This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

Unexpected behavior when dimming White LEDs with TPS61043

Other Parts Discussed in Thread: TPS61043

Hello everybody,

I am working with the chip TPS61043 to control the backlight white LEDs of a LCD and I am observing a strange behavior. We want to use the pin CTRL to dim the lights. For design constraints the PWM signal that dims the LEDs is generated by a microcontroller (by Phillips model LPC2214) and has a frequency of 50kHz (which is the upper limit according to the datasheet) and the duty PWM is modified from 0 to 256 steps. To control the LEDs a bit more current is needed (TPS61043 is limited to 60mA so we added a MOSFET to get a gain in current, as you can see in the attached file. At the end the maximum current is almost 80mA) R213 is 0 ohms. So far, everything pretty straightforward.

When we connect an external signal generator (we are using the model RIGOL DG4162), the dimming is 100% linear but when we connect the PWM signal directly from the microcontroller to the CTRL pin, the brightness observed in the LEDs changes abruptly when passing from 125 to 126 (which corresponds to an 50% duty cycle). The current that flows through the LEDS  I measured when using the signal generated by the microcontroller:

Duty: 48.75% -> 14.54 mA

Duty: 49.14% -> 14.84 mA 

Duty: 49.53% -> 62.72mA (this is more than 4 times the previous current!!!! A sudden increase in the intensitiy of the backlight can be appreciated)

Duty: 49.92% -> 63.33mA

Duty: 50.31% -> 64.24mA (from this point on, the variations in current are very smooth)

The same duty but generated with the external signal generator:

Duty: 48.75% -> 43.63  mA

Duty: 49.14% -> 43.93 mA 

Duty: 49.53% -> 44.24 mA 

Duty: 49.92% -> 47.27 mA

Duty: 50.31% -> 47.49 mA (the variations are always very smooth)

The signal generated by the micro and measured at the pin CTRL doesn't present any anomalies (correct duty cycle and abrupt rising and falling edges). I tried connecting an OPAM in follower emitter between the micro and the TPS with the same odd results. I also tried driving a MOSFET with the micro and connecting the drain to the CTRL and a pull-up of 1kHz. I also tried with a BJT (thinking that maybe the micro wasn't giving enough current). I also try solding a 0 Ohm in R215 and R213 (to short the pins RS and FB) and connect R220 with RS with the same odd results.

Another thing that i tried is using the FB pin and calculating the resistors needed with the equations from the datasheet (page 15), but I cannot manage to turn off the LEDs completely (which is one of our requirements) so this option is out of the question. 

I cannot find an explanation or a solution to our problem. Any ideas?

Thanks a lot.