Power supply characterization

[Matt Giarra]

I am interested in better understanding DC power supplies used in receivers/amplifiers. What are some approaches to characterizing DC supplies used for audio equipment, including relevant metrics? I understand the blunter characterizations commonly performed for generic/bench top power supplies, e.g., ripple, stability, etc. I am more interested in characterizations/metrics that are particularly relevant for audio, like frequency response, which one might not typically perform/measure for supplies with applications besides audio.

 

[drjim]

Power supplies don't have a "Frequency Response" of any sort. They take AC power, and convert it to pure, stable DC power. That's all. The terms used to describe current specs for power supplies have change a bit, from "Ripple, Hum, and Noise" to "Periodic and Random Disturbances", but that's just a semantics thing.

All you care about is the purity and stability of the DC power.; Frequency response has no meaning in this application.

 

[heyraz]

Look at this:

http://www.r-type.org/articles/art-144.htm

 

[petes'picks]

There are some books on amplifier design that cover the subject of power supplies for audio. Here is a nice video with some good information.

 

[petes'picks]

I found good information in this article.

https://www.tnt-audio.com/clinica/ssps1_e.html

 

[gadget73]

Power supplies don't have a "Frequency Response" of any sort. They take AC power, and convert it to pure, stable DC power. That's all. The terms used to describe current specs for power supplies have change a bit, from "Ripple, Hum, and Noise" to "Periodic and Random Disturbances", but that's just a semantics thing.

All you care about is the purity and stability of the DC power.; Frequency response has no meaning in this application.

they can have a sort-of frequency response, but generally its an artifact of poor decoupling. Motorboating is an example of that. Can either happen from bad parts, or just bad design in the first place. Get the wrong mix of stuff and the supply nodes will "wag" at a rate that results in oscillation.

 

[Matt Giarra]

Power supplies don't have a "Frequency Response" of any sort. They take AC power, and convert it to pure, stable DC power. That's all. The terms used to describe current specs for power supplies have change a bit, from "Ripple, Hum, and Noise" to "Periodic and Random Disturbances", but that's just a semantics thing.

All you care about is the purity and stability of the DC power.; Frequency response has no meaning in this application.

Thank you for the response. I may be misunderstanding your post, but generally speaking, power supplies certainly do have frequency response in regards to their abilities to vary their outputs in response to time-varying loads. The following application notes should make that clear:

Keysight Technologies: Power Supply Control Loop Response (Bode Plot) Measurements

Texas Instruments: Considerations for Measuring Loop Gain in Power Supplies

Analog Devices: [Understanding] Power Supply Loop Stability and Loop Compensation

Digikey: Understanding Switching Regulator Control Loop Response

Electrical Design News (EDN): Testing a power supply – Stability

In any case, my question is about what approaches are uniquely or especially insightful for characterizing audio amplifier power supplies, not which common ones aren’t.

 

[drjim]

You're talking about Transient Response. That falls under the "Well Regulated" or "Stable" category. I suppose you could call it "Frequency Response" of a sort, but I don't.

 

[rcs16]

Power supply rejection ratio is a frequency dependant characteristic. A power supply and a circuit both have a psrr spec. Sensitive circuits that do not have good psrr require the ps to have it or ps artifacts get through into the audio. Ps Transient response applies to both the input and the load, they are separate tests. Ps requirements depend on the circuits they are powering and the specifications required. Many low level audio circuits are class A amplifiers, where ps current draw us near constant so ps load change transients are not so important where as AC line voltage and transients need to be stabilized and suppressed.