0
\$\begingroup\$

I have an SMPS producing output of 12V,1Amps. It is giving a voltage ripple (I am not sure if ripple right term for this) as given in image. enter image description here I donot have permission to make any changes to this SMPS Circuit. I use this 12V as an input to a switching regulator that output 5VDC as given circuit below.

enter image description here

When I measure the 5V in oscilloscope it still shows the same ripple (about 4V pk-pk) as seen earlier in 12V input. This is causing microcontroller in my board unable to flash sometimes.

enter image description here enter image description here

When this ripple is reduced to around 500mv pk-pk (This I do by disconnecting external load connected to SMPS) it flashes without issues. Please suggest some solutions that I can do on my regulator circuit shown above to nullify this even when load connected.

My circuit after regulator draws <250mA of current. SMPS Frequency : 100kHz Switching Regulator frequency: 60kHz Oscilloscope images are taken with DC coupling. Yes, surely there is issue with SMPS,but as I said, I donot have permission to alter anything in it.

I have tried to increase capacitance C66 from 470uF to 1000uF, But no use. I tried by adding various combination of low pass filters at output. Still not even a mm/mV drop in this ripple.

\$\endgroup\$
12
  • \$\begingroup\$ I think that's called ringing rather than ripple. Looks like it is 30MHz. You could try a ferrite bead. That seems like a serious problem with the SMPS itself though. It's supposed to be 5V but is slamming the output with 8V. Anything running off that actually needs that 5V is going to have trouble and no one is allowed to do anything about it? \$\endgroup\$
    – DKNguyen
    Commented Aug 26, 2021 at 18:19
  • \$\begingroup\$ As DKNguynen said, at 30MHz ferrite beads on your 12v line could help. Also, note that electrolytics dont have much use at those frequencies. Ceramic and PP film caps are useful. You can also try putting a 0.1uF ceramic cap very close to the chip being flashed. And a ceramic or pp film cap at your 12 v input after the ferrite bead. It looks like the 12v supply is automotive grade. Not really a clean output. \$\endgroup\$
    – kalyanswaroop
    Commented Aug 26, 2021 at 18:49
  • \$\begingroup\$ Show us where you're picking off the ground and 5V from the circuit. If the ground wire from the cold end of C67 to the ground of your microprocessor has current from the 12V SMPS circuit flowing in it, then your problem isn't a dirty 12V supply -- it's a dirty ground, from ground bounce. \$\endgroup\$
    – TimWescott
    Commented Aug 26, 2021 at 19:05
  • 1
    \$\begingroup\$ If the scope probe is near the inductor or the loop where high switching currents occur, it might also couple capacitively or magnetically to the probe. Use the ground spring on the probe, the crocodile clip lead is useless at these frequencies. \$\endgroup\$
    – Justme
    Commented Aug 26, 2021 at 19:31
  • \$\begingroup\$ @DKNguyen : Thanks for the hint, I only had normal inductors with me. I have placed order for ferrite bead NFZ15SF102SZ10D and it should reach me in 1week. Will update you after testing it. \$\endgroup\$
    – Vishwesh GM
    Commented Aug 28, 2021 at 5:02

0

Reset to default