sonic difference between linear and switch mode PSU?

[outoftune]

i have been seeing a lot more audio products using SMPS these days...just wonder if anyone has ever done an A/B comparison between a linear PSU and SMPS with the same circuit...and if there is any audible difference?

 

[mjrippe]

As long as the PSUs are *properly filtered* and only supplying DC voltages there will be no sonic difference between the two.  Really, DC is DC.  So long as the supply can provide the correct voltage and current all will be well.

 

[soapfoot]

As long as the PSUs are *properly filtered* and only supplying DC voltages there will be no sonic difference between the two.  Really, DC is DC.  So long as the supply can provide the correct voltage and current all will be well.

You forgot-- "as long as the supply responds perfectly to dynamically changing loads." More important in power amps, of course, but... while "DC is DC," if the supply sags under load or changes behavior in response to transient demands, then it's also introducing "AC."

It's not a reach to consider the possibility that linear and SMPS might respond or behave differently in this regard, and therefore cause a difference in certain circuits.

Another thing to consider is any gear that has its own high-frequency stuff (clocks, etc) going on. I've heard (but not witnessed) one or two horror stories of certain digital guitar pedals that supposedly can have audible intermodulation between the clock frequency of the device and the switching frequency of a SMPS like the "one spot," etc.

 

[pucho812]

I have seen enough digital gear that has smpsu and not enough proper filtering causing all kinds of issues. For example an avid 192 does not have proper filtering which can be picked up on a scope. nothing I want in my audio if you know what I mean

 

[Brian Roth]

I have found switcher supplies to be less reliable than linears.  Not exactly a "sonic difference"...unless you include "no sound at all" <g>.

Bri

 

[ruffrecords]

I don't know about sonic differences but I tried them for a heater supply. Even the ones enclosed in metal frames seem to spew radiated interference all over the place and the conducted stuff is even worse. Most of them don't like large capacitive loads such as you get with well decoupled audio. I have tried them and now I would not touvh them for audio.

Cheers

Ian

 

[JohnRoberts]

i have been seeing a lot more audio products using SMPS these days...just wonder if anyone has ever done an A/B comparison between a linear PSU and SMPS with the same circuit...and if there is any audible difference?

There is actually an objective measurement you can perform, PSRR (power supply rejection ratio) that is exactly what is sounds like. Be sure to measure PSRR in the frequency range where the PS is making noise. Then is is a simple calculation to determine how much PS noise is showing up added to the audio path.

JR

 

[soapfoot]

On the subject-- I've heard many SMPS supplies actually acoustically radiate noise.  My current cell phone charger is one-- it radiates a high-pitched squeal that changes in amplitude (but not frequency) when a load is presented to it (i.e. when my phone is connected for charging versus the supply just plugged into the wall by itself).

The noise appears to have changed from when the supply was new (louder? lower pitch?). It still charges my phone just fine.

I don't know enough about switching supplies to be able to speculate on a cause or whether this is normal-- the noise level and pitch is similar to that generated by the picture tube of an old CRT television. My last phone charger did this as well, and I think I've observed it in a few other places, a few other times.

If something was making this noise in a control room (particularly multiple things), then it would absolutely drive me up the wall. 

 

[outoftune]

great information thank you for the replies!

 

[mjrippe]

Actually, I agree with everything that has been posted here and I personally avoid SMPS whenever possible.  I just wanted to clearly answer the original question and dispel any myth that the type of supply, correctly implemented, would have an effect on the sound 

 

[ruffrecords]

Despite everything I said about SMPSUs, back in the 70s we used one at Neve to generate 48V phantom power from the 24V desk supply. However, it was contained in a 100% screened box and was heavily filtered.

Cheers

Ian

 

[moamps]

Five tubes, how many SMPSs  8) ?

 

[JohnRoberts]

Bad design is bad design. A poorly executed switching power supply is not the fault of the switching supply. It can be a little more difficult to manage switching noise, but there are obvious benefits in reduced size and heat dissipation. A fairly simple calculus to perform if the benefit is worth the extra cost (noise filtering and attention to layout.)

There is more than one kind of switching supply. They don't all radiate magnetic fields from a storage inductor. I did a low current 48V step up supply for a 1U DSP mic input using a variant cap doubler/tripler but driven from a higher frequency square wave, so I could use tiny SMD caps. There are several variants on the cap based charge pump topology. While they can still make noise, a different kind of noise.

JR

 

[emrr]

I have used 48V switchers for phantom power near transformer coupled preamps, with no discernible noise contribution.  In practice many have lower output noise than a comparable linear, with the noise performance best at lower loads, opposite of many linears. 

 

[Brian Roth]

I know that SSL had a huge filtering box that was part of the SMPSU on their 9000 series consoles.  I found one mention of that unit, apparently filled with a bunch of L and C components, and that filter box is 9 RU in size and weighs 200 KG.

There is a new aftermarket SMPSU for the SSL 4000 desks, and the designer mentioned having a total of 100,000 uF of filtering caps as part of LC networks.

I guess if you throw enough filtering and shielding at the problem, noise becomes a non-issue.  Just hope the unit never gets hit by a large voltage surge from the AC mains.

Bri

 

[Rochey]

Many of the old SMPS's ran at low frequencies, some of which even switched at rates within the audio band. (hence the squealing). However, I see switching rates in the MHz range now... potentially more radiated than conducted noise, but much easier to filter out too.

 

[frans]

There was a comparison of sorts and a story about something like that a while back in Resolution mag - I'd have to sort through them to see which issue. If anybody is interested...?

 

[soapfoot]

I know that SSL had a huge filtering box that was part of the SMPSU on their 9000 series consoles.  I found one mention of that unit, apparently filled with a bunch of L and C components, and that filter box is 9 RU in size and weighs 200 KG.

Interesting-- particularly since one of the advantages of SMPS is usually said to be size and cost.

 

[JohnRoberts]

I know that SSL had a huge filtering box that was part of the SMPSU on their 9000 series consoles.  I found one mention of that unit, apparently filled with a bunch of L and C components, and that filter box is 9 RU in size and weighs 200 KG.

Interesting-- particularly since one of the advantages of SMPS is usually said to be size and cost.

It is... Near the end of my tenure managing a mixer engineering group one of the larger mid-sized mixer that was previously using an external power supply was able to build in a much smaller (and universal voltage) switching supply. Contrary to all the radiated noise fears from high edge rates, look at the magnetic circuit. Instead of 100 or 120 large current (magnetic flux)  pulses per second, there were tens or hundreds of thousands of tiny magnetic pulses per second. So smaller magnetic flux per pulse, smaller magnetic field. Sorry I don't recall any details only that SMPS were already delivering on their promises well over a decade ago. The power supply was several cubic inches and easily nested inside.  Back then it was that cheap, but the cost savings from eliminating the separate PS chassis, power cable, and hassle factor, was well worth it.

Now Peavey was no SSL, but SSL was no  Peavey...  8)

JR

PS: As i recall we had hired one full time switching supply expert and kept him busy on multiple projects. between mixers and power amp groups.

 

[3nity]

For my 1272s and 1290s preamps i have used 2 switching psus.
1 x 24V and another x 48V.

Never had a problem. the reason maybe cuz its 4 feet way on an external case!?

good luck.