LED flying saucer

[lbenson]

I recently got from Dealextreme several of the "flying saucer" LED lights shown in the pictures. These are for camping (in my Oz-made Kamperoo tent trailer). They run off of 4 AAs. I would like to put a DPDT switch in them so that I can switch to 5 volts regulated down from a 12-volt auto battery (actually a separate non-lead sealed 18AH). I am concerned about burning them out, since there is not a resistor in sight (and I'm fairly sure that there are not resistors contained in the switch, tho there could be).

The switch has 3 "on" positions--outer ring, inner ring, both rings. The LEDs are wired in parallel. Since there appear to be no resistors, can it be that the device relies upon the discharge rate of the battery to keep from burning out the LEDs? I unsoldered and removed one of the 3-LED inner PCB strips, and hooked it up through a 220R--it seemed not as bright as on the device. I then hooked it up without the resistor--the lights were much brighter, and within 5 seconds were getting warmish, so I assume the LEDs are not resistorized.

There are 24 LEDs in the outer ring and 12 in the inner. Sorry, Dippy, no data sheets are available. How can I safely wire up the 12-volt battery, regulated down to say 5 volts with, say, a LM2940CT-5 1-amp regulator (beefy enough? max 36*20ma = 720ma)?

 

[Dippy]

Oh dear. No Data Sheets as that would add 5 pence to the cost. Can you phone the factory in China?
Without it sitting in front of me I haven't got a clue.

As it's not a PICAXE question I would briefly suggest that a simple regulator from 12V to 5V may not be good. I would use a regulator and then use an 08M to provide a duty cycled supply to the thing to be safe. Then you could all sorts of effects too, especially with a bit of hacking. Flashier than just using a 555. Good luck.

Not that I would want to disturb the quality soldering...

 

[jodicalhon]

What might the internal resistance of the batteries be?

(V unloaded - V loaded)/current should give something close, I imagine.

Just a thought...

 

[Tom2000]

This is the second mention I've read lately discussing commercial LED light fixtures that don't use current limiting resistors. (The first was in sci.electronics.basics, and referred to an Osram Dot It Light.)

With just three LEDs, I surmised that there was a slim chance that, with Sylvania's huge purchasing power, the parallel-connected LEDs might have been very closely matched. Unlikely, but nothing else came to mind. (And the 3-AAA power supply was another problem...)

With the number of LEDs in your fixture, it's highly unlikely that the manufacturer could rely on matching. Yet, no external resistors in sight. QED: the LEDs must incorporate internal limiting resistors, similar to the 5 volt red LEDs that Peter Anderson sells.

My supposition has nothing to do with your problem. I just find all this very odd.

Back to your problem...

Perhaps you could use a 6 volt regulator, such as a 7806, and switch directly between the 4-AA battery pack and the regulator output. DPDT, as you mentioned. One pole selects the LED power source, the other pole switches +12 to the regulator.

Good luck!

Tom

Edited to add: if you only have a 5 volt regulator available, you can raise its output voltage by placing a couple of diodes between the regulator's ground terminal and circuit ground to lift the whole package above ground potential. Two silicon diodes will give you about 6.2 volts output from a 5 volt regulator.

 

[kevrus]

How about using an LM317 (variable voltage regulator) configured as an adjustable current regulator which could be set to a safe current level, maybe measure the current drian from the AA batts to use as a guide.

 

[lbenson]

Tom--I've used PHAnderson's resistorized LEDs by the dozen, so that was a thought. But my experiment of trying the disconnected strip and having the LEDs get warm in 5 seconds made me think that they were not resistorized.

Kevrus, Jo_C--good ideas about the battery current--the LM317 might be able to emulate the batteries.

Dippy--oops, sorry, in my haste I forgot to mention the picaxe08M which is supposed to turn the lights off and on and dim them using the mini TV-style remote.

 

[lbenson]

Here is what I did for a first pass. Running off of 4 AAs my amp meter showed 186ma on the ring of 24 LEDs, 218ma for the ring of 12, and 218ma for both rings. The latter reading doesn't make sense to me, but they all put me in the ball park. For the 12 it comes to 18.17ma per LED. If they are 25ma LEDs (who knows), then the 12 could take 300ma. I wanted to try my LM2940CT-5 regulator, so if I wanted to provide 300ma at 5 volts that would require 16.67 ohms of resistance. I didn't have that, so I used two 33 ohms in parallel. That did not look as bright as the other, so I experimented until I found that parallel 12.2-ohm resistors gave pretty nearly the same light as the battery-powered light.

Thanks to Dippy and others, I have a current-limiting bench power supply, tho not a Thurlby Thandar. I started it at 5 volts and 300ma. I cranked the volts up to 12Volts, and the current rose to 340ma (at around 5.2 volts) and never got higher. The LEDs stayed cool, or very slightly above ambient; the resistors got warm--not hot (only 1/4 watt--I must increase) and the regulator got hot. I put a heat sink on it and it still got hot, not too hot to touch, but too hot to hold.

So it looks like this option would work with appropriate heat sinking. I found a nice site on using LM317s with LEDs, but it suggested that the main reason for that was to replace resistors alone in a situation where a battery might drop and so dim the lights. I don't understand how this could substantially differ from the setup I have made.

I am merely a hobbiest, so none of this stuff comes naturally to me. Is my math right? If anyone can suggest a better way to do this--run this thing off of a 12-volt battery--I will be happy to consider it, and try it if within my constraints of time and parts. Did somebody mention a Picaxe 08M switching power supply?