LED's as test load?

I don't have any 1N400x diodes on hand... I do have several white LED's though. Now I'm not sure what the forward voltage curves for the LED's are exactly, but when I ran 20mA through one it was at about 3V.

3V is about where an LPC-815 tops out according to charts I've seen...

would it make sense at all for me to use an array of 5 of these LED's in parallel (and one 1Ohm resistor in series) as my test load?

also, suppose I did have the regular diodes. How do I figure out how many of them to use? The LD's Vf ranges from about 1.9 to 3.2. I'm aiming for whatever point I expect the diode's voltage drop to be at?

The problem with LED's is that they can only take about 20ma and they will cook, and if you put 5 of them in paralelle the forward voltage drop goes way down.you stated the forward voltage drop of a laser diode is 1.9 to 3.2 it closer to 2.7 and as high as 5.7 for the 445 diodes.
so you want to know how many regular diodes to use for a test load well there forward voltage drop must add up to the voltage of the diode you are setting the driver for.
example if you laser diode has a forward voltage of 3.0 volts and each 1n54xx diode has a forward voltage drop of .5 volts @ the current you want then you will need 6 1n54xx diodes in series to add up to that 3.0 volts for your laser diode.
did I say that right some one else step in if I messed that up...

arcaneinsomniac said:

I don't have any 1N400x diodes on hand... I do have several white LED's though. Now I'm not sure what the forward voltage curves for the LED's are exactly, but when I ran 20mA through one it was at about 3V.

3V is about where an LPC-815 tops out according to charts I've seen...

would it make sense at all for me to use an array of 5 of these LED's in parallel (and one 1Ohm resistor in series) as my test load?

also, suppose I did have the regular diodes. How do I figure out how many of them to use? The LD's Vf ranges from about 1.9 to 3.2. I'm aiming for whatever point I expect the diode's voltage drop to be at?

Click to expand...

Okay that makes sense, but i mean where along that voltage curve for the LD should I set the test load for? Suppose I had some hypothetical LD which starts out lasing at 2V and goes up to 5V at its highest output. I reckon that I'd need to refer to its chart or find out in some way what its voltage ought to be at the current that I intend to pump through it.

and I see now the LEDs isnt a terribly good idea after all. Putting them in parallel was to get them to share the current load but I don't think that's what would actually happen (not without putting a resistor with each one and it's too cumbersome at that point).

I guess i'll just go to radio shack.

You've got the right idea about looking at voltage curves. The thing is, you'll never really see that much of a range per LD.

658 LOCs (LPC-815, 826, etc) tend to run from around 3.2 to 3.8 if you really push tem.
405 tends to run up near 6V
445 tends to run around 4.5 and up perhaps as high as 5V, but I think typically more like 4.75V

It's not quite as straight forward as just dividing the LD voltage by the voltage drop expected from your diodes. Why? Because the diodes themselves will also have a voltage curve. So at 400mA (what you might be testing a 658 at) your diodes will each drop less voltage than at 1600mA (what you might be testing a 445 at)

Safe diode counts:
658 - 4 diodes, regardless of the current you're testing (within reason, up to perhaps 600mA)
445 - 5 diodes, as long as you're testing them at reasonably high current. However, if you were to test a 445 at more like 250mA, then you'd probably need 6 diodes to hit the corresponding 445 Vf at that current. (because each diode would be dropping more like 0.7V instead of closer to 1V)
405 - 6 diodes, pretty much all the time

* Side-note. I'm breaking from accepted / tested wisdom here, so ignore this comment OP. But I think in theory, if you were to be simulating a 12x 405nm running at like 200mA, you might actually need 7 diodes in a test load.

arcaneinsomniac;

I have used high-power white LEDs in my test loads (about 3.3VDC, like a Red).

I have added diodes (standard & Schottky) to get the final voltage I needed for BluRay lasers.

LarryDFW

Yea I second Larry's post. I use Cree LED's all the time as test loads. My XP-G's combined with some rectifier diodes in series is usually how i set my drivers. In fact, I have 4 testloads from R2's R5's Q3's where i solder on additional diodes to get 3.2V (no extra diodes), 4.5V, 5.4V, and 6.1V. all those LED's can handle brief instances of over 1.2A, i got a MC-E LED that is soldered to 1 1n4007 diodes to acheive 4.7V and that one can handle up to over 2A.

why all the high power LED's as testloads? isnt that a waste? NOPE...because lasers are cool and worth it + XM-L's and P7's are better, which ended up in those flashlights...what else am i supposed to do with those crappy 200 lumen LEDs????

steviedezie said:

crappy 200 lumen LEDs

Click to expand...

damn kids these days.... :spank:

BShanahan14rulz said:

damn kids these days.... :spank:

Click to expand...

They're crappy only by comparison I guess. No sense leaving them sitting around not getting anything done. Pretty thoughtful really.

Since my LED's are actual crappy ones that will fry above 100mA (It turns various colors. Then there's a POP as it goes), I probably shouldn't bother with them. As I understand it, diodes have an exponential I/V curve. If I wire up several, and for whatever reason their properties do not quite match up, all (most?) the current will flow through the weakest link and fry it. But wouldn't that then open the circuit at that point, and the current automatically redistributes to the rest of the LEDs? This sounds like a cool experiment for me to try out.

Is it even possible to calculate how the current flows through a couple of diodes wired in parallel? When I was doing this for class we would operate on the assumption that a diode was either ON or OFF. But the actual behavior is quite a bit more complicated.

BShanahan14rulz said:

damn kids these days.... :spank:

Click to expand...

hehehehe :eg: :eg:...once you go p7 and XM-L's you cant go back... even my MC-E which used to be king is collecting dust.

EDIT: oh besides, im not THAT young... i remember when incandescant flashlights were all we had...xenon was the XM-L of its day

how come you haven't trashed your P7's yet? Those are junk, by your standards, by now....

And I'll trade you 5x1n4004 (~3.5V) for an XP-G (3.3V?)
You'd be getting a better deal, more volts! ;-)

P7's and a XP-G is hardly the same... 500-700 lumens VS 200 lumens is a big difference my friend...

also, if a XM-L and P7 can be bought for ~$14 + $6 driver, why go Q2.3.5 or R2.4 etc etc for $6-$10

I used to use a 1W cree LED in line with 1N4001 diodes for my
test load, but only to see any flicker etc in the driver output.

IgorT had mentioned this to me via PM which is a good idea
to detect any anomalies in a boost driver output..

I then got a DSO scope to test the drivers and have since stopped using
an LED in the test load, but still a good idea for those who test
their drivers with a DMM...

As long as you are building proper current sources, power leds make fine testloads. I've used mediocre efficiency leds out of cheap flashlights often enough to do a quick test. The optical feedback is nice, and can give you a hit of power-on and -off peak behaviour if you don't have a scope to properly measure it.

I agree with Benm...when testing drivers I always recommend for people to buy a regulated power supply. Hell i got mine brand new from an electronics shop for $90. 0-3A and 0-15V full adjustable. can be set to act as a CC or CV source. Has saved me from frying lots of drivers and diodes. one time i hooked it up to a phr directly...no driver, set it as constant current mode... cranked it up to 5.1V and sure enough the damn thing lased...and kept lasing. it was only in an aixiz module so after a minute it got pretty warm. still good to know how stable this PSU is...it powered a phr directly and didnt kill it. still used it in a build that lasts even today.

moral of the story...get a good power supply and then test drivers/diodes...batteries as power sources just dont cut it sometimes.

steviedezie said:

get a good power supply

Click to expand...

I hear ya. I bet somethin like this ought to do the job

I definitely like the idea of being able to hook a current regulated power supply directly to a diode to check it out. And it would prove useful for any number of other tasks.

bingo...thats a great PSU. I have this one here. similiar to the one you linked to except that it does only 15V. same price. I just had the Velleman locally and bought it locally instead of waiting for shipping.