Driving a Red emitter from a supercap

[steveo_mcg]

I've got a couple of 100f 2.7v supercaps just sitting here waiting for a project. By my calculations 729 joules should give over 20 minutes of 500 mw of power (729/(.5*60)), that is spherical cow reasoning in reality but if i could get 50% of that i'd be happy.

The biggest problem i can see is the voltage range the capacitor works through, i think i need a buck/boost driver but i don't know which one or where to source one from, I'm in the UK so a UK source or Honkers/China would be best. The wider the range of the driver the more power can be extracted from the cap. There is a DX driver that works down to about 1v but when the cap is fully charged at 2.7v it would melt a red emitter.

Any ideas?

 

[DM51]

Re: Driving a RED emitter from a supercap

Welcome to CPF, steveo_mcg.

I think your thread will fit better in the Batteries/electronics section, so I'll move it there now.

 

[steveo_mcg]

Re: Driving a RED emitter from a supercap

Cheers mod, missed the electronics section in my searches...

 

[tylernt]

Boost circuits have a hard time with input voltages below 0.9v or 0.7v because that's the minimum "turn on" voltage for most transistors. Some can keep going down to 0.4v if they were turned on at or above 0.9v, but won't start up at that low voltage when first switched on. You'll either have to give up this part of the supercap's capacity or use two supercaps in series to cut your unusable portion in half.

There is a DX driver that works down to about 1v but when the cap is fully charged at 2.7v it would melt a red emitter.

The emitter should be fine if you are using a current-regulated buck AND boost driver. The boost part would allow operation down to low input voltages, and the buck part would limit current to a safe value.

 

[steveo_mcg]

Boost circuits have a hard time with input voltages below 0.9v or 0.7v because that's the minimum "turn on" voltage for most transistors. Some can keep going down to 0.4v if they were turned on at or above 0.9v, but won't start up at that low voltage when first switched on.

That's interesting to know, thanks. I don't think it would impede the performance too much for this application since the cap would be fully charged when turned on and would discharge down to the voltage where the chips just couldn't work.

You'll either have to give up this part of the supercap's capacity or use two supercaps in series to cut your unusable portion in half.

Of course! Thank you! If I rearrange the configuration to parallel the amount of energy stored remains the same but i don't need to worry about the buck+boost driver a simple buck driver will suffice and when the voltage drops below what the driver is rated to it will go on direct drive but since its a red led it should stay lit for a couple of minutes after that. As an added benefit the buck driver should be much more efficient than a boost.

Thanks mate, i got blinkered in to keeping the Farads up and doing some thing clever in the electronics.

With that thought in mind what would be the best buck regulator? Needs to work down to about 2v and should ideally be fairly low output 100-300 ma.

edit: would this do? http://www.kaidomain.com/ProductDetails.aspx?ProductId=9607

 

[Mr Happy]

The energy stored in a capacitor is proportional to the square of the voltage, so three quarters of the energy is stored in the upper half of the voltage range. That means your capacitor will discharge 75% between 2.7 V and 1.35 V. You don't need to worry too much about very low voltages if you have an efficient driver. This would be especially effective if you put two capacitors in series and used an efficient buck driver.

Note: You would need to use a buck driver, not a linear driver like the 7135. Linear drivers waste too much energy when you have a meagre energy source.

 

[steveo_mcg]

The energy stored in a capacitor is proportional to the square of the voltage, so three quarters of the energy is stored in the upper half of the voltage range. That means your capacitor will discharge 75% between 2.7 V and 1.35 V. You don't need to worry too much about very low voltages if you have an efficient driver. This would be especially effective if you put two capacitors in series and used an efficient buck driver.

Note: You would need to use a buck driver, not a linear driver like the 7135. Linear drivers waste too much energy when you have a meager energy source.

Would i be able to adapt a dx 3256 i've a couple left over. They output about 800ma could i drop that to about 300? I suppose the question is could i adapt it easily? I'm guessing the minimum voltage rating is for running a white emitter?
http://www.dealextreme.com/details.dx/sku.3256

 

[Mr Happy]

I understand the low current modification involves unsoldering and replacing a tiny SMT resistor on the board with one of a different value. A very fiddly job but possible if you are up to it.

As to whether it would work with less than 3.6 V input given a 2 V emitter I do not know. Perhaps it would and you have nothing to lose by trying.

 

[steveo_mcg]

I understand the low current modification involves unsoldering and replacing a tiny SMT resistor on the board with one of a different value. A very fiddly job but possible if you are up to it.

As to whether it would work with less than 3.6 V input given a 2 V emitter I do not know. Perhaps it would and you have nothing to lose by trying.

An excellent point.

I've done a small amount of smt soldering before, a couple of small led's that wasn't much fun but its do able.

Thanks for both of your inputs on this.

 

[Capo_au]

Note: You would need to use a buck driver, not a linear driver like the 7135. Linear drivers waste too much energy when you have a meagre energy

Any chance you could elaborate on that?

Thanks! :thumbsup:

 

[tylernt]

Any chance you could elaborate on that?

Linear drivers drop the current by bleeding some of it off as heat, which is wasteful. Buck drivers simply chop up the current into bite-size pieces using PWM or similar scheme, so nothing goes to waste (except a tiny bit to drive the chopping circuit itself).

 

[TorchBoy]

Re: Driving a red emitter from a supercap

Note: You would need to use a buck driver, not a linear driver like the 7135. Linear drivers waste too much energy when you have a meagre energy source.

Actually if your meagre energy source has the right voltage then linear regulators are very efficient. The heat generated by burning off the excess volts is analogous to the power consumed by buck drivers. It's not waste, it's a cost of regulation. Buck drivers just do something different when converting input power into heat, and it's not necessarily a "tiny bit" of power they consume. Some have quite a high cost of regulation.

The AMC7135 though drops out of regulation under 2.7 V so I can't think how it could be used.

Steve, are you committed to be using only one red LED?

 

[steveo_mcg]

Re: Driving a red emitter from a supercap

Steve, are you committed to be using only one red LED?

I wouldn't go as far as commited i had thought about a red cree from dx (1776). I'm running one currently on my road bike with a couple of AA's and a resistor and its more than bright enough at about 200-300 ma tbh its quite bright when the batteries start to run low and it drops to 100ma.

What did you have in mind?

http://www.dealextreme.com/details.dx/sku.1776

 

[TorchBoy]

Re: Driving a red emitter from a supercap

I was wondering if putting two in series (thereby giving a higher total Vf) would open up a few more possibilities. I might ponder some more.

 

[steveo_mcg]

Re: Driving a red emitter from a supercap

Perhaps, but i think it would reduce the effective power available from the supercap. Two in series would be ~4v the caps max v is ~5.4v. Running down to 1.8v would give about 40% of the available energy in the cap but down to 4v would only give 7% so that wouldn't be ideal.


I've been pondering the above.... If I didn't already have the caps 4p1s (50f) would be a better arrangement. The same total amount of energy would be available but the choice of drivers would be much simpler and by switching one of the emitters off a little more time could be eeked out.

 

[Mr Happy]

Re: Driving a red emitter from a supercap

Actually if your meagre energy source has the right voltage then linear regulators are very efficient.

But in our case we don't have the right voltage. We have a capacitor starting out at 5.4 V and being discharged down to ~2 V, releasing about 85% if its energy. A linear driver in this case would be dropping 5.4 V to 2 V at the start of the run making it only about 37% efficient.

 

[steveo_mcg]

Just done a quick test with a spare driver i had lying around, with 4 fresh cells it works fine on the red emitter, maybe a little hot, with a Vout of 2.4v but from one supercap at Vin 2.7v it won't even direct drive. I need some more test leads to try running them in parallel to bump Vin a bit higher and find out where it starts to operate but the fact that it wont work at all at 2v means this driver is probably a non starter since too much of the energy is unavailable.

 

[TorchBoy]

Re: Driving a red emitter from a supercap

I'll try again. Linear regulators are normally not well suited to use with capacitors more because of a capacitor's greatly-changing discharge curve than because they are "a meagre energy source." Linear regulators can still be well suited to some meagre energy sources.

 

[Mr Happy]

Re: Driving a red emitter from a supercap

I'll try again. Linear regulators are normally not well suited to use with capacitors more because of a capacitor's greatly-changing discharge curve than because they are "a meagre energy source." Linear regulators can still be well suited to some meagre energy sources.

Fair point.