# Basic question - Led power supply

#### taj1987

Jun 14, 2011
5
Hello,

I am at a dead end and could really do with some help.

I need to power Luxeon LEDs (forward current 700 mA, forward voltage 6.84 V) for an experiment but have absolutely no idea how to power them. I have acces to universal power sources, can I use one of these and attach it with crocodile clips?

The LED can be found at http://uk.farnell.com/lumileds/lxhl-pm02/led-luxeon-v-green/dp/1106605?Ntt=lxhl-pm02

I have found LED power supplies such as http://www.wydels.co.uk/Product-Cat...ivers/Luxeon-LED-Power-Supply/PLDCC-350-12-24 but have no idea whatsoever if this would do the job.

I apologise for my lack of knowledge but have contacted all of the suppliers of the LED and power supply that I could find but have had no help

#### poor mystic

Apr 8, 2011
1,074
Hi Taj
your led's have an equivalent resistance of nearly 10 Ohms at the operating point, and if you put another 10 Ohm resistance in series with each led, and supply the circuit with a little less than 14V, the current should look after itself.
Since you have an adjustable supply, you can raise or lower the supply voltage until you get just the required current flowing through the led.
Notice though that the 10 Ohm resistors will get very hot. Don't burn your fingers! They'll consume 5 Watts of power in the circuit so I'd use resistors capable of handling 10 Watts.

#### taj1987

Jun 14, 2011
5
That's great, thanks for your help.

I have another question. Do all LEDs need resistors, such as mounted LEDs?

#### poor mystic

Apr 8, 2011
1,074
I guess you could voltage control an led but I'd feel insecure with that arrangement because as the device heats up I expect that the applied voltage would drive more current into it.
So a resistor is always used and the voltage is left to float around if it wants to.
I'm not sure what you perceive as being different about the operating requirements of 'mounted' led's; as far as I know they're all the same kind of cat.

Last edited:

#### poor mystic

Apr 8, 2011
1,074
... the thing is that every unit of current can be expected to produce a unit of light, which is the germane consideration here.
Analysis of power dissipated by the device will reveal that the relatively small variation in power observed in the led due to voltage variations when driven by a resistive supply is a lot more acceptable than the variation in power due to current increase when the led is driven at constant V.
So we choose the resistive supply in preference to the constant voltage supply both to maintain an even rate of illumination and to protect the device.
This question can be further studied, should you be so minded, in the manufacturer's data sheet which will allow you to discover the current/voltage/temperature proerties of the device.
Or perhaps not, it is well-known among electronics designers that led's need to be current-driven and perhaps the manufacturer won't bother with the useless information.

#### BobK

Jan 5, 2010
7,682
The best way to drive an LED is with a current controlled power supply. There are LED drivers available that do this. Try dealextreme.com for very cheap drivers. You do have to wait about 2-3 weeks to get anything from them though.

Bob

#### poor mystic

Apr 8, 2011
1,074
BobK really is right on the money with tthat answer, but if I got this right Taj is conducting an experiment of some kind here which does not depend very much on efficient use of supplied power.
So in order to get his light up-and-running quickly and get on with the experiment I chose the high-voltage-plus-resistance option.
Perhaps I am exactly wrong! Others on this forum have noticed my fallability and I am in no position to adopt a pontific attitude.
Just as well, it was getting to be a bit of a strain being right all the time...

#### taj1987

Jun 14, 2011
5
But I would still need a resistor if I used a current controlled power supply, is this right?

#### poor mystic

Apr 8, 2011
1,074
No!
if you control current you control the led. Current control is what they're all about. So if you do have a settable current supply use it by all means... no resistor will be required.

#### taj1987

Jun 14, 2011
5
Ah right. So I could connect the current controller straight to the LED, set it to 700 mA I assume, am I correct or is this another obvious error on my part?

Thank you so much poor mystic, and Bob for all of your help. I'll see what equipment I can get my hands on.

#### taj1987

Jun 14, 2011
5
And can I use this LED like the type with the 2 long pins (one slightly shorter than the other as I have discovered on this journey)? I mean can I attach one clip to one side of the LED?

#### poor mystic

Apr 8, 2011
1,074
Hi again Taj
Yes. According to my reading of the data sheet 700mA is the correct current for the device, so if you supply that current to the led directly from your power supply the led will run safely.

Jun 10, 2011
443
Look for any lab-quality DC power supply that has the requisite voltage and current and has a constant current mode. I've got a number of such things on my bench and I find I use constant current mode often (maybe even more than constant voltage). My favorite power supply is my HP E3615A, as it has 10-turn pots for both current and voltage, so I can quickly adjust them to what I want. With such a supply, there's no need for a resistor in series with an LED. For example, a short while ago I needed to measure the optical spectrum of a red LED, so I just connected it to the HP supply, dialed the current setting to be 20 mA, connected the LED and did my work. Simple -- and there was no way I'd damage the LED.

By the way, for you folks who prototype things, a constant current supply is quite handy. At first turn-on, set the current to zero and set the voltage knob to the supply voltage you want to run the circuit at. Then slowly increase the current knob and watch the current. You can feel chips to see if they're getting hot; if not, then increase the current. The compliance voltage also tells you something -- if you see it stuck near zero as you're turning up the current knob, you may be looking at a short. Power things down and do some continuity checking.

Finally, most lab supplies have an overvoltage protection or a crowbar circuit. On the E3615A, there's a screwdriver adjust on the front panel that lets you set the maximum voltage the supply will output. If you accidentally turn the voltage adjust knob too far, there's an SCR which will "crowbar" the output to zero and the supply shuts its output off (you have to power cycle it to turn the output back on). This may save your bacon some day, especially if you're a klutz like me who accidentally bumps the wrong knob sometimes.

#### poor mystic

Apr 8, 2011
1,074
As for how you clip the leads to the device you need to download the technical datasheet from Farnell I suppose, since there doesn't seem to be any clear indication of the package lead arrangements in the first page of the product decription.

#### poor mystic

Apr 8, 2011
1,074
Daddles knows what it is to work in a civilised environment, which I am forgetting...
Of course everything he says is true, and even amateur electronixists do better with access to good current-controlled equipment such as his HP supply. What a lot of hassle he is saved by access to such equipment.

Jun 10, 2011
443
poor mystic, your post reminded me what it was like prior to the times where I had decent lab equipment at my desk. I apologize if my post sounded a bit hoity-toity, as I had to make do for quite a while with a Heathkit power supply I built in the 70's. I spent quite a bit of time (years) searching for lab-quality stuff that was amenable to my home hobbyist's budget (I found that HP power supply on ebay for $100US delivered and it was in nearly new condition). Many of us hobbyists have to make do with less. That said, if you're going to fool around with electronics, either as a hobbyist or professionally, having one or two lab-quality DC power supplies at your bench will save you an enormous amount of time. The other one I use a lot is a B&K 9130, a nearly-ideal power supply for an engineer's bench. Quite versatile, but also a pretty penny (I got mine in a trade at a good discount and I didn't have to come up with any cash). Also, with a wall wart, an op amp, some resistors and a 10-turn pot, and a MOSFET, it's not hard to make a lovely current source for yourself. I've never used one, but I suspect a nice power supply for bench use would be the B&K$139US model 1550. I'm going to get one for my son-in-law. Because it's a switching supply, note the lowest output voltage is 1 volt.

Last edited:

#### (*steve*)

##### ¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,505
It is important to mention that the LED may need a heatsink. I didn't read enough to know if it is mounted on a heatsink.

Replies
1
Views
804
Replies
1
Views
809
T
Replies
4
Views
985
T
Replies
4
Views
1K