Help - I want to power a 5 mm LED from 1.5V alkaline cell

[EngrPaul]

For reference, theres another method to this design but its significantly more complicated in terms of passive components required, but its still a 2x 2N3904 Signal transistor circuit

This looks like the schematic for the circuit I made and is pictured above. :thumbsup:

 

[Calina]

Illum:

• What is the LED current with this circuit?
• What is the efficiency?
• Will it work with two LEDs in series?
• How about two NiMH (input voltage around 2.5V) and two LEDs?

You are faster than me. I too was wondering about your first two request.

By the way, thanks Illum.

 

[Calina]

This looks like the schematic for the circuit I made and is pictured above. :thumbsup:

Great! Got to love this forum.

Do you have any idea of the efficiency of this circuit?

I also like the Zetex integrated circuit, I'm not sure that I'll find it locally though. 70% efficiency and $0.62 is hard to beat :laughing:.

 

[EngrPaul]

Do you have any idea of the efficiency of this circuit?

Actually, no I don't. However I'm capable of measuring it. Let me know if you need to know and I'll break out the equipment.

 

[Calina]

Actually, no I don't. However I'm capable of measuring it. Let me know if you need to know and I'll break out the equipment.

Need is actually a big word but it would be nice to know :naughty:.

A week ago I didn't know where to start and now...
Gosh, I'm still in the same predicament, for a different reason. To many options :devil: !

What would be the most flexible or the most efficient circuit ???

 

[Illum]

reminds me of something I made awhile ago
Max756 8-DIP set to 3.3V, squeezes out about 3.6V 120ma from a 1.5V, which I had connected to a CREE P4 and mounted everything on top of an aluminum bar where I would then connect via some sort of a tie down to picture frames.



While I bought add an AA holder to both sides it worked so well I had it connected to a homemade "CR123A battery box" where the plug is connected to. So I get to drain both AAs, and CR123As while keeping some light to a family photo:thumbsup:

Lighting done by L4
Off


On


Unfortunately for this particular project I forgot to mod in a switch, so either the battery comes off, or it gets drained. And its been a bit brutal to NiMHs which I forgot I set them in.

While most of the parts I used in this particular project is salvaged... with the exception of the Perf board, DIP holder, Max756, and the CREE P4, I only measured about 60% efficiency, peaking to about 80% at 3V, and quits at 0.7V

the CREE P4 is very efficient on a low current, I have a 2xP4 USB gooseneck light [another homemade] that shares a current or just 150ma but can light up my entire computer from above :thumbsup:

 

[datiLED]

I have built quite a few of the circuits that Illum posted the schematic for. I have done it with through hole and SMT components. I even made a PCB for surface mount components to stuff into some 3 for $1 flashlights that my wife picked up for our son. (He was 3 at the time.)

He often left the flashlight on, and with the incan bulb, it would dim out in about an hour. With the LED circuit, it would last the better part of a day. We usually found it on during that time, and would twist it off for him. It cost several times what the light was worth to build the circuit, but it saved us on batteries.

I have also picked up a few OM5448 IC's from Hendon Semiconductor. The datasheet indicates that they are very efficient (70-85%) when built per the datasheet (check out page 8). Read the datasheet, and pay attention to the value of R1 in the schematic on page 7. It needs to be lowered to 5.6k for a white LED. There is even a two level driver schematic on page 11. I have not had a chance to play with these yet, as there is too much going on in my life to mess with personal projects like that. It looks promising, though.

 

[Illum]

SOT-23 packages are rather difficult to solder by hand...unless you've made a PCB board that has contact pads for it to sit on.

Perf boards and through hole components are easier to solder, but sometimes relatively more expensive...plus sometimes by a ill-optimized combination of components and switching at a high frequency may contribute to certain inefficiencies in the design...

Personally, anything above 80% at such criteria are just creme on the cake

 

[Calina]

Here is your solution for $0.62. I also included a link to the data sheet for the driver.
http://www.mouser.com/Search/ProductDetail.aspx?qs=7EMYER6HMn0PEIRolSjI1g%3d%3d

Data sheet: http://www.diodes.com/datasheets/ZXSC380.pdf

Great! PEU just posted a circuit and a chart using this driver (actually the ZXLD381), it couldn't be simpler: https://www.candlepowerforums.com/threads/226903

 

[PEU]

I didn't see this thread earlier, the zxld381 based driver is pretty simple to assemble and its really small. It gives the impression that is current regulated, but its not, the pulses are current regulated whatever that means in the real world...


Pablo

 

[Illum]

I didn't see this thread earlier, the zxld381 based driver is pretty simple to assemble and its really small. It gives the impression that is current regulated, but its not, the pulses are current regulated whatever that means in the real world...


Pablo

yes it is, but in SOT-23 as well, mouser didn't have them in stock so I ordered the 380, at mere $0.62 a piece

I used the 380 design, its not thru-hole, so you'll have to be a little innovative in soldering to perf boards




Zetex ZXSC380 + 100uH inductor = [Vout = 3.5V (19ma)] at 1.5V in. Efficiency is about 70% so it'll pull about 64ma from the 1.5V cell but as long as your not using an AAA you can expect a long life from it.

If I step up the input to 3V it can supply up to 70ma with 90% efficiency to boot, pulling just under 91ma:twothumbs

 

[Illum]

 

[paulr]

I got some Dorcy 1AAA lights from woot for about 3 bucks each shipped. It should be pretty easy to scrounge the circuit board from one and connect it to a D cell.

 

[tylernt]

That Zetex 381 IC driver looks great but the bottom of http://old.nabble.com/to-92-LED-driver-td26835073.html says that it overdrives the LED with excess peak current. Indeed, the datasheet shows a current peak of about 6 times the average current!

Might be a good idea to add something to smooth those peaks out...

 

[Illum]

how heavily it overdrives it will depend on the frequency of its pulse, often influenced by the selection of the inductor.
According to several datasheets, T1-3/4 LEDs have a nomial input current of 20ma, but if applied in pulses, its tolerant up to 100ma:candle: