SEPIC Design for MP3425

Greetings,

A few years ago, I designed the MP3425 into a SEPIC configuration into an audio product. Recently I’ve been getting a few back for unstable supplies that are causing audio noise in the 40kHz range, when the input voltage is between 6-9V. I wouldn’t describe myself as a power supply guy, but I tried a bunch of stuff in an attempt to stabilize the supply within the design constraints.

I’ve tried:

  • A feed forward cap (no difference seen)
  • 0.1uF input/output caps closer to the part (no difference seen)
  • Changing the comp cap (no difference seen)
  • Changing the comp resistor (changes the frequency of oscillation, but not stable input range)
  • Raising C501 (makes the voltage across C501 more stable, but no change in instability)
  • Swapping the IC (fixes it)
  • Swapping the problem IC back onto the board (brings the problem back)
  • Trying a 1:1 coupled inductor instead of the 2x uncoupled inductors (no difference)
  • Adding 0.1 Ohms in series with the output cap (no difference seen)
  • Lowing the output cap (no difference seen)
  • Changing the switching frequency (moves the instability around, but it still exists)

It seems like the problems just follow the ICs, which is odd because when I was swapping ICs, I was just swapping 2x ICs used on the same board for different supplies, which means they were likely from the exact same tape and reel. Good ICs seem to be good across the entire range.

Is there anything quick and obvious I’m doing wrong?
Picture below for the circuit.

Thanks for any help!

Hi Josh,

Thank you for posting on the MPS Technical Forum.

It is my understanding that when you swap the ICs, the design works. Did you try putting the “problem IC” into the other rail to see if it works there? And what is the input on that other rail?

I don’t see anything obviously wrong, but SEPIC design can be complex.

Thanks,

Allison