Power supply testing proposal

[6speed]

I realize a competent piece of gear will not require the world's greatest power supply, but I thought I would propose a power supply test anyway. Play multiple tones (including 60Hz) from an amplifier through the power supply and record what comes out the other side. Other than wanting to do this at a high voltage, is there anything that would invalidate it?

 

[Doodski]

I realize a competent piece of gear will not require the world's greatest power supply, but I thought I would propose a power supply test anyway. Play multiple tones (including 60Hz) from an amplifier through the power supply and record what comes out the other side. Other than wanting to do this at a high voltage, is there anything that would invalidate it?

When @amirm tests a amplifier to clipping that is a test of the power supply. Not sure what you mean play tones from a amp through a power supply and see what comes out the other end.

 

[6speed]

External power supplies. Wall warts. Anything aftermarket. Fodder for people arguing about linear vs switching 12V or 5V supplies. Test them in isolation and not their impact on the performance of any given device.

 

[watchnerd]

Play multiple tones (including 60Hz) from an amplifier through the power supply and record what comes out the other side.

Ummm....

The power supply isn't normally in the signal path.

How would this work?

 

[Frank Dernie]

External power supplies. Wall warts. Anything aftermarket. Fodder for people arguing about linear vs switching 12V or 5V supplies. Test them in isolation and not their impact on the performance of any given device.

I don't see the point in that, personally.
It is the complete device that outputs sound into our systems. Testing a PS on its own tells nothing useful so is a waste of time and effort IMO.

 

[Jimbob54]

 

[watchnerd]

I don't see the point in that, personally.
It is the complete device that outputs sound into our systems. Testing a PS on its own tells nothing useful so is a waste of time and effort IMO.

It would be like testing engines without being hooked up to gearing or a drive train.

 

[6speed]

People buy power supplies, either because they are assembling their own product or think that replacing an existing one with another one will make any difference. The reality is that there are many applications for any given power supply, and while the performance of a complete system is what matters, it would be nice to know the capabilities of a specific part.

If you build your own preamp and research opamps, are you going to look at the data sheets or are you going to only look at the test results of complete products that happen to use those opamps?

If you want to use a switching supply for your preamp and don't want to design one from scratch, do you want to buy one off the shelf with proper filtering that is on par with a Benchmark amp, or one that spews garbage all over the audible spectrum?

 

[Doodski]

Ahh you mean use a lab type AC power supply and then measure the ripple or switching noise at the output of a walwart or brick type supply. The filtration and voltage regulation are a solved issue mostly and have been for some time.

 

[6speed]

With no PS filtration test results, I think it is only safe to assume they are solved in concept, not practice. Look at Amir's AVR test results...they are all over the map.

 

[Killingbeans]

Play multiple tones (including 60Hz) from an amplifier through the power supply and record what comes out the other side.

The thing is, what you get from a wall socket is more like an AM radio station from hell riding on a 50/60Hz carrier wave.

Besides, the ripple on the output of a SMPS is mostly leftovers from the switching process (self noise). I can't really imagine that much gets through from the input?

 

[watchnerd]

Ahh you mean use a lab type AC power supply and then measure the ripple or switching noise at the output of a walwart or brick type supply. The filtration and voltage regulation are a solved issue mostly and have been for some time.

Might as well throw in comparison to a battery based UPS in there while you're at it just for giggles.

 

[Doodski]

With no PS filtration test results, I think it is only safe to assume they are solved in concept, not practice. Look at Amir's AVR test results...they are all over the map.

Now you are talking a AV receiver power supply. Those are all over the map for other reasons. Too many power amplifiers connected to one power supply is the main one and loading the power supply down. @amirm tests those power supplies well. If this was a shop repair or test and cal with the top and bottom open then one can start probing around with little extra effort but otherwise it's usually a closed case test. It's rare that a tear down occurs. I've asked for a AC current measurement at the AC mains with the latest NAD 7 ch amplifier so we could see what one module is drawing and now you want to monitor the loaded rail voltages, see ripple and switching noise bandwidth etc. It's lotsa extra work and it can be a can O wurms.

 

[Jimbob54]

Might as well throw in comparison to a battery based UPS in there while you're at it just for giggles.

And portable USB batteries for USB powered DACs - they get touted as "quieter" all over the shop- I suspect they are not, or not all. But whether any noise makes it past the input more depends on the component.

 

[6speed]

Filtering meant for switching noise would be just as effective on AM frequencies. That should only add merit to any testing...assuming Amir can extent a multitone signal and sweep that high.

Batteries would be a valid reference point.

How about the effects of an expensive PS Audio regenerator on a cheapo power supply? Did it remove any AM grunge that made it through the SMPS filtering.

@Doodski I did not mean to refer to AVR power supplies but rather make an analogy to the results of existing AVR testing i.e. the DAC, preamp, amp results via analog, HDMI, coax inputs are inconsistent between models, generations, manufacturers, etc. You cannot predict or take much for granted. Just because a problem is solved on paper does not mean the affliction has gone away.

To make another comparison, look at how much discussion there was about the ESS IMD hump. Just because everyone is now supposed to know the solution does not mean it won't continue to show up in future products.

 

[amirm]

I realize a competent piece of gear will not require the world's greatest power supply, but I thought I would propose a power supply test anyway. Play multiple tones (including 60Hz) from an amplifier through the power supply and record what comes out the other side. Other than wanting to do this at a high voltage, is there anything that would invalidate it?

Power supplies make for nasty loads for an amplifier to handle. It may go into protection or get damaged.

I do have a programmable AC generator where I can modify the source frequency or provide a chopped signal (simulating a dimmer). I have tested a few power supplies this way and it doesn't seem to provide much insight.

 

[Doodski]

There is or was a guy testing PC power supplies the way you are suggesting. It's a complete rigorous test suite that delves into everything. It requires more equipment, many more tests and more space for the test lab and shipping gear incoming and outgoing. I think it's a bad idea.

 

[6speed]

Power supplies make for nasty loads for an amplifier to handle. It may go into protection or get damaged.

Would that be a problem even if the current out of the PS under test is low? I thought PS Audio regenerators work on the same principle--they're just power amps meant to output an unusually high voltage for their rated power.

I do have a programmable AC generator where I can modify the source frequency or provide a chopped signal (simulating a dimmer). I have tested a few power supplies this way and it doesn't seem to provide much insight.

Unless that can sweep the source frequency...

 

[6speed]

There is or was a guy testing PC power supplies the way you are suggesting. It's a complete rigorous test suite that delves into everything. It requires more equipment, many more tests and more space for the test lab and shipping gear incoming and outgoing. I think it's a bad idea.

If that's an unavoidable hurdle, then I agree.

...or we talk that guy into it.

 

[watchnerd]

There is or was a guy testing PC power supplies the way you are suggesting. It's a complete rigorous test suite that delves into everything. It requires more equipment, many more tests and more space for the test lab and shipping gear incoming and outgoing. I think it's a bad idea.

I just don't see the point of testing them outside the system in question.

If it's an internal supply, it's built-in and nothing can be done about it.

If it's external, just swap between power supplies and test the resulting impact on audio signal per usual.