20 V peak on a buck converter output

Question

I noticed that an LM5007 buck converter set to 3.3 V on startup shows a peak of 20 V for 150 ns, and then a smaller peak after 100 ns. I know that the feedback resistors have the wrong values.

Below are screenshots from the design files.

Not measured by me, try zooming in on it.

Are you sure that’s real and not an EMC spike that your scope is picking up? — Frog, Feb 11 '23 at 08:39
Your C45 does not have that 1R damping resistor in series with it, as shown in the LM5007 data sheet. Also, is a load connected for the measurement? — electroGeek, Feb 11 '23 at 09:17
No load is connected, I already told my client to try it that way. — Marek Bílý, Feb 11 '23 at 11:18
@electroGeek maybe it's a tantalum. Hard to say at this distance. OP: is red layer a solid ground plane? — Tim Williams, Feb 11 '23 at 14:49

score 1 · Accepted Answer · answered Feb 11 '23 at 11:43

That measurement setup is inadequate for high bandwidth measurements. Long wires are picking up noise from external fields. In this case it is probably the changing magnetic field of the (unshielded) inductor during the switching event generating that noise.

To measure the actual output voltage, try using significantly shorter wires, ideally use the ground spring to measure directly at the output capacitors: _{Image borrowed from here}

20 V peak on a buck converter output

1 Answers1