Other Parts Discussed in Thread: LM3404, LM3404FSTDIMEV
Hi,
We have worked with the now fairly old LM3404HV part for many years. Works fine. We also have dimmed using 20kHz PWM, also works fine.
Recently we have added one dimming control that uses much lower PWM frequency, around 500Hz, and then we see one borderline issue on some parts pulled from a production batch:
- At certain duty cycles on the DIM-pin (10-15%), the LM3404 might delay the start of switching for quite a "long" time (from Rising edge of the DIM pin), and this is random from cycle to cycle, leading to a very visible flickering and nervous light output. Some parts does not exhibit this beahviour. I have measured a lot on this, and it is clearly seen that when DIM goes high, the parts does not start switching until x us has elapsed, and x is varying randomly.
Everything looks perfect, DIM signal is super-clean and nice, Vcc on pin 7 is also perfect. I tested changing the 10nF bootstrap capacitor to 100nF and then all problems goes away. I have verified this on 4 units that all showed the issue.
This is a smaller production batch, so it's not super-critical, I just wanted to raise the question if 100nF as bootstrap is OK to use or if that might lead to other concerns? Can 10nF be too little if using lower PWM frequencies on the DIM-pin?
Thanks,
Martin