Hi,
I'm experiencing a problem with a product we've recently designed which sometimes produced audible noise at half the PWM dimming frequency (= 25KHz). It seems that if the falling edge of the 'DIM' input coincides with the falling edge of the switcher mosfet drive pulse it can result in a single shortened (or sometimes extended) pulse, as if there's an internal race going on between these edges. This disturbance can then cause alternate DIM cycles to produce small (< 1V) modulation of the anode voltage during the off period of the LED strings, and in turn this creates some noise at 12.5KHz that some customers complain about (although my hearing is too shot to detect it!).
I'm pretty happy our control loop is stable, but I could well believe the behaviour is caused by poor grounding around the current feedback pin. However, I've tried various filtering methods there to no avail.
Has anyone encountered similar behaviour?
Thanks,
Keith