question about mA draw

papasan

Enlightened
Joined
Mar 25, 2001
Messages
621
Location
Northern Virginia
i have an array of 17 white nichias in parallel hooked up to 3 nimhs. the nichias are from various sources, basically all of the white leds i have left over from past experiments.

the thing is, after about a minute after closing the circuit the leds start to pull 900mA @ 3.55V. that's 53mA per led!

i'm curious why the leds would pull so much from a 3.6V power source. is it something about hooking so many up in parallel or something? i wouldn't think it would pull this much unless i drove them to 4+V.

would this, over time, damage the leds and reduce their half-life? i would imagine it would, but i remember something about over-voltage killing leds faster and these are technically at voltage specs.
 

lambda

Flashlight Enthusiast
Joined
Jan 6, 2002
Messages
1,795
Location
Iowa
papasan,

With them all hooked up in parallel, the ones with lower voltage drop might even be drawing more than you think. And the ones with a higher voltage drop are drawing less current.

So, long term effects, no real good way to tell. But, I have a white 5mm been running for months at 100ma. Hasn't craped out yet.......no guts, no glory.......
 

papasan

Enlightened
Joined
Mar 25, 2001
Messages
621
Location
Northern Virginia
i would have thought that the voltage drop differences would even out with 17 leds. perhaps i'm just extremely lucky and have alot of low-voltage leds.

i've forgotten how much i like 5mm nichia leds. alot of nice spill-off light to the sides that just isn't there with a luxeon. the whole basement here lights up nicely with this cluster running. and the cool shadows you can make with 17 light sources =).
 

Lux Luthor

Flashlight Enthusiast
Joined
Nov 10, 2000
Messages
1,944
Location
Connecticut
<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by papasan:
...i've forgotten how much i like 5mm nichia leds. alot of nice spill-off light to the sides that just isn't there with a luxeon. the whole basement here lights up nicely with this cluster running. and the cool shadows you can make with 17 light sources =).<HR></BLOCKQUOTE>

This is so true. The Luxeons just don't replace 5mm LED arrays for general lighting. They have almost NO spill light.

The only way I've been able to use them is by shining them at the ceiling. I have put them in headlamps, but the lack of spill light is unacceptable to me, so I stick with my BD Moonlight with its 4 LEDs for this.
 

Brock

Flashaholic
Joined
Aug 6, 2000
Messages
6,346
Location
Green Bay, WI USA
Another thing to keep in mind they won't run that way for long because of the voltage drop as the batteries are used. I bet even after about 10 minutes you would be about the 700mA range. Were the batteries fresh out of the charger?
 

papasan

Enlightened
Joined
Mar 25, 2001
Messages
621
Location
Northern Virginia
the batteries were unused but sat out of the charger for a few days before i used them, they probably didn't have the 1.4V or so they would have had right out of the charger. the 3.55V reading was taken while they were under load a couple of seconds after they reached 900mA output.

interestingly, the load started as 650mA or so and rose to 900mA after a minute or two. after an hour it was still over 800mA but i neglected to take the voltage reading. just some informal tests to make sure i got my soldering good while waiting for the electronics to put in there.
 

Brock

Flashaholic
Joined
Aug 6, 2000
Messages
6,346
Location
Green Bay, WI USA
wow, must be some good batteries. I have noticed that LED do have a "ramp up", they tend to pull a bit more power after a short time. I take all my reading after they have run for 60 seconds.
 

Wim Hertog

Enlightened
Joined
May 18, 2002
Messages
271
Location
Bornem, Belgium
I have just the same problem with the design of my 21 LED headlamp. I've ordered 20 white LEDs from ebay and half of them draws 20-30 mA per LED @ 3.75 V and the other half draws 50-65 mA per LED. I have one that draws even 95 mA @ 3.75 V!
LEDs also draw more power when they get hot, so when you power up your LEDs, they will start to warm up and draw more power!

Lambda: Does your LED running @ 100 mA still produce as much light @ 20 mA as a new LED produces?
I think that it will degrade rather fast @ 100 mA.
 

Jonathan

Enlightened
Joined
Dec 14, 2001
Messages
565
Location
Portland, OR
<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by papasan:

interestingly, the load started as 650mA or so and rose to 900mA after a minute or two. after an hour it was still over 800mA but i neglected to take the voltage reading. just some informal tests to make sure i got my soldering good while waiting for the electronics to put in there.
<HR></BLOCKQUOTE>

What you may be seeing is a thermal run away effect.

As a LED heats up, its forward voltage drops, meaning that the voltage needed to push a constant current through it will go down. Another way of saying this is that as a LED heats up, the current which flows through it when supplied with constant voltage will go up.

What makes the LED heat up? The _current_ flowing through the LED causes heating, and the heat can only be carried away so quickly, so the junction gets hotter.

This leads to thermal run-away: If you connect the LED to a perfect voltage source, then a certain amount of current will flow, and the LED will heat up, and its forward voltage will drop, so more current flows, and the LED heats up some more,....

Usually we don't use a perfect voltage source. Instead we use something like a battery with internal resistance. As the current starts going up, the battery output voltage drops, and limits just how high the current can soar. The system reaches an equilibrium, where the battery voltage at current X matches the LED voltage at current X.

But when you have LEDs in parallel, you have a situation where the power supply is capable of supplying _much_ more current than a single LED can tolerate. As far as the LED is concerned, the current which the supply could feed it is 'infinite'.

When you connect your LED array, all of the LEDs are cold, at room temperature. Some LEDs have lower voltage drops than others, so they carry _more_ current than the others. These LEDs get hotter, faster than the others, and the forward voltage differential gets even greater. As the LEDs heat up, the total current flow will increase, but the 'hot' LEDs will take more and more of the current.

Push this situation too hard, and the 'hot' LEDs will fail, leading to a chain reaction where the next 'hottest' LEDs will take the power and fail.

I'd suggest that you feel around on your set of LEDs while they are running, to see if any are particularly hot. You may also want to try doing something to 'heat sink' the cathode side of the LEDs; this will help keep the junction cool.

-Jon
 

papasan

Enlightened
Joined
Mar 25, 2001
Messages
621
Location
Northern Virginia
that's funny...in a past discussion of thermal run-away i stated just about what you said and was shot down. "resistance increases with heat" i was told.

with it running some of the negative leads get warm, but not really hot. they are all soldered to a radio shack pc board (a graph of holes with copper surround) and are all jumped with big gobs of solder. not great heat sinking, but it's adequet. i don't plan on letting the array get past 600mA or so with the controlling circuit there.
 

Jonathan

Enlightened
Joined
Dec 14, 2001
Messages
565
Location
Portland, OR
<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:</font><HR>Originally posted by papasan:
[QB]that's funny...in a past discussion of thermal run-away i stated just about what you said and was shot down. "resistance increases with heat" i was told.
[QB]<HR></BLOCKQUOTE>

Well...it depends upon the device, and you have to look at the datasheets. The specification of interest is 'temperature coefficient of forward voltage'. If the device has a negative temperature coefficient of forward voltage, then the voltage drop goes down as the temperature increases.

If you look at the datasheet for the Luxeon 1W LEDs, they are reported to have a temperature coefficient of forward voltage drop or -2mV/C, and the forward voltage will drop as the temperature goes up. I don't have the corresponding number for Nichias...but I strongly suspect that the number is negative for them as well.

-Jon
 
D

**DONOTDELETE**

Guest
All lumenescent materials have a negative thermal coefficient, otherwise thermal runaway would occur immediately.

The coefficient is actually dependent on the current through the device. The higher the current, the higher (that is, less negative) the coefficient.

As current increases and the coefficient reaches zero, thermal runaway occurs.
 

Jonathan

Enlightened
Joined
Dec 14, 2001
Messages
565
Location
Portland, OR
Duggg,

I think that you and I are using 'thermal coefficient' differently. I am describing the 'thermal coefficient of forward voltage'. This is simply the change in forward voltage _at constant current_ with temperature change.

A negative thermal coefficient of forward voltage does not imply stability. It simply means that the forward voltage will drop if the temperature goes up. If a device with a negative thermal coefficient of forward voltage is connected to a stiff voltage source, then current will be expected to go up as the junction temperature rises, which could actually lead to thermal runaway if other factors didn't limit the current flow.

On the other hand, this same device with a negative thermal coefficient of forward voltage, when operated with a regulated _current_ source, would be inherently stable, since the forward voltage would drop as the temperature increased, meaning that the total power delivered to the device would _decrease_.

Finally, a device could be rated as having a thermal coefficient of _current_ at a constant voltage, which would have the opposite sign as the thermal coefficient of voltage, at least for devices which consume power (LEDs).

-Jon
 
Top