've spent on LEDs and caps! still it's not as much fun is it!<br>
<br>
I did buy an excellent LED bike light from LIDL with a single LED mounted outside the reflector similar to the B & M unit, It runs from 4 AA cells and draws 420mA and 210mA on the two brightness settings, I did consider just connecting the dynamo directly to the (re-chargeable) cells through a bridge rectifier.<br>
<br>
Pete....Sheffield
 
Last edited by a moderator:

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
Perhaps it would be better if we used (and a lot cheaper) some re-chargeable batteries?<br>
......
Pete....Sheffield

that's the path I've chosen.

The plus side is that I get long run times and fairly bright standlights. The negative is that the circuitry is more complex and I use a switch to disable the standlight when I park the bike (don't want to completely discharge the AA nicad). A CMOS timer could probably be added to get rid of the switch, I suppose.

A lot of the issues with standlights come down to basic math. Any power used to charge the supercap or battery comes at the expense of the headlight (or taillight). The standlight run time is related to the size of the storage device as well as how effectively you can utilize the energy in the storage device. It also depends on how bright you want the standlight to be.

There's also the question of the duty cycle of the standlight, i.e. what percentage of the time will the standlight be running. If you spend a lot of time stopped at intersections, then it's going to be very difficult to keep the standlight powered.

While it's fun to tinker, sometimes it helps a lot to do the calculations and see what it even possible. This might avoid wasting time and money on ideas that don't work out.

regards,
Steve K.
 

SimonInd

Newly Enlightened
Joined
Oct 19, 2011
Messages
20
I have a XM-L at the front and a XP-E (Red) at the rear running off a Shimano hub dynamo. I've found that a bunch (4) of 5.5v 4F supercaps that I picked up on eBay work brilliantly in parallel as a standlight for the XP-E, giving me bright rear light for 10 minutes or more.

The XM-L drains the caps instantly so for the front light I'm going to need to discharge them through a resistor, probably using a circuit like this to get a some more voltage across the caps.
http://www.candlepowerforums.com/vb...ight-circuits-(Will-this-circuit-work-)/page3

The capacitors ovbviously have a reasonable internal resistance as they don't stop the rear light coming straight on when I start up, and they do take some time to fully charge.

The rear standlight experience leads me to believe that supercaps are the way to go, I'll report back once my front light is done.

Cheers
Simon
 

pantaz

Newly Enlightened
Joined
Oct 19, 2011
Messages
9
I've been looking at various bike light projects for ideas to retrofit my old (ancient) halogen unit. I wasn't even thinking about supercaps until I spotted this thread!

I'm likely going with a higher output main headlight, but a supercap circuit would be an excellent backup -- no worries of exhausted batteries!

I'll be sure to post updates if I learn anything that isn't already covered here.
 

jdp298

Newly Enlightened
Joined
Mar 22, 2011
Messages
115
Location
With your sister, somewhere you wouldn't like
Back to the drawing board. Have found some 50F 2.3v (yes, fifty!) caps on Farnell for sensible prices. So, incorporating the cct Steve K appears to have made work, here:

http://www.candlepowerforums.com/vb...ight-circuits-(Will-this-circuit-work-)/page3

I can cheerfully get on with things. But for a couple of questions:

1) How do you calculate the Inductor and I(sense) resistor? I've been looking at the datasheets all evening and got nowhere. After all this time I'm not even sure how the ZXSC310 even works in the first place. I want to use a single 3.8v 700mA 3W LED or a 1W 3.5v 350ma one. Obviously the Standlight element could be lower current, but I want to know how to choose.

2) In Steve's cct, here:

http://www.flickr.com/photos/kurtsj00/5420960597/in/set-72157621965148305/

I get holding the shutdown pin low when the wheels are moving. But the datasheet says 0.7v is needed to enable it. How do you get round this?
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
...
I can cheerfully get on with things. But for a couple of questions:

1) How do you calculate the Inductor and I(sense) resistor? I've been looking at the datasheets all evening and got nowhere. After all this time I'm not even sure how the ZXSC310 even works in the first place. I want to use a single 3.8v 700mA 3W LED or a 1W 3.5v 350ma one. Obviously the Standlight element could be lower current, but I want to know how to choose.

page 9 of the datasheet (my 2002 copy, at least) has a table of suggested inductors. They are all 68uH and rated for about 0.5A. I ended up with a slightly larger inductance and higher current rating part that I got out of Digi-key. I picked the part about 8 or 9 years ago (I think..), so I'm not sure exactly how that happened.

page 10 of the databook tells how to pick the current sense resistor. This sets the peak current in the inductor, which then sets how much energy is stored in the inductor. I think I experimented a little bit with different values, which I do recommend.

I don't think there will be a problem with either LED, but the current out of the zxsc310 will in the 10's of milliamps, not the 100's.


...
2) In Steve's cct, here:

http://www.flickr.com/photos/kurtsj00/5420960597/in/set-72157621965148305/

I get holding the shutdown pin low when the wheels are moving. But the datasheet says 0.7v is needed to enable it. How do you get round this?

The datasheet says that the Enable pin can have a bias current of up to 1uA positive or negative. My experience has been that the pin sources current, and if you don't connect anything to it, it will float high. That's why I just added a switch to ground. You could add a high impedance resistor from the Enable pin to Vcc just to be sure that the pin is forced high. I think I avoided this to minimize the drain on the battery when the bike is sitting around.

regards,
Steve K.

edit: as far as how the zxsc310 works, it's a simple boost converter. There's a decent wiki page on this type of converter.
http://en.wikipedia.org/wiki/Boost_converter

I'm avoiding calling the zxsc310 a switching regulator, because it's not really regulated. The output varies a fair amount as the input voltage changes.
 
Last edited:

SimonInd

Newly Enlightened
Joined
Oct 19, 2011
Messages
20
Quick update for those that are interested, I've managed to come up with a simple circuit that gives > 5 minutes standlight for both front and rear LEDs seemingly without wasting any current from the dynamo when you're on the move.

It uses a CREE XM-L for the front and a red CREE XP-E for the rear, connected in series so they can both get 0.5A when the dynamo is running. I then have 2 x 4F 5.5v super cap in parallel with the XM-L via a 47R resistor with a 1N4001 (D1) to limit the maximum voltage accross the cap as per the circuit linked to in my previous post. Where this circuit differs is that second LED, the rear, has 4x4F 5.5v supercaps in parallel with it and the 47R resistor has a 1N4001 (D2) in series to prevent the rear's supercaps from discharging back through this resistor and D1.

The result is very bright lights as soon as you pull off (I'm guessing the supercaps can't suck in too much current) and once they're charged, e.g. after a minute or two of riding, you have standlight functionality for both front and rear.

I can't upload a diagram at the moment but will when I get the chance, and if you're wondering about why I'm using 4F 5.5v supercaps it's because you can buy 10 for about £14 delivered on eBay which seemed like a good starter pack for experimentation...

Now I'm going to stuff all this is some nice chrome vintage lights to finish the job...

Cheers
Simon
 

Steve K

Flashlight Enthusiast
Joined
Jun 10, 2002
Messages
2,786
Location
Peoria, IL
The price for the supercaps sounds very good! Of course, I've only bought one supercap and that was a couple of years ago. Can you tell us the details of the supercap, such as manufacturer, model, p/n, etc?

For comparison, when I search Digi-key for a 4F cap, it shows me a 3.3v part from Taiyo Yuden for $4.38. The diameter is 10mm and length is 22mm.

regards,
Steve K.
 
Top