CREE LED, current limited?

[mayor]

Hello,

I'm using a current-limiting circuit to test out a CREE CXA1304 LED. It's max rated current is rated at 1A.

I've tested the current-limiting curcuit, set to 400mA, with a resistor as a load. Works.

BUT! When connecting the LED, I get a max of under 200mA, as though the LED had a high enough internal resistance.

For people experienced with such LEDs, does this sound like a defect (or a zapped LED, though I was careful with ESD band and all), or does the LED need to warm up to its 85C operating temp? Anyway, it seems odd to a power LED newb.

[The Soulman]

What was the voltage across the resistor and how how much is it across the led?
Perhaps your current source is hitting its voltage limit.

[Belrmar]

For people experienced with such LEDs, does this sound like a defect (or a zapped LED, though I was careful with ESD band and all), or does the LED need to warm up to its 85C operating temp? Anyway, it seems odd to a power LED newb.

LED's as all diodes , when they get warmer their foward voltage decreases, so it may be a problem with the supply that can't get enough voltage to push 400ma acros your LED

[floobydust]

Cree CXA1304 LED come in 9,18,36V class.
Your power supply may not have enough compliance voltage; it does not have enough voltage to drive the LED at full current.
Depending on the LED voltage, a dummy load resistor would be like 18V/400mA= 45R.

[Kalvin]

Increase the voltage until you get 400 mA current through LED.

Ps. the LED will get warm or hot quite fast without proper heat sinking, so remember to check the LED's temperature but don't burn your finger.

[Whales]

as though the LED had a high enough internal resistance.

Yes the LED has resistance.  It's non-linear and depends on many things (current, temperature, etc), so it's very inconvenient to calculate.

Instead of working out the resistance of an LED you typically assume the LED is going to have a fixed voltage drop across it whilst operating.  Read the datasheet to find out what this voltage drop will be at the current level you want to run it at.  A single white LED will be somewhere around 3V, more if there are many of them inside.

[Nusa]

I'll assume you're talking about the 9V part. You'll notice the datasheet mentions 400mA @ 9V @ 85C and 400mA @10.5V @ 25C. So yes, temperature matters. That max 1A rating assumes your thermal design can keep the case temperatures within the operating range. Datasheet looks like it covers the concepts reasonably well.

[mayor]

Hi all, thanks for the responses!

1. Yes, sorry, it is the 9V version of the LED (I've measured its Vf: between 9.25V and 10V iirc)
2. Temperature is under control in this test (pretty big heat sink, checked with thermocouple, measured at under 50C)
3. I checked the datasheet for the LED, and its Vf is pretty constant across a wide temp range

The power supply is essentially an LM317 (itself driven from an ATX power supply's 12V output) in constant current configuration with a 3 ohm resistor.

Is my understanding of this configuration wrong? I kinda thought there was enough margin for the 9V drop at more than 200mA?

[Whales]

Daft question: have you measured voltages whilst it is operating?  Make sure that eg the 12V atx rail is not doing something silly like sagging down to 10.5V.

[mayor]

Good point. I never actually suspected it might be doing something wonky. Will check, thanks.

[mayor]

I've made some measurements with and without measuring current in series:

Without measuring current:
Power supply output 11.8V
LED 8.8V
LM317 output 9.11V
0.9V across 2.8ohm resistor (across adj and out of LM317)
This is more than 300mA (still not the ~400 I expect)

While measuring current:
Power supply output 11.5V
LED 8.6V
LM317 output 9.17V
0.5-0.6V across 2.8ohm resistor
Roughly 200mA

I am surprised by the impact of the multimeter in series (on 400mA range), and I suspect my breadboard for numbers that seem to jump around a lot.

Still not clear why this isn't delivering the expected current (but yes, higher voltage from the source looks like a good idea) I tried another LED for good measure, same results.

[mikerj]

Still not clear why this isn't delivering the expected current

The reason is clear to see in your measurements; your current source is incapable of delivering enough voltage to set your desired LED current, because your supply voltage is too low to cover the dropout voltage of the LM317 and the voltage drop across the current sense resistor. 

In this state the LM317 can not regulate current, which is why introducing your multimeter into the circuit causes the current to drop even further.

[mayor]

Still not clear why this isn't delivering the expected current

The reason is clear to see in your measurements; your current source is incapable of delivering enough voltage to set your desired LED current, because your supply voltage is too low to cover the drop output voltage of the LM317 and the voltage drop in the current sense resistor. 

In this state the LM317 can not regulate current, which is why introducing your multimeter into the circuit causes the current to drop even further.

Cool- thanks all for the help. I'll find 15V to feed the LM317 to finalise this test.

[ketimoi]

Cool- thanks all for the help. I'll find 15V to feed the LM317 to finalise this test.

Update the post if you able to figure out the problem. Interest to know the solution.
I did similar type experiment with the led cree flashlight. I bought the flashlight from this blog Best Tactical Flashlight Reviews 2018 (Budget Light with Rich Feature). But, I was not successful in that experiment. Although, yours and mine project isn't similar. But, want to know more about your project solution.

[mayor]

Using a 3A 18V source for the LM317, I was able to get more current going through the LED, but still not what I expected.

I indirectly measured current by the voltage drop across the 2.8 ohm feedback resistor, which read 1V. That's roughly 360 mA, so better than the 200 from before. That is still short of the 430mA or so I was expecting. I do read ~430 when the load is a low value power resistor.

I think the next steps will be to solder this up on a perf board and ditch the breadboard.

[mayor]

So I am really stumped. I thought the LM317 might somehow be oscillating, but it isn't. The output is totally flat.

There's 18V coming into it, and current is set by a ~2.4ohm resistor on the output, as stated above. Still getting less than anticipated in the LED.

Anyone have other suggestions or measurements I should take?

[mayor]

Is the LM317 scorching hot?

It would eventually become so if I let it run for a bit. But it reaches the same temp as when using the dummy load. In this latest test, the drop across the LED is a very stable 9.2V.

[mikerj]

Using a 3A 18V source for the LM317, I was able to get more current going through the LED, but still not what I expected.

I indirectly measured current by the voltage drop across the 2.8 ohm feedback resistor, which read 1V. That's roughly 360 mA, so better than the 200 from before.

1v is well outside the specified reference voltage range for the LM317T.  If you have constructed this on a breadboard, then you are very likely  getting a significant voltage drop across the contacts inside.  Instead of measuring across the resistor, measure the voltage between the output and adjust pins on the LM317T itself.  If that is close to 1.25v then the regulator is operating correctly and you need to look at removing unwanted resistance in the sense resistor circuit. 

If the voltage across the LM317T pins is 1v, then you have an out of specification regulator, maybe a counterfeit part.

[mayor]

Using a 3A 18V source for the LM317, I was able to get more current going through the LED, but still not what I expected.

I indirectly measured current by the voltage drop across the 2.8 ohm feedback resistor, which read 1V. That's roughly 360 mA, so better than the 200 from before.

1v is well outside the specified reference voltage range for the LM317T.  If you have constructed this on a breadboard, then you are very likely  getting a significant voltage drop across the contacts inside.  Instead of measuring across the resistor, measure the voltage between the output and adjust pins on the LM317T itself.  If that is close to 1.25v then the regulator is operating correctly and you need to look at removing unwanted resistance in the sense resistor circuit. 

If the voltage across the LM317T pins is 1v, then you have an out of specification regulator, maybe a counterfeit part.

Cool, thank you- I'll def check that out later today. On the other hand the circuit behaves as expected with a resistor as a load instead of the LED.

[mayor]

@mikerj: voltage across adj and out is on spec at 1.25V. Still only 1V across the feedback resistor!

[mikerj]

@mikerj: voltage across adj and out is on spec at 1.25V. Still only 1V across the feedback resistor!

Right, so now you know where the problem lies; something in the path between the LM317T and the resistor is introducing additional voltage drop.  Measure between the Adj/Ouput pins on LM317T and the respective pins of the resistor to find out which side it's on (may be on both sides). 

If it's on a breadboard then look no further, that will be the root cause of the problem.