Re: Interesting reading about spread-spectrum, also for on-board designs

On Mon, 24 Jun 2024 10:35:42 -0400, Joe Gwinn <joegwinn@comcast.net>
wrote:

On Sun, 16 Jun 2024 07:01:41 -0700, john larkin <jl@650pot.com> wrote:
>

On Sun, 16 Jun 2024 05:26:35 GMT, Jan Panteltje <alien@comet.invalid>
wrote:
>

On a sunny day (Sat, 15 Jun 2024 08:43:12 -0700) it happened john larkin
<jl@650pot.com> wrote in <esar6jdro3r7ki70t17jsmpo48qkkg05na@4ax.com>:
>

On Sat, 15 Jun 2024 05:57:23 GMT, Jan Panteltje <alien@comet.invalid>
wrote:
>

On a sunny day (Fri, 14 Jun 2024 07:55:14 -0700) it happened john larkin
<jl@650pot.com> wrote in <45mo6jhcq8kisjmbrom8i1r17ljr6g4qu8@4ax.com>:
>

On Fri, 14 Jun 2024 07:43:53 GMT, Jan Panteltje <alien@comet.invalid>
wrote:
>

Interesting reading about spread-spectrum, also for on-board designs:
>
>
https://www.electronicdesign.com/technologies/power/article/55041243/monolithic-power-systems-choosing-the-proper-parameters-in-frequency-spread-spectrum-fss-design

>
That's cool, except for the usual dreadful fake pcb image.

>
Fake?
Sure some fun zig-zag loops in some tracks... ?
Why would anybody bother making artwork like that when you can just take a picture?

>
I've wondered that. And when they use a picture, it tends to be some
ancient crufty Dip-package board.
>
I have rarely used wiggle traces to add delay, but the fake board
images are full of them. And glowing traces. And vias without holes.
>
Hey, the electronic designers here could post pics of our prettiest
real circuit boards.
>

>
>

We have a bunch of switchers that use our own FPGAs as the
controllers, and I need to start up a project to make them
spread-spectrum.

>
I find it interesting, not much experience with spread spectrum,
but a lot with wideband FM modulation like for the old Ampex video recorders... VHS, Umatic, Betamax too.
So with a bit of luck you can send your commercials on a few hundred kHz :-)
>
>

We don't want the ss modulation to show up in the DC outputs as ripple
or anything audible.

.

>

RF wideband filtering, inductors.. capacitors...
>

>
The jitter pattern needs to create spectral spreading at high
frequencies but not change averages at low frequencies. There might be
some math involved.

>
Yes I'v heard about maaz
I do not see many tennis players use math to calculate how to move and where to point the ball.
Would take too long.

>
Given the speed of measurable electrical nerve impulses, most sports
should be impossible. Recognizing an image, one of millions, in
milliseconds is even more improbable.
>
Something else is going on.

>
In massive parallel:  Fit, extrapolate, intercept.
>
Joe Gwinn

Even massive parallel has layers of logic, and chemistry is slow.

I suspect that the nerve impulses that we can detect electrically are
not the real story. It is assumed that the pulse timing conveys the
data, which makes no sense.