LM317 Constant Current questions

[audioguru]

You test LEDs without using a current-limiting resistor?
Do you also jump out of airplanes without using a parachute?

 

[HarveyH42]

It was only for a very brief second to make sure of which set of leads was anode and cathode. I didn't expect it to actually explode on contact, figured it wouldn't even get warm.

Anyway, the experiment was a miseriable disapointment, the two LEDs were very dim. Will make some measurements after work and see why.

 

[eblc1388]

It was only for a very brief second to make sure of which set of leads was anode and cathode.

Yes, also a good method to test a fuse to make sure it works.

Speaking of which I must add that LEDs I own from various manufacturers do not follow a standard on the polarity indication.

Long/short leads or a flat on the LED body will not always indicate the correct polarity.

Or would I say most do but a few don't.

 

[audioguru]

It was only for a very brief second to make sure of which set of leads was anode and cathode. I didn't expect it to actually explode on contact, figured it wouldn't even get warm.

Over-current can blow-up an LED much faster than a split second. The chip is very small and will reach melting temperature long before its case begins to get warm. Their reverse voltage rating is only 5V so it might also blow-up when its supply is reversed.

I have installed and tested hundreds of LEDs and never blew one up.

 

[HarveyH42]

I wouldn't have done it with a good 9 volt battery, the spent one should have be okay, or at least not enough to send pieces flying. Could have lost an eye from that explosion. Got lucky, figured I'd share my stupidity and hopefully spare someone the grief.

 

[Hero999]

A good way of testing LEDs is to connect the cathode to -V, moisten your finger and use it to bridge the gap between the LED and +V of the battery. The current drawn is too small to hurt you or the LED which should glow quite dimly but you should be able to see it.

 

[audioguru]

You can't find an ordinary resistor?

 

[Hero999]

Yes but I normally don't bother searching for one.

 

[HarveyH42]

Okay, finally succesful run. I made some measurements, got weird ass numbers, so figured I did something wrong. Did some more research...

Instead of driving the LEDs from the output pin of the LM317, I need to come off the Adjust pin (wrong side of the resistor). Also learned that I needed 3 volts above what the LEDS needed, so for the 5.4v volts from the NiCds, only had enough for 1 LED. Left it running at 7:30pm, it's still bright at 1:30 AM.

One odd note, measured 64.8 mA with the meter between the LED and neg. battery. The resistor is 10 ohms.

 

[eblc1388]

One odd note, measured 64.8 mA with the meter between the LED and neg. battery. The resistor is 10 ohms.

Nothing odd. This measuring method is only good if your LM317 input voltage is higher than 5.4V, perhaps 7V or more because you have introduced extra resistance(the DVM) into the circuit and might have exceeded the regulating range of LM317 to maintain its output. Proper method is to measure the voltage across the 10 ohm resistor and calculate the current.

If your constant current circuit is working, you should get a constant 1.25V across the 10 ohm resistor, which also means a constant 125mA current. Anything less indicates that the LM317 input voltage is too low for regulation.

 

[audioguru]

Use an 18 ohm resistor to avoid burning out your LED when make the input voltage to the LM317 high enough.

When you have an input voltage that is high enough for the LM317 to function properly, measure the voltage across the resistor to determine the current and measure the voltage across the LED to see how different it is from its spec.