Switching Power Supply Noise Suppression

Thread Starter

Curt Carpenter

Joined Jun 25, 2018
37
I have a +/- 15V switching power supply that is generating about 75mv of noise in my system. Does anyone have any general advice on how I can minimize that? I've tried a variety of filters without much luck. Most of the noise in in the MHz range, with spikes at the switching frequency (I suspect) of the power supplies.
 

Thread Starter

Curt Carpenter

Joined Jun 25, 2018
37
hi Curt,
This video covers some interesting points on SMPS noise reduction.
Maybe helpful.
E
Thanks. I haven't tried snubbers on the +/- supplies, but will give it a shot. The video was helpful with other ideas too that may help me figure out what's going on, so thanks again.
 

Thread Starter

Curt Carpenter

Joined Jun 25, 2018
37
Where are you seeing this noise, from where to where?
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.
 

crutschow

Joined Mar 14, 2008
34,285
It's like ground bounce, but it happens regardless of how I load the supplies.
Does the noise affect the circuit operation?
If not, perhaps you can just live with it.

Switching supplies are known for generating high frequency common-mode ground noise that's hard to suppress.
On a low-noise switching design I did (<100μV), I had to use to an added LC filter on the output plus common-mode chokes on the input and output, to get the low noise I needed,
 

Thread Starter

Curt Carpenter

Joined Jun 25, 2018
37
Does the noise affect the circuit operation?
If not, perhaps you can just live with it.

Switching supplies are known for generating high frequency common-mode ground noise that's hard to suppress.
On a low-noise switching design I did (<100μV), I had to use to an added LC filter on the output plus common-mode chokes on the input and output, to get the low noise I needed,
Thanks. I'm sure not having any luck suppressing the noise I have. And it's getting into some power op-amps (about 7 gain) which of course makes the problem worse depending on the load on the amplifiers.

I'm going to try some of the ideas in the video from Eric Gibbs, but am thinking I'm going to have to go to a linear supply, which is a major change of plan.
 

DickCappels

Joined Aug 21, 2008
10,152
On one project (broadcast TV camera) I had to add a small filter board outside the power supply enclosure. This not only filtered the power supplies but also included a choke in the ground line.

An often helpful trick to greatly eliminate stray return paths is to put a common mode choke in series with the ground and power supply lines. One way is to pass all ground and power conductors through a single core.

Remember that when dealing with high frequencies use chokes with as little capacitive coupling between the terminals as possible. Often fewer turns are better.
 

Thread Starter

Curt Carpenter

Joined Jun 25, 2018
37
On one project (broadcast TV camera) I had to add a small filter board outside the power supply enclosure. This not only filtered the power supplies but also included a choke in the ground line.

An often helpful trick to greatly eliminate stray return paths is to put a common mode choke in series with the ground and power supply lines. One way is to pass all ground and power conductors through a single core.

Remember that when dealing with high frequencies use chokes with as little capacitive coupling between the terminals as possible. Often fewer turns are better.
Thanks, I'll try that.
 

Thread Starter

Curt Carpenter

Joined Jun 25, 2018
37
How low noise do you need?

I'd try an added LC filter followed by a common-mode choke first.
I think I could live with anything under 6mV, which shouldn't be that hard, but nothing I've tried so far has gotten me below about five times that. I have the awful feeling that it's something really dumb that I'm just not seeing.

Time to tear out the power supply and load it up without all the circuitry and try Dick Cappel's ideas. That will take a day I'll never see again...
 

DickCappels

Joined Aug 21, 2008
10,152
If it doesn't have to be a switching supply, use a linear supply. No point in making all that noise if it might be a problem.

A later broadcast camera was a studio camera that had no switching power supplies because I was tired of getting the noise out. The hulking transformer must have added 10 pounds (4.5 kg) to the camera head which made it feel like a real studio camera. Noise was not a problem.

On my first job in electronics we had a 12V 3 amp power supply - think 1971 when cooling fans and 19" racks were common). The regulators were based on the LM723 and had a 5V P-P (visual) test spec. It was probably more like 1 mv RMS.

If you can find a way to avoid making the noise you are way ahead of the game.
 

Thread Starter

Curt Carpenter

Joined Jun 25, 2018
37
If it doesn't have to be a switching supply, use a linear supply. No point in making all that noise if it might be a problem.

A later broadcast camera was a studio camera that had no switching power supplies because I was tired of getting the noise out. The hulking transformer must have added 10 pounds (4.5 kg) to the camera head which made it feel like a real studio camera. Noise was not a problem.

On my first job in electronics we had a 12V 3 amp power supply - think 1971 when cooling fans and 19" racks were common). The regulators were based on the LM723 and had a 5V P-P (visual) test spec. It was probably more like 1 mv RMS.

If you can find a way to avoid making the noise you are way ahead of the game.
It may come to that. At this point, it's more about the thrill of victory and the agony of defeat... The video that Eric Gibbs pointed me to has made the whole thing, despite the frustration, a worthwhile leaning experience at any rate.

If I can't beat the noise with the power supply on the bench, I will go the linear PS route. But I have the bad feeling that once I get the switcher out, it's going to work fine.
 

Thread Starter

Curt Carpenter

Joined Jun 25, 2018
37
I don't see why that would be a bad feeling. It would be your backup plan, and it would probably solve your problem if it persisted to that point.
It would be a bad feeling because it would indicate that the source of my noise problem is elsewhere in the system. There are some 3A power amplifiers in there...
 

Thread Starter

Curt Carpenter

Joined Jun 25, 2018
37
I've given up, and will convert to a linear supply. No matter what I tried, I was never able to get the spikes generated by the switching supply below 30mV peak despite adding line filters, snubbers and trying assorted filters. Maybe a better switcher would help: I noted that the noise increases as the supply warms up over about 15 minutes of use.

I appreciate the inputs. It has been an education!
 

Thread Starter

Curt Carpenter

Joined Jun 25, 2018
37
Did that include common-mode chokes?
Yes. Also low pass LC filters on both + and - supply lines. The noise between the switching spikes was acceptable, but the spikes were a problem I just couldn't resolve. I thought I had them cut down to something I could live with before I discovered that they crept up as the supply warmed up.

Edit: Meanwhile, with a linear supply, the noise is down to 10mV p/p mostly RF from somewhere which seems OK for now.
 
Last edited:
Top