Simulating load on LED drivers.

[lookimback]

I don't have a particular board I'm working on, but I'm trying to figure out if I could use a diode in place of a string of LEDs so I can get a board to power up and thus continue testing/repairing. This would be for when I no longer have a backlight assembly and most likely will be one of those all in one combo boards. If for instance a circuit was designed to drive 11 3v LEDs, would a 33v zener in reverse bias be an equivalent load?

[mariushm]

Re: Simulating load on LED drivers.

The tv will probably also have a circuit to detect how much current goes through the chain of leds at a certain brightness level ... to detect if some leds in the chain are shorted or something like that.
So i'm not sure if just using a zener would be enough.

It would be easier to just get a bunch of leds and wire them in a prototyping board along with some jumpers or DIP switches to short out segments or individual leds as desired.

[lookimback]

Re: Simulating load on LED drivers.

So, even if I knew the current and output voltage at 100% brightness, it could potentially not work if the brightness were to be turned down? As for the LEDs in a proto board, I had thought of that and have just wired strips together to get close to what I needed. I was just hoping for something cheap and simple. Something where I could maybe even add a single LED in series as an indicator, and could accompany the board when it's sent back to a customer as proof of functionality. I deal with a lot of people who seem to be completely unaware that there is something other than the boards they see when they remove the back from their TV, i.e., LEDs, and that's why they're TV no longer works.

[JDC]

Re: Simulating load on LED drivers.

Not sure if this helps but I do know that a 3.3v 2A zener diode can be used in place of an individual TV LED.

[tom66]

Re: Simulating load on LED drivers.

You could use a 33V zener diode but remember that if the diode conducts 0.2A (a fairly typical LED current) then the diode would need to dissipate 16 watts. I am not sure where you would find a 16 watt zener diode.

Best bet would be some kind of custom current sink or electronic load unit but they are pretty expensive.

[lookimback]

Re: Simulating load on LED drivers.

You could use a 33V zener diode but remember that if the diode conducts 0.2A (a fairly typical LED current) then the diode would need to dissipate 16 watts. I am not sure where you would find a 16 watt zener diode.

Best bet would be some kind of custom current sink or electronic load unit but they are pretty expensive.

I was afraid of that. Would it be possible to use a resistor to limit current, but somehow supply the required current to the sense circuit with a parallel circuit? Also, what would happen if I just used a resistor instead of a diode?

[tom66]

Re: Simulating load on LED drivers.

I was afraid of that. Would it be possible to use a resistor to limit current, but somehow supply the required current to the sense circuit with a parallel circuit? Also, what would happen if I just used a resistor instead of a diode?

Well you could use a resistor if you knew the operating current and operating voltage of the backlight. Ohm's law: R = V / I and power rating would be P = I^2 * R or P = V^2 / R (you'll get same result for both). Resistor would be different for most TVs so no one size fit all option. A resistor in series with the zener diode would not solve your power dissipation issue.

[lookimback]

Re: Simulating load on LED drivers.

Well you could use a resistor if you knew the operating current and operating voltage of the backlight. Ohm's law: R = V / I and power rating would be P = I^2 * R or P = V^2 / R (you'll get same result for both). Resistor would be different for most TVs so no one size fit all option. A resistor in series with the zener diode would not solve your power dissipation issue.

In most cases, I would have at one time had a working backlight, so I could just start keeping record of current and voltage. But, I would still have the dissipation problem, even if I limit current. So the resistor would be dissipating most of the heat then right?

[tom66]

Re: Simulating load on LED drivers.

In most cases, I would have at one time had a working backlight, so I could just start keeping record of current and voltage. But, I would still have the dissipation problem, even if I limit current. So the resistor would be dissipating most of the heat then right?

The resistor would dissipate all of the heat. Instead of converting electrical power into 15% light and 85% heat (LEDs are about 15% efficient), you are converting 100% of it into heat.

The voltage/current ratings of LED strings/power supplies are often printed on the board. Also, you could get it within a fair range and it will still work.

[lookimback]

Re: Simulating load on LED drivers.

I think maybe I'm going about this the wrong way. Since my goal is to prevent the undercurrent protection from shutting the board down, maybe I should just supply the required current to the negative terminal of each LED connector. So, using the 5v rail, if the required current was 20mA then I'd need a 250Ω resistor which would dissipate 0.1 watt. I should then be able to get a voltage reading between positive and negative of the LED connector (since it isn't being shut down). This wouldn't exactly load test the circuit, but it would at least let me know it worked. Alternatively, maybe I could just make something like a decade box with zeners, but then I'd need 2 for most modern TVs. And to complicate it more, some TVs use 3v LEDs and others use 5V LEDs.

[tom66]

Re: Simulating load on LED drivers.

It wouldn't work because the undercurrent detection is almost always done on the high voltage side, not the low voltage side. It regulates the voltage to keep the current within limits and will trip off with either UVP or OVP so you need to keep it happy.

The low voltage side is only used for current balancing between individual strings.

[lookimback]

Re: Simulating load on LED drivers.

It wouldn't work because the undercurrent detection is almost always done on the high voltage side, not the low voltage side. It regulates the voltage to keep the current within limits and will trip off with either UVP or OVP so you need to keep it happy.

The low voltage side is only used for current balancing between individual strings.

Damn, I thought I was on to something.

[lookimback]

Re: Simulating load on LED drivers.

I think the easiest, and cheapest thing I can do is similar to what Mariushm said. I could substitute all but one of the LEDs with zeners of equivalent value, so I have my indicator. Then, instead of the dip switches I could just use a dual row header and a header shunt jumper. Or, if I wanted to get creative, maybe I could make a custom slide switch.

[lookimback]

Re: Simulating load on LED drivers.

If I wanted to do something more complex, could I use a MOSFET and do something like a transistor constant current circuit with an adjustable output current?

[tom66]

Re: Simulating load on LED drivers.

If I wanted to do something more complex, could I use a MOSFET and do something like a transistor constant current circuit with an adjustable output current?

Yes you could

If you want some ideas, see the schematics here


(this was a project I designed about 6 years ago)

You'd need a much bigger transistor and a heatsink, but you could build this with a potentiometer as an input and use it to regulate a sink current

You would probably need four duplicates of this circuit for a four-channel unit, but they could be driven off one potentiometer. It may be possible for the channels to be put in parallel, but I do not know how well this will work.

[lookimback]

Re: Simulating load on LED drivers.

Yes you could

If you want some ideas, see the schematics here


(this was a project I designed about 6 years ago)

You'd need a much bigger transistor and a heatsink, but you could build this with a potentiometer as an input and use it to regulate a sink current

You would probably need four duplicates of this circuit for a four-channel unit, but they could be driven off one potentiometer. It may be possible for the channels to be put in parallel, but I do not know how well this will work.

I like this idea because it wouldn't require me to know how many LEDs the board is designed to drive. I could see me getting that wrong and blowing something up.
I've seen some boards with 5 channels.
Four duplicates is probably good enough for what I'm likely to encounter though. Would this be able to make this work with something which drives between 9 and 60 LEDs which could be 3 or 5v. That's going to me a lot of wattage to dissipate.

[tom66]

Re: Simulating load on LED drivers.

Yes you'd need a very big heatsink to do it, but it is possible.

Transistors are very good at dissipating heat. You can dissipate 200W in a standard TO-220 transistor with a large heatsink. It'll get hot, but be within operating characteristics.

[lookimback]

Re: Simulating load on LED drivers.

Could this be powered by the LED driver, or would it need it's own power source?

[tom66]

Re: Simulating load on LED drivers.

Could this be powered by the LED driver, or would it need it's own power source?

The logic would need its own power source but a 9V battery would likely work (just make sure to choose an opamp that works with a Vdd of 9V)

The rest of the power would naturally come from the LED driver, as you are trying to load that, but that shouldn't power the logic, because it's very easy to get into a situation where you will end up oscillating the LED driver if you run your control logic from the same power supply. Think about if your current sink draws too much current; this will cause the op-amp current to reduce, but cause the LED driver voltage to rise...

[lookimback]

Re: Simulating load on LED drivers.

Ok, that makes sense.