minimum draw 5mm solutions

I´m looking to make a seven candle christmas decoration, but I´m looking for to make it run on a couple of C-cells... I want to use 5mm high power LEDs. But I don´t know anything about connecting seven of these to a few easily hidden batterys!! I´m looking for a long batterytime... preferably a month constantly on...
What is your suggestions for me to do?

I built a light for my brother-in-law's large humidor. It was two strings of 5 each 5mm Cree LEDs. Each string ran off of 2 D alkalines. Each string was wired in parallel and pulled something like 10mA total at 3V. I forget the exact current draw (notes at home) but it would run for like 5 years straight. I didn't even install a switch because of this.

You could do the same with C cells. Running for a month should be no problem. You can get LEDs and a battery holder at Radioshack.

I have to say it would be easier, quicker and possibly cheaper (if you bought it from an electronics surplus 'site) to just use an AC-DC adapter, then no battery expense either. For example a 24V (few hundred mA current rated not because you need that much current, just because it's what is commonly available at low cost) DC adapter, 7 LEDs in series (typical forward voltage per LED ~ 3.2V) and a 1KOhm resistor in series would give you roughly 10mA through the LEDs.

Here's an example transformer,
http://www.allelectronics.com/make-a-store/item/DCTX-243/24VDC-300MA-WALL-TRANSFORMER/1.html

Those who are mathematically inclined will note that the math doesn't add up, that with a 24VDC supply you would use a smaller resistor in series to put 10mA through the LEDs. However, the example is an unregulated supply so it will be floating at near it's peak voltage with only 10mA current so it's peak voltage is about...

24V * 1.414 - 1.4 (internal bridge rectifier forward voltage) = 32.5V

(32.5V - (7 LED * 3.2 forward voltage)) / 0.01A = 1010 Ohm resistor... round down to 1KOhm

Over the years I've built quite a few 5mm LED lights for use in live theatre. They are designed to be powered by 9v batteries that have already powered a body mic in a show and no longer have enough capacity to power another one. Many theatres have boxes of these batteries sitting around just waiting for a way to be useful.

I make two kinds of lights to utilize these 'free' batteries:

The first type of light is basically a beacon. The power runs from the battery/resister/clip up 10" of insulated twin conductor to a standard female mini phone jack. An LED is installed in the end of the male mini phone jack. This is a perfect setup for when a singer must walk down a set of stairs into the dark house during a performance. One or more of these beacons may be held with gaffer's tape at the corners of the steps leading down from the stage to the audience. The 10" leads allow the battery to be swung underneath and hidden by masking. Each LED may be masked with tape to prevent any in the house from seeing it. It doesn't need a switch as the light goes out when the phone jack is unplugged.

The second light of this type is simpler and is used for lighting drop boxes in quick change shows and both backstage and onstage prop shelves. These use blue LEDs and 9v spring battery clips. I epoxy the LED and the resistor into a 1/2" pvc pipe cap that's been screwed to the existing hole in the battery clip. They may be held at any angle to illuminate any item or items with magnets or velcro.

Both types of lights rock as stairways, drop boxes and prop shelves are often on wheels and the batteries will drive the lights anywhere. A stage manager or prop person may have light to load the props for act two offstage and then push the shelf unit into place onstage without having to run a cord or worry about whether the lights will work. The same is true for the stairway beacons. In many shows the stairs may be moved into place not long before it's intended use so having them battery powered not only makes their deployment simpler but also much more reliable than if they were plugged into the wall.

I know a AC-adapter would be easier, but the thing is.. The light will be standing in a place where no cables are possible... It has to be driven by batteries.

And I guess a couple of 5mm LEDs connected in parallell would do well. What is the normal voltage to run these 5mm LEDs at?

You'll need the LED spec sheet to be sure of the "typical" forward voltage, but generally for ultra-bright white it is about 3.2V.

That's for them to have their rated output, since you want this to run for a month they will have to run at lower current than that needed to meet their ratings, and so the forward voltage requirement will also drop.

However, to get a useful amount of light for a month off 2 x C cells, you're needing to get all the capacity out of the battery pack you can, so direct drive is out of the question as the cells will drop below whatever the critical voltage is for the amount of light you need.

This means you need at least a boost regulator, but most of those are designed to drive high powered LEDs, they current regulate at several hundred milliamps or more while you are needing less than a dozen mA.

I'm sure there are driver boards out there somewhere but I don't know of any, and there's bound to be some DIY circuits if you web search for something like "white LED driver schematic".

If you went with more than 2 cells in your pack or used Li-Ion battery(s) instead so the nearly drained battery pack voltage was still above the minimum voltage your LEDs need, a simpler circuit again using a current limiting resistor as with the AC-DC supply could be possible, though efficiency would suffer if the voltage difference is large.

Given the amount of batteries you want to use and an entire month of runtime, you'll have to be pretty precise about it, a tight control of current level to make it run that long at the max brightness possible.

Supposing a roughly 4000mAH C-cell battery pack, and 720 hours in a 30 day month, 7 LEDs, you will end up with less than 0.8mA per LED which is going to be quite dim, dim enough that I suspect you will find it unworkable, that you simply must use more than 2 x C cell batteries. Even if you used only 2 LEDs that's still under 2.8mA per LED, significantly brighter than 0.8mA but still not very bright at all... consider that one typical 5mm LED isn't all that bright at a typical rating of somwhere between 15mA and 30mA (20mA most common, 30mA usually requires short leads and thick copper traces to 'sink heat away).

A month is a really long time to continuously power a light off of batteries unless you use more or higher capacity cells in the battery pack.

Considdering the math, I guess I´ll have to go with more than two cells... However that is not a problem, as long as I can fit it inside the candleholder... I guess four or five D-cells is possible... but as you say, a driver is needed. Problem is, I haven´t found a suitable one. I guess a seriesconnection is possible if I go for a higher number of cells, but I still need currentlimitation

If you use enough series batteries that they are higher than the forward voltage of the LED (or however many LEDs in series you have, their summed forward voltage drop), things get easier.

Take a look at the following DIY circuit, though you may need to put potentiometers in place of the fixed resistors it uses and experiment with resistance values to get the current down low enough that it runs for a month on whatever battery pack configuration you settle on.

http://superpositioned.com/2006/03/02/a-low-power-led-flashlight-with-pvc

(and web search for other white led driver schematic) The other option is just go to an electronics supply house like http://www.digikey.com and use their search function and selection menus to drill down to specifics for an LED regulator chip to DIY or a finished driver board (probably pretty expensive for what it is).

Here's a solution!

Get a GD Buck/Boost Converter Blank and choose the option to have Wayne solder in a .47 Ohm resistor for R1, and nothing for R2. Definitely check with Wayne first, but that should give you a GD 106. Wire up your LED's in parallel, and they'll get approx 15mA each. Run it on 3 alkies.

This setup should get you a nice balance of brightness, runtime, and LED life. You could even choose the .56 Ohm resistor to make a GD 90, which will give a little under 13mA to each LED.

The GD has a contact pad for battery +, so with a little modding ingenuity, it could be rigged directly into a Radioshack battery holder, making contact with the last battery in series. Either that, or just solder a wire onto the pad.

For seven 5mm LED's on a few alky batts, that's how I'd do it!

Bimmerboy said:

Here's a solution!

Get a GD Buck/Boost Converter Blank and choose the option to have Wayne solder in a .47 Ohm resistor for R1, and nothing for R2. Definitely check with Wayne first, but that should give you a GD 106. Wire up your LED's in parallel, and they'll get approx 15mA each. Run it on 3 alkies.

Click to expand...

Unless my math is off, even if the driver were 100% efficient it won't run more than a single-digit # of days off 3 alkalines with 15mA per LED. If it were ran only in the evenings it might run long enough....

Cheap Quick and Short term.

3C Alkaline cells, 7 X 100ohm resisters and 7 leds wired in parellel.

buy 18 white leds , 7 to fry accidentally, and a few spares.
get 10 X 100ohm resisters, just in case you lose a few.
put the 3 c cells in series, stacked , or get a Series holder at the electronics store.
solder each of 7 resisters to one of the legs of the led.
wire the leds and resisters all the same
hook the batteries in series making about 4.5v then parellel this with your led resister stack.

|--{~}---@---|
|--{~}---@---|
|--{~}---@---|
|--{~}---@---|
|--{~}---@---|
|--{~}---@---|
|--{~}---@---|
|+-{=]{=]{=]--|
(3 bats in series)

Runtime about 60 hours, till it dwindles down to very dim, and just keeps on going (or glowing) losses about 30%

------------------------------------------------------

if you have a simple meter

|--@---|
|--@---|
|--@---|
|--@---|---{/~}---- to battery negative
|--@---| Variable resister
|--@---|
|--@---|
|+{==]{==]{==]-

wire the leds simply in parelel, then use one single variable resister of 20+ ohms, preferably a good wire wound, so it lasts, probably you will find a nice 100ohm easily.
roll the variable resister to its highest resistance setting
put the probes for the voltage meter on one of the leds, and read the voltage, as you slowly turn the variable resister up, stop when it reads about 3.1-3.2v or untill the leds turn blue and smoke starts pouring out .
Or put a ampmeter between the last connection to the battery, and set the miliamps to about 105ma, Remembering that the Meters OWN resistance will exist, so when you pull the meter out and reconnect the battery you will slightly readjust to the same visual levels.
the Resister will be at about 15ohms with new alkalines

when it starts going dim, just visually recognise about how bright it is supposed to be and slowly turn it up, eventually when the cells are neer dead, you will be at the end of the resistance.

-------------------------------------------------------
after i look at trying to run it that long (168hours), an option of using rechargables really looks like a minimum (AC being a good idea again), or running them well below max spec, like at 5ma each or something. but then again my math is pretty bad.
http://led.linear1.org/led.wiz but you can check my math.

After you hook up the variable resister, which gives you cheap easy full control, you might find that as you bring it slowly up, you can stop with it looking more like a candle as opposed to a flashlight, and it will probably make it the whole 168 hours. i would just guess, feel out about where i want it, and let the nature of batteries do thier own thing.

notes: dont turn the resister "full Up" with full batteries, as then you will need the spare leds but then again that is what they are for

the "math" is based on WHITE leds, if your using Yellow or Amber Leds and not White ones, you could probably use just 2 Cells, and you would have to recalculate the resistance, but that would be a good idea. with yellow Leds and 2 cells the math stuff doesnt change that much, but it should be checked at least.

cheap quick candle #2
grab some plastic pipe from the hardware store, or tap plastics.
shove 2 AA bats into Each "candle" tube, wire up a resister to the led, run a wire down the tube to connect the battery negative, have the resister reach to the top of the battery, send it off to china to have them make 10,000 of them, and buy 7 of them for $1 at the doller store runtime at 5ma 15 days or so.

Scattergun said:

I´m looking to make a seven candle christmas decoration, but I´m looking for to make it run on a couple of C-cells... I want to use 5mm high power LEDs. But I don´t know anything about connecting seven of these to a few easily hidden batterys!! I´m looking for a long batterytime... preferably a month constantly on...
What is your suggestions for me to do?

Click to expand...

Can you still get the LM3909 IC?

This the classic micro-power LED driver; you could power all the LEDs from a single 1.5V cell for months at a time.

The secret to battey life is a flash rate above persistence of vision and a short duty cycle. This IC does the lot. My link also shows some of the circuits. Modern high-brightness LEDs open up a whole new world for the LM3909 - if you can find a source.


Edit - you can make your own with discrete transistors.

Scattergun said:

I´m looking to make a seven candle christmas decoration, but I´m looking for to make it run on a couple of C-cells... I want to use 5mm high power LEDs. But I don´t know anything about connecting seven of these to a few easily hidden batterys!! I´m looking for a long batterytime... preferably a month constantly on...
What is your suggestions for me to do?

Click to expand...

Check out this simple solution: 7 LEDs driven at 20mA from 1 NiMH cell.

J_C said:

Unless my math is off, even if the driver were 100% efficient it won't run more than a single-digit # of days off 3 alkalines with 15mA per LED. If it were ran only in the evenings it might run long enough....

Click to expand...

Not sure why I left that out. Off the cuff, without doing precise math, approx. 2.5 days continuous on three C batts? Nonetheless, I have some additional thoughts on the matter overall, as I think some of the suggestions could be missing the mark for this project.

Scattergun said:

I don´t know anything about connecting seven of these to a few easily hidden batterys!!

Click to expand...

Thus, the main reason for suggesting the GD. It's great to learn how to read a schematic, buy all the components, build your own circuit, experiment, tweak, etc. I'd like to finally get around to doing some of this myself. But considering the above statement, is it worthwhile for someone who just wants to make a Christmas decoration? Only Scattergun can answer that, but I personally wouldn't unless I was looking to make the project itself a learning experience in building switched circuits. Otherwise? Meh... too much work.

What say you, Scattergun?

As a side note... now, I could simply be confused here, but how is a boost driver with a single battery (either alkaline, or NiMH) going to power a seven LED decoration with useful light for any real length of time?

In narrowing down the choices, the GD blank is the only ready-made driver I can think of with customizable current at the point of purchase, making it almost a no-brainer as it can be perfectly tailored to suit seven parallel 5mm LED's with zero fuss. It'll work great on three alkies, offer good efficiency, plus it's a buck/boost, and should be able to suck all the useful capacity out of the batts. Using somewhat rough math again, assuming 10mA per emitter (using a .68 Ohm for R1), 3 D cells, and, say, 80% driver efficiency (likely higher according to spec), it should get a solid 100 hours in regulation, probably more... I think... I'm no EE by a long stretch.

Thoughts? Opinions?

datiLED said:

Check out this simple solution: 7 LEDs driven at 20mA from 1 NiMH cell.

Click to expand...

If you look at where they list that 20mA, it looks more like that is just the datasheet spec for the LEDs, not the actual drive current as it also includes "3.2V" which wouldn't be the total series voltage needed for 20mA through 7 of them.

That's a good thing though, if it put 20mA through 7 then we're back to only a few days of runtime. I think lctorana has the right idea about rapid cycling, but unless I looked at it wrong the LM3909 circuit seems to take significantly longer to (re)charge the doubling cap than the allowable on-time duty cycle, I don't see how the circuit can meet the goal unless the LEDs flicker a lot, to have high enough output that persistence of vision is enough, the average current would be too high to run months from a single cell and have it appear both constantly on instead of blinking, and bright enough to notice like a simulation of a small candle.

That might be good, a little flickering is what candles do, but how much is hard to say. It seems that circuit would come closer to the idea with pretty big (like D cell) and two of them to approach the max 3.0V rating the discrete version is rated up to.

As for the GD emitter, I wonder if it's possible to get it unpopulated and add your own resistors (probably someplace like Mouser, Newark, Allied Electronics would carry them) to lower the current even more. while fractions of a mA average current is probably too low, 10mA might be more than needed if enclosed in something semi-translucent.

Wait a minute, I have a solution. Switch the batteries when they run down. Like we do with any other battery powered thing.

I still think Ictoranas solution is the most appealing, since it would give me the runtime I need. Any source for that driver??:thinking:

Scattergun said:

I still think Ictoranas solution is the most appealing, since it would give me the runtime I need. Any source for that driver??:thinking:

Click to expand...

the LM3909?
That IC has been discontinued for over a year now :shrug:

I'd look for ideas here as well: https://www.candlepowerforums.com/threads/225483
http://www.educypedia.be/electronics/circuitsled.htm

Would LM555 be a suitable replacement?

Scattergun said:

I still think LCTorana's solution is the most appealing, since it would give me the runtime I need. Any source for that driver??:thinking:

Click to expand...

Build your own?