Cooling is the key for this part. It can do the current, but you have to suck the heat away.

I've been doing a bit of research on power LEDs lately, I'll see what I can dig up when I'm in front of my computer.

 

cool, thanks.
I used to build a lot of computers, so I got an aluminium heatsink off an old motherboard:
http://i760.photobucket.com/albums/xx243/i84cookie/heatsink.jpg

I have some contact paste too.

The heat sink needs to be fairly small to fit into the projector. Plus the projector has a fan, I think it should be suitable. I'll be using the LED for up to 10 min at a time, roughly.

 

Looks like a good plan overall.

This thread I'm trying to figure out how to do it home made style. Post #4 of that thread has some interesting links.

I've also come across this, which is probably more what you're looking for. They might have retailers in UK.

http://www.national.com/analog/led

This thread is kinda long, but we did some parts research for the USA. You might be able to springboard off it.

http://forum.allaboutcircuits.com/showthread.php?t=29749

And another link on AAC.

http://forum.allaboutcircuits.com/showthread.php?t=30192

As you can see you're not alone, the number of applications seems to be growning too.

Someone else sent me another forum link you might find handy.

http://www.candlepowerforums.com/vb/portal/index.php

 

thanks bill useful stuff. I'm getting a database error from the candle power forums, maybe they're having technical difficulties...

I looked through some of the popular LED brands and there doesn't seem to be one that matches the 1000lms that my one promises (or so it says), so I'm still keen to get this noname brand LED going.

Ok so if I get this 18v 2.22a power supply, then buy two 700ma buckpuck's, connect them in parallel so it gives me a total of 1400ma, which is perfect, what kind of resistors to I need to use for this circuit?

 

No! It doesn't work that way, you have to power devices with their outputs connected together and it will fry one or both.

Like I said earlier, use the LM317 as a temporary solution, and take the time to get a better one. The LED doesn't care where the amps come from

 

oops, ok. I read in another forum where someone was looking for a 1500ma buckpuck, and someone replied saying to put two 700ma ones together!

Alright I'll get the LM317 and it's little heat sink.

To complete the circuit do I need a resistor?

 

Yes, to "program" the LM317. The 317 will also need heatsinked, since it will get hot, use a different heatsink.

Use qty 12 10Ω ¼W resistors in parallel to set 1.5A of current. The resistors in this configuration are equivalent to 0.83Ω at 3W. 10Ω ¼W are a lot easier to get, and cheaper to boot. You can substitute with higher wattage resistors, but not lower wattage units.

You will be running this regulator at it's absolute max, so heatsinking is critical. Use the full works, including thermal paste. While this part is rated for 1.5A, failure rates for parts go dramatically up when you hit this knee. It has lots of built in protections though, so while it may burn out it won't take the lamp with it.

 

Without a detailed datasheet you are just guessing. Its max voltage rating is 15V. Its max current rating is 1.8A. Then its max power will be 27W which is much more than its 20W rating.

Maybe it can dissipate 20W when its case is cooled to 25 degrees C with liquid nitrogen or something that is much better than a huge heatsink. A big high velocity fan?

 

Actually I did pull up the datasheet, Fairchild says in excess of 1.5A, but I didn't see a definitive number.

You will note I'm advocating it as a temporary measure until better can be bought or acquired.

 

Audioguru was referring to the rated power dissipation of the LED itself.

It's ratings were stated as 20W.

So, that means both the voltage drop across the LED and the current through it needs to be monitored during experimentation to obtain the power dissipation, to ensure that the supplied current doesn't exceed it's ratings.

Unless one likes to watch expensive LEDs pop.

 

Yeah, but the OP showed his planned arrangement, which I think will work. A CPU heatsink properly modified should work fine, especially with the fan present. The LM317 is pretty predictable, you think it needs monitored?

 

Well, if he's going to be running around 1400mA through it from an 18v supply to a 14v LED, it'll get mighty toasty. That's around 5.6 Watts.

The LM317 really isn't that much of a concern though; they're relatively cheap. It's the LED that I'm concerned about.

We don't know for certain what it's Vf will be at a given current. But it seems to me that it has a package rating of 20W.

If our OP runs 1400mA through it and it has a Vf of 12v, that's only 16.8 Watts - well within specifications (providing the heat sink is big enough).

However, if he runs 1400mA through it, and it winds up having a Vf of 15v, that's 21W, which is over the rated specification.

 

Actually, I can't get that 18v 2.2a psu, so I'm going to go with this one:
http://cpc.farnell.com/stontronics/3a-301wp15/psu-international-15v-2a/dp/PW02465

slightly less powerful at 15v 2a, maybe it'll be less risk?

Thanks for the diagram Bill. So twelve 10Ω ¼W resistors in a parallel circuit. That's cool, will these do? http://des-parts.co.uk/index.php?ma...id=110&zenid=d6e6b9e64ca207ef8d3286ebd7f70a58

A couple of questions, Vin and Vout, is in - and out +?

Also, how do you recommend constructing this circuit, on a breadboard or loose wires taped together...?

 

I've been known to just solder everything together and seal it with epoxy putty, but in this case it has to be heatsinked. There really isn't a choice.

The resistors were exactly what I was thinking about. You can tweak the current by adding or subtracting a resistor, each resistor is 0.125A. Qty 12 X 0.125 is 1.5A.

Wookie brings up some legitimate concerns, you may want to start smaller, say 1.25A, and measure the LED Vf (how much voltage the LED actually drops). I highly respect Wookie's opinions.

You will need at least 18VDC for the power supply, since the lamp is rated for 15V Vf max, to guarantee it will work. I don't think that power supply will work. As long as the power supply has enough current (2.0A is good, so is 3.5A, I wouldn't use 1.5A) it doesn't matter.

I understand why you want to drive this unit to its max. Just remember that there is a penalty in reliability and lifespan. Having said that, I'd do pretty much the same thing.

 

ok, I've ordered everything I need, well I guess I might need to buy some wire. Plus nuts and bolts to attach the heat sinks.
Thanks for all your help guys. I'll let you know how i get on.
I'll probably get back on here when everything arrives, I may need advice with the circuit construction.

I found a better power supply, I can't believe I didn't think of checking ebay before! Its 18v 2a, exactly what I need, i guess.

 

I was looking at wire, will this stuff be suitable?:
http://www.maplin.co.uk/Module.aspx?ModuleNo=6198

I wanted to ask, if I was going to test this circuit, will it be safe to briefly turn on the LED without the heat sinks attached?

 

I wouldn't. 15 seconds could fry them, you might get by with a very few seconds, but is it worth the risk? It is your call though. I think you'll be surprised how fast they heat up, and how slow they cool down.

 

ok, i won't do that then. What about that wire?

 

Not sure. Seems a bit light, but the insulation is probably more important. I'm wondering whether it will melt or not. What do you think your maximum length will be, and the total length carrying over an amp?

 

I can't imagine the total length being more than a foot or a little more. There'll be over an amp through the whole thing, and 2 amps before the regulator. But the spec on this wire says it can handle up to 6 amps. It also says maximum working voltage is 1000V rms, what does that mean?