3W star LED module paradox

Thread Starter

mjhilger

Joined Feb 28, 2011
118
Hi all,
I have several 3W & 5W white LED star mounted pieces. There is no drive circuit on the star, just the LED. I have been playing around with these using PWM to change the brightness, etc. I am using lab equipment so all is very tightly controlled. I had run one of the 3W for 30+ hours on my heat sink at about 2.0W. I had the voltage into the circuit around 4.1 volts to mimic Li-ion battery I will eventually use. I wanted to see if the Li-ion would cause any noticable (measurable) changes, due to internal resistance etc. So I removed the lab supply and used one 18650 cell. In moving some things around the PWM sig driving the gate of the HEXFET got loose and the LED got zapped with the full voltage for a short period of time.
Here is where I get confused. The module has been running at around 3.6v 600ma; but it seems after the zap this particular LED now has much higher voltage drive requirements @ 600ma it now requires 7.00v to drive.
I could understand if it had burned out; or if some of the individual cells in the module stopped working and the current requirements were reduced at a lower light output, but how does the voltage requirement go up? The band gap should not have changed - could it? I'm just wondering if anyone has an explanation?
 

Wendy

Joined Mar 24, 2008
23,415
Tell me you used a resistor please. At no time do you not use a current limiter such as a resistor, LEDs are current operated devices. Without current regulation you have likely blown the LED, it sounds like it has been zapped.

You must regulate the current at 600ma, this is not optional. It is not voltage, but current that is critical, use too much current and the LED dies.

A quick tutorial on low power LEDs. What applies for them also applies for their higher power cousins.

LEDs, 555s, Flashers, and Light Chasers
 

Thread Starter

mjhilger

Joined Feb 28, 2011
118
Hi Bill,
I'm well aware of the way LED's work and no I didn't use a limiting resistor as the PWM limits the power and will eventually be in a feedback loop with a .33 ohm resistor monitoring the current. The goal of the design is to create a high efficiency driver. When I'm done with all I want to do, if the micro or the FET fail and blows the LED, then so be it. I'm well within operating ranges of the monitoring devices and FET.

This is why when the full voltage was applied to the LED for a couple of seconds and it didn't blow; but changed V-I characteristics that I am confused.
 

Wendy

Joined Mar 24, 2008
23,415
Actually, you don't. A resistor is not optional, and PWM does not limit current. Even with PWM you MUST have a resistor, this is very, very fundamental.

I have a chapter on PWM in my article that I referred you too, chapter 5 - The 555 and PWM.

There are no circumstances you can do without a resistor or other current limiting equivalents, such as a constant current source. Feed a powerful battery into a LED and it will be damaged, always. Low power batteries (such as CMOS backups) have an internal resistance that will let you get by without, the resistor is inside the battery.

PWM allows you to adjust a light level from the maximum level, but the resistor sets the maximum.
 

someonesdad

Joined Jul 7, 2009
1,583
Bill's advice is simple and important. The PWM does not limit the current if you increase the voltage. Even if a particular on-time of the PWM signal is only one the order of a few microseconds, the instantaneous current can be high enough to damage the device. Hence the specs in the datasheet -- typically both average current and peak current.

On the other hand, if neither you nor your company mind paying for the blown device(s) and associated downtime, that's fine too. I suppose if you were trying to eke out every bit of efficiency, then you could justify leaving out the resistor as long as you accept the risk of a blown or damaged device. It's your call.

Now, as to whether the assumed over-voltage condition was the cause of the change in the V-I characteristics, that's hard to say without taking the device apart and doing a failure analysis. However, if I had to hypothesize, one reason I'd toss out is that some metallization got too hot, melted, and perhaps increased resistance enough to account for what you're seeing. It could also depend on whether there are other active devices on the assembly besides the LED(s).
 
Top