Re: spread-spectrum model

On Thu, 18 Apr 2024 13:16:04 -0400, Joe Gwinn <joegwinn@comcast.net>
wrote:

On Thu, 18 Apr 2024 08:26:56 -0700, John Larkin
<jjSNIPlarkin@highNONOlandtechnology.com> wrote:
>

I'm designing a switching power supply module and could reduce EMI by
going spread-spectrum on the switching frequency. The simple one below
reduces things by 20 dB. Probe the SS node and FFT.
>
The ss inside switching reg chips is no doubt more sophisticated. In
an FPGA, we could do some sort of pseudo-random thing.
>
On a multi-channel power supply, there may be some small advantage to
have a separate spread per channel. That would be easy.

>
I'd check for cross-correlation as well, so no ganging up in systems
using multiple channels in some signal path.

When my engineers get too fussy about stuff like that, I remind them
"it's just a power supply."

>
Depending on details, the problem could manifest itself as peaks or
ripples in the time domain, your beloved homeland.
>
Joe Gwinn
>

TI has a couple of intersting appnotes

https://www.ti.com/lit/pdf/slyt809

https://www.ti.com/lit/SLVAF18


Their little TPS54302 type parts have radical looking PWM, but the
final DC is super clean. Nice trick.

https://www.dropbox.com/scl/fi/8rytjiwp4hmt2ypgk9bk4/DSC06826.JPG?rlkey=4qipduct0ptrhei07ijdxpsca&raw=1

https://www.dropbox.com/scl/fi/kf2kxbxih6xjbx8uv2o0d/TPS54302_spectrum.JPG?rlkey=rd3diu5nvhasfn7228m8yk665&raw=1

We may get some EMI from switching rise/fall ringing too, in the
hundred-MHz ballpark. It would help to de-phase that too.