Does anybody have experience with these LED drivers?

This specific board seems to be the "standard" driver used in many millions of Chinese made LED light fixtures. I purchased a few of these drivers, and most seem to work just fine. That is they take in 100W on the AC, and convert it to around 3000ma @ 30v on the output.

But some drivers that I have only take in 85W on the AC side, and put out about 2600-2700ma on the DC output. The LEDs connected to these drivers are visibly dimmer, especially when side by side with a driver that takes the full 100W.

Any idea why this is happening? Just poor tolerances of the components on the board?

 

I think you've answered your own question.

caveat emptor

 

They did claim 3000ma +/- 5% on the spec sheet. So 2600ma is out of spec.

Does this look like a typical switching power supply? Think there is a single resistor somewhere that can be adjusted to "up" the current limiting circuit?

 

On the back:

(U3) STMicroelectronics M101AI Switching Controller
(U1) STMicroelectronics L6562 TRANSITION-MODE PFC CONTROLLER

 

What's the Vf on your LEDs. Maybe it's LEDs and not driver problems.

 

It's the standard 100W LED designed for these power supplies. Or these power supplies are designed for these 100W LEDs. The standard array of 10-series by 10-parallel. 30-32v forward @ 3000ma.

 

Looks like R33 and R36 on the board are the shunt resistors, 0.043ohms each. Then a lead going down to the M101AI which measures voltage drop.

 

Have you checked to see if all drivers have exactly the same board layout and component values? Perhaps there has been a board revision?

 

Have you checked to see if all drivers have exactly the same board layout and component values? Perhaps there has been a board revision?

It's possible, but they are all spec'ed 3000ma. So anything outside +/- 5% is out of spec. Just have to figure out what resistor to adjust.

 

My guess is they just cut it to close. Your LEDs can be 32 to 34 volts, and while the spec on the regulator says 30 to 36 volts, the data sheet has a recommended maximum voltage of 32 volts. Even if they ran it at it's absolute maximum rating of 36 volts it doesn't leave much for current sense and such.

 

M101AI might be up to 32v only, but it's not known if it's also doing voltage control. It's most definitely doing current control, using the R33 and R36 shunt resistors in parallel (0.043ohms each).

I have one supply that takes 105W from the wall, and outputs exactly 100W through the LED. Ain't bad efficiency.

 

So the resistors are 0.043% ohms each (R043), and since they are 3 digits, that means they are 5% tolerance. Right?

And there's 2 of them in parallel, so it's a 0.0215 ohms together. If the supply is currently providing only 2600ma, then 2600ma across 0.0215 ohms would be 55.9mv voltage drop across the these resistors. Right so far?

So if I want the M101AI to provide more current, I'll need so solder another resistor across the 2 already there, so the voltage drop decreases, and the M101AI compensates by sending more current to get the voltage drop it thinks is appropriate.

How does this plan sound?

 

It might be easier to modify R2 or R3 than the low value resistors.
http://www.st.com/web/en/resource/technical/document/datasheet/CD00001116.pdf

Yes, I'm sure it runs in current mode, but you still need enough voltage to turn the LEDs on hard enough to demand the current.

 

It might be easier to modify R2 or R3 than the low value resistors.
http://www.st.com/web/en/resource/technical/document/datasheet/CD00001116.pdf

Yes, I'm sure it runs in current mode, but you still need enough voltage to turn the LEDs on hard enough to demand the current.

I've tried connecting a big resistor parallel with the LED, and the total current remains the same while the LED dims and resistor gets hot. So it's def tuned to 2600ma.

Which R2 or R2 are referring to? The pictures of the circuit board or some schematic from the spec sheet?

 

The spec sheet. You would have to find the ones for your board.