LED Power Supply

Thread Starter

Garoad

Joined Apr 24, 2012
26
I'm just about done a little battery powered LED "christmas light" project, which requires about a 12V power supply.

I know battery voltage drops a bit after some use long before they're considered "dead", so I was thinking of using a 12V regulator like a LM7812, and 10 D cells in serial instead of 8 D cells in serial. My thinking being that this should extend the "maximum brightness" period of the lights without burning them out with an excessive 15V.

Good idea, bad idea, or overkill? The light set is only 20 3mm LEDs which slowly change color (on their own), 4 parallel strings of 5 LEDs per string (w/4 resistors). My guess is that the LM might burn a little extra current, so it might be a little self defeating, but not sure whether that's significant.
 

wayneh

Joined Sep 9, 2010
17,496
Are your D cells rechargeable? Most such are really just AA batteries in a larger package. You can save space and cost by just using AAs. Just sayin'

The regulator is a waste of time, IMHO. It alone will require an extra battery in series to cover the dropout voltage, and can only reduce current by burning off some itself.

A better approach overall is to use a joule-thief circuit like the one in all those garden lights you see. That allows the circuit to function at a more consistent current (brightness) even as the battery voltage fades.
 

Thread Starter

Garoad

Joined Apr 24, 2012
26
Someone here (I think Bill) mentioned before that a joule thief circuit is pretty inefficient (I had asked about it before), so I opted not to go that route a while back (although it's still an option).

I wasn't planning on using rechargeable D's (too expensive!). I had about 10 leds on a few C's last year and that lasted about a month (8hrs/day weekdays), so I'm guessing a much larger number of D's should last a whole season. These are what I have.

Those garden lights (I opened one up before) only seem to have a tiny inductor in there that I could see (wasn't able to figure out H value with any certainty). You think I should just add one of those to the overall circuit? (How many henries? I have a bunch of inductors sitting around, mostly those small resistor looking types, a few different H values.)
 

wayneh

Joined Sep 9, 2010
17,496
The solar circuits are meant for only 1 AA, so I wouldn't expect them to last at 12v or to handle the current you need. But a scaled up version of that circuit could work. It may not be very efficient, and I suppose it's a foot race whether you win going with higher efficiency or by extracting more out of the batteries. The principal advantage of the circuit in those solar lights is being able to use a single AA, reducing cost.

Using the 7812 regulator with 10 or more cells does achieve one thing - you can hold brightness for an extended period. At the higher starting voltage, you could have extra LEDs instead of the regulator, so the regulator adds a big loss but it will work. The loss continues even when the regulator cannot hold voltage anymore due to battery voltage sag- the output voltage will still be less than the input voltage, meaning it's burning power, and you'd be better off without the regulator at that point.
 

Thread Starter

Garoad

Joined Apr 24, 2012
26
I suppose I could get really fancy and design a circuit that "turns off" the regulator once the voltage drops. Then turn on a joule thief like circuit once the voltage drops very low... hehe, but this is sounding more like a science experiment. I guess I'll wait to see if anyone else weighs in with other ideas.


I never heard that before. Are the Ah ratings the same?
I've heard it (but figured this was cheap ones), but sources like "wiki" make the claim that D cells are higher mAh.
 

tracecom

Joined Apr 16, 2010
3,944
I've heard it (but figured this was cheap ones), but sources like "wiki" make the claim that D cells are higher mAh.
Duracell make the following service claims for their alkaline CopperTop batteries with a 10 Ω load.

AAA 10 hours
AA 22 hours
C 65 hours
D 145 hours
 

bountyhunter

Joined Sep 7, 2009
2,512
I'm just about done a little battery powered LED "christmas light" project, which requires about a 12V power supply.

I know battery voltage drops a bit after some use long before they're considered "dead", so I was thinking of using a 12V regulator like a LM7812, and 10 D cells in serial instead of 8 D cells in serial. My thinking being that this should extend the "maximum brightness" period of the lights without burning them out with an excessive 15V.

Good idea, bad idea, or overkill?
A 7812 reg needs about 14V input to regulate which means over most of the battery's life, the 7812 will just be a 2V waste of voltage to the powered circuit.

If the LEDs have series resistor, the current should not vary much with the battery dropping from 15V to 11V.
 

Thread Starter

Garoad

Joined Apr 24, 2012
26
Yeah I should probably just go with the original plan of 8 D cells (already have the holders), 12+V, get this one done and move on to a more elaborate project with a 555 or something.

This is the current circuit:

(I'm 99% sure I'm recalling the actual resistor ohm ratings.)

But the LEDs are these cool 3mm slow color change lights I found on ebay. Predictably they get a little dim when they turn blue, which is why I started worrying about keeping the voltage as close to 12V as possible. In a lit room they'll probably appear to go dim when they turn blue, once the batteries are weak, but ah well.

I'm actually pushing it putting 5 of these in serial, because there's a big noticeable difference if I do fewer. With 5 like this, they do a sparkle/blink thing (which I thought actually looks cooler), vs a much smoother, gentle change if I do one resistor per led. That also makes me concerned about the possible effect of voltage drop. It's hard to describe.
 

bertus

Joined Apr 5, 2008
22,270
Hello,

Do you have the specs of the leds?
Color changing leds do have a chip inside.
They are probably to be used on a fixed voltage.
The current may change during the color change.

Bertus
 

wayneh

Joined Sep 9, 2010
17,496
I never heard that before. Are the Ah ratings the same?
Frequently, yes, but it varies. AAs are such a commodity that a D cell can be made much more inexpensively if you "cheat" and just use a AA in wolf's clothing.

The rechargeable Energizer D cells at Walmart are 2200 mAh. That's less than the AAs I buy.

But I just did a quick search and see there are many, many D-cell options at 10-12,000mAh. That's much more than any AA, so there is such a thing as a "real" rechargeable D.
 

Wendy

Joined Mar 24, 2008
23,415
Where did you get those LEDs? They sound interesting.

I'm spread thin, but I do honor requests to review threads.
 

Thread Starter

Garoad

Joined Apr 24, 2012
26
I don't have much on specs, these came from Ebay (Hong Kong!).

Color-changing LEDs are not supposed to be connected in series. The "sparkle/blink" you see is them being destroyed.
Really! That means this has turned into an experiment after all I guess (how long will they last...) I'll probably have to go back to the drawing board then, for these color changing lights, because I was trying to minimize use of resistors & wiring.


Bill - Go to ebay and search for "color change led", select "Business & Industrial" as the category, search, and then scroll down a bit. That should get you there. They're hard to find by just a general text search. It takes a few weeks for them to arrive in the US, though.

I think these are the type I bought a while back.

Looks like this guy (name is "hktaiyuen") has 5mm LEDs of the same type now, too.
Ebay Search link
 
Last edited:

Audioguru

Joined Dec 20, 2007
11,248
Some of my solar garden lights have color-changing LEDs. They work from a voltage so they must have a current-limiting circuit built-in.
 
Top