led driver low frequency oscillation help

[pipelined]

Hello everyone,

I'm having problems with an led driver I built.

Circuit Description:
The LED driver board has three boost circuits: one for Red, Green and Blue. The input voltage is 24v and depending on the LED bus, the outputs can swing from 40v to 90v. Each one of these circuits has a 3843 controller and all were set to run at 200kHz. Since the current needs to be held constant (600mA), I used a sense resistor for the current, and fed this signal to the Vfb pin. Since the internal Ref voltage is 2.5 volts, I placed a pullup resistor from this pin to Vref of the 3843. As for the Cs pin ramp, I buffered the Rt/Ct pin to a voltage divider and sized the voltage divider such that the voltage here is less than 1 v.

Problem:
When I connect the circuit to the actual LED bus, the current regulates to the one I set, but there is a 3kHz oscillation riding on the output current, with a pkpk value of 40mA. During startup to around 5 mins at ambient, this oscillation doesn't appear.

When I used a resistive load, this oscillation doesn't appear. BUT, if a I place the actual board in a chamber and heat it up to 85C, the oscillation doesn't happen

I'm hoping anyone out there can help give me some suggestions/tips on how to determine the cause of this. I'm not sure why an LED will cause the regulator to act like this.

Thanks and Best Regards,
Pipelined

 

[Myndex]

I'm assuming that the LEDs on the LED bus are all in series?

Can you upload a schematic?


Since a diode's Vf isn't constant or linear, and varies due to temp, I'll say that's where to look first.

I'm wondering if the 3KHz is related to the switching speed of the LEDs.


Andy

 

[pipelined]

Hi Andy,

Yes, the LEDs are in series. My idea was to make a constant current source to control the brightness of the individual LEDs in the series. To dim the LEDs, I apply a PWM signal w/c turns a switch on/off. This switch(MOSFET) is in series with the LEDs.

The oscillation occurs even if I haven't applied the PWM yet(duty cycle = 100%). By the way, if ever I apply the PWM, it is at 600Hz.

Thanks,
Pipelined

 

[Myndex]

ummmm... You lost me here - you have a mosfet in series with a number of LEDs in series?? Or are you trying to control each individual LED?

Schematic pic?

A

 

[sdmays]

If you have created a constant current source with your 3483 to drive the LEDs, you should not break the flow of current with anything, including the MOSFET. I am speaking of your 600Hz PWM.

If on the other hand, you have created a voltage source with a current limiter to drive the LEDs, then you can safely break the flow of current to your LEDs.

Dumb Question: Do you have a cap across your LED strings? Is the cap sized appropriately to feed a clean, DC, feedback voltage to your controller? If your cap is not sized for the LEDs, the voltage across the LEDs will drop to a point where the impedance looking into the LEDs will approach infinity, and will actually block current until the voltage reaches high enough to turn the LED's back on.

What kind of converter are you using? Forward, Flyback, Boost,...