Well, I found lithium rechargeable batteries that are 14500 size, which is the same as AA and they are 3.7V, 900mAh. Now I can go back to a very simple current limiting circuit, just running the 8 LEDs in parallel with the battery. I think this makes my life easier. I'm still going to continue looking into the various joule thief circuits as I'd like to be able to use a more readily available and common battery.

 

I'm still trying to understand ways to boost the 1.5v or 1.2v from an AA battery. Would a DC-DC step-up converter work for this type of application? Seems like a pretty simple solution. Any pros or cons to using these?

 

You have not calculated how long a single battery cell will last at the very high current you are talking about.
A modern 1.2V Ni-MH AA cell is 2000mAh (they are rated at 2200mAh for an end voltage that is much too low to power anything). As it runs down the LEDs will be slowly dimming the entire time.
The Lithium cell needs to have a circuit that disconnects the load when its voltage drops to 3.0V.

If you have eight 3.5V white LEDs at 25mA each then they use a power of 8 x 3.5 x 0.025= 0.7W. A voltage stepup circuit uses power for its operation which might be an additional 0.1W so the battery must supply 0.8W.

The current from the 1.2V battery is 0.8W/1.2V= 667mA which the battery cell can provide for "only" 2000/667= 3.0 hours. At the end of 3 hours the LEDs will be so dim that they will barely be visible in the dark.

 

3 hours would almost be acceptable if brightness were maintained, but it sounds like it wouldn't be.

The Lithium batteries have a protection circuits, so I think they should be ok.

Still hoping to find a way to power with regular AA batteries...

 

Im also looking for a good circuit to power LEDs from a single AA. From what ive gathered, a step-up voltage regulator would increase the battery life, but be aware that regulators have minimum input voltages. The lowest ive seen is 0.2v, but generally theyre about 1v, so using a single AA battery would power the circuit until it reaches two 3rds of its life, so lets say a <snip> AA has 1500mah, that gives 500mah until the circuit dies. Im looking to build a circuit of 10 LEDs at fV:3.3v fI:20ma, so my circuit is around 200mah, so using a regulator would give me 2.5 hours of battery life. 2 AAs however would give me around 10 hours (i think....). My problem is both confirming that a regulator would solve this problem, and sourcing the right component for the job. Any help from someone in the know would be greatly appreciated.

 

A linear voltage regulator wastes power by throwing away excess power as heat.
A single Ni-MH battery cell drops to 0.9V as it runs down. You would need a circuit to stepup the voltage to at least 3.3V when the battery is 0.9V then the stepped up voltage would be 5.1V when the battery cell is fully charged at 1.4V. A low dropout voltage regulator set with an output of 3.3V would throw away 1.8V at the current in the LEDs which is 55% the power used by the LEDs.

Your 10 LEDs use a total of 200mA at 3.3V which is 0.66W. The stepup circuit uses about 0.1W and the regulator wastes an average of 0.18W so the total power from the battery is 0.94W. The average voltage from the battery cell is 1.2V so its average current is 0.94W/1.2V= 783mA. A cell rated at 1500mAh will last LESS THAN 2 hours.

 

I thought the point of a regulator is to maintain (for example with a 3.3v regulator) 3.3v no matter what the input voltage? So there would be no differing output voltage from the regulator using a drained battery or a fully charged battery. Am i wrong about that? Also, yeah i understand youd get a couple hours using a single AA, but with a second AA the input voltage would be ~2.5v, and if i used a regulator with a minimum input voltage of 1v, wouldnt the circuit run for around 7 hours? Im not sure about the method i used to work that out though Apart from a regulator, ive found no solution to using a single (or two) AA battery to power 10 LEDs for anywhere over a couple hours. I could just keep replacing my batteries but there must be a better solution...

 

An ordinary 3.3V linear regulator needs an input that is at least 1.5V more so the input must be at least 4.8V.
A low dropout linear regulator needs an input that is at least 0.2V higher than its output.
These linear regulators do not boost an input voltage, they reduce it by throwing away extra power as heat.

Nobody makes a 1V regulator. You must use some kind of voltage stepup circuit then regulate (reduce) the stepped up voltage. But when the voltage is stepped up the current from the battery is also stepped up.

 

hmmm i understand. How about a dc-dc converter to step up the AA to 3.3v? Surely its not uncommon in a project to want to get many LEDs lit using a low power supply, i dont understand why methods to do so are so elusive

 

Also, after talking with someone with more experience than myself, he advised that a current boost regulator might be my best option. Could you give some info on that?

 

There are voltage stepup circuits that regulate the current with pulse-width-modulation.
LEDs regulate their own voltage and must have the current limited or regulated so these stepup circuits are ideal. The output transistor or Mosfet switches on and off so it barely gets warm. I think Maxim make the ICs.

 

Surely its not uncommon in a project to want to get many LEDs lit using a low power supply, i dont understand why methods to do so are so elusive

To paraphrase your question, "Surely its not uncommon in a project to want to travel in time, i dont understand why methods to do so are so elusive"

It's the laws of nature, and in your case it's tough to light up a whole bunch of LEDs for a long time from a single battery.

That said, if I was trying to do this, I think I'd go buy a bunch of solar lights for less than $2 apiece and see if you can use the circuits in them. These days they're using patented, specialized circuitry. They'll run a single LED all day off a charged AAA battery (not the cheapo one that comes in it). What happens if you add more LEDs in parallel to one of those circuits? Or, what happens if you parallel multiple LED drivers off a single battery? No need to reinvent the wheel here, just tweak it a little.

 

Interesting thought wayneh. I actually did a quick search and found this:

http://www.dealextreme.com/details.dx/sku.44463

A little solar powered string of 60 LEDs. I'm definitely curious about what kind of circuitry is inside that thing, but don't know if I'm willing to spend the $17 to try to reverse engineer it.

 

Aside from the fact that every one of us travels in time constantly, and time travel in itself is not impossible, i dont think im being ridiculous in what im trying to do. All i want to do is to fully utilise a battery using my LED circuit, and not just have the circuit cut off after a very small amount of time due only to the LED drop voltage. Electronics is all about workarounds, there must be a simple solution, or at least there must be a definitive "simplest" solution to my problem....

 

I asked this question to Maxim technical support and this was their response:

For your design, a step-up DC-DC converter would probably be the better solution. We have a few different devices that may work dependent on the load current. The MAX1760 can supply up to 800mA and the MAX1674 can supply up to 300mA though it's more efficient. Here are links to their datasheets:

http://datasheets.maxim-ic.com/en/ds/MAX1760-MAX1760H.pdf
http://datasheets.maxim-ic.com/en/ds/MAX1674-MAX1676.pdf

Regards,
Maxim Technical Support

 

is there a UK equivalent to maxim?

 

I can buy a few British parts here in Canada.
Can't you buy American parts (Maxim) in Britain?

Farnell probably sells them.

 

Very likely it would get stuck.

Take any of the circuits, and add a Schottky diode between the LEDs and the circuit output, and a 1uF or larger cap across the LED string. If your current is quite low (<40mA average, 80mA max) you could use 1N914/1N4148 switching diodes.

I arrived at this forum looking for a way to drive a colour changing LED from a Joule Thief. This suggestion is exactly what I wanted -- it works a treat.

I'm using 1N5819 Schottky diodes and 1uF tantalum bead capacitors, all from eBay.