Switching Power Supply Noise Suppression

BobaMosfet

Joined Jul 1, 2009
2,110
Thanks for the reply.

The noise is pretty much everywhere, and it behaves a lot like the sort of thing you see with a ground loop (although I haven't been able to identify such a loop so far). On my 'scope's 20mv range, I can see the noise wherever I probe, and can tune it a little by adding or moving ground conductors. It's like ground bounce, but it happens regardless of how I load the supplies.
You're going to see digital noise on the ground. Ground is never perfectly 0V in a digital environment. Digital circuitry pulls current in fits and starts. Bypass caps help with this, but it's there, and you may also be coupling your incumbent supply frequency (In the US, it's 60Hz AC). It can take some clever effort to minimize it, if it's a pain. You can also use isolation techniques for portions of your circuit that are impacted.
 

Thread Starter

Curt Carpenter

Joined Jun 25, 2018
37
You're going to see digital noise on the ground. Ground is never perfectly 0V in a digital environment. Digital circuitry pulls current in fits and starts. Bypass caps help with this, but it's there, and you may also be coupling your incumbent supply frequency (In the US, it's 60Hz AC). It can take some clever effort to minimize it, if it's a pain. You can also use isolation techniques for portions of your circuit that are impacted.
Thanks. I was pretty careful to keep the analog and digital grounds as separated as I could, and haven't seen anything that I could identify as digital noise. If it was there, it was being swamped by the power supply contribution. I'm sensitive to the problem though: back when high-speed CMOS logic was just coming on the market, I discovered you could build a pretty good transmitter using a 74HC04 with the inverters wired up in a ring and a length of wire attached to Vdd and cut to a quarter wavelength of the oscillating frequency.
 

Thread Starter

Curt Carpenter

Joined Jun 25, 2018
37
Just to follow up...

I kept beating on it, and was finally able to to get my switching power supply noise down to 5mV. The key turned out to be keeping the grounds from my +15 and -15 supplies totally isolated from each other until until after each supply went through its common mode choke, then rejoining the grounds and running the + and - rails through low-pass LC filters. With the two grounds connected directly together at the supplies, it seems to have created a ground loop or something of that sort (no diagrams for the supplies to study, so not sure exactly what was going on.) The rise in noise level as the supplies warmed up seems to be cured too.

Meanwhile though -- got my linear supply noise down to under 1mV, just barely visible on my scope. Lesson learned!
 

DickCappels

Joined Aug 21, 2008
10,152
Very good outcome! You can learn a lot chasing noise but you have to pay a high price for the lessons. Congratulations on getting to this point.
 
Top