Can a Linear Power Supply Inject Noise Into its AC Circuit?

[Ron Resnick]

A question for people who use linear power supplies for servers or streamers or other audio components . . .

Do you have any concern that your linear power supply is injecting or propagating noise or some other EMI nasties to the AC circuit into which the linear power supply is plugged?

Do you try to isolate in some way your linear power supply from your audio components?

Do you plug your linear power supply into a different circuit and the one into which you plug your phone in stage or DAC or preamp?

 

[John T]

A question for people who use linear power supplies for servers or streamers or other audio components . . .

Do you have any concern that your linear power supply is injecting or propagating noise or some other EMI nasties to the AC circuit into which the linear power supply is plugged?

Do you try to isolate in some way your linear power supply from your audio components?

Do you plug your linear power supply into a different circuit and the one into which you plug your phone in stage or DAC or preamp?

Hi Ron, Long time no talk! I think our last conversation was in regards to the 6 Pack Array, I do hope this finds you well: Interesting Question, I keep my LPS separate on a dedicated circuit, away from all other Audio Circuits including the 5 that feed my Torus WM 100. My LPS powers the Taiko Switch and my Router. I have had noise introduced via the LPS when sharing a circuit. I could easily experiment and try the LPS in a feeder that goes into the Torus, but things are sounding so good I just leave well enough alone! I could quite simply have the Horizon/Extreme share the same circuit which would free up one circuit but I've come to the realization if it sounds good "leave it alone"...

 

[John T]

Hi Ron, Long time no talk! I think our last conversation was in regards to the 6 Pack Array, I do hope this finds you well: Interesting Question, I keep my LPS separate on a dedicated circuit, away from all other Audio Circuits including the 5 that feed my Torus WM 100. My LPS powers the Taiko Switch and my Router. I have had noise introduced via the LPS when sharing a circuit. I could easily experiment and try the LPS in a feeder that goes into the Torus, but things are sounding so good I just leave well enough alone! I could quite simply have the Horizon/Extreme share the same circuit which would free up one circuit but I've come to the realization if it sounds good "leave it alone"...

 

[Ron Resnick]

Hi John!

Thank you for this report!

 

[Puma Cat]

I think it’s important to understand that not just the LPS for digital-domain devices, but the power supplies inside our “analog” amplification components themselves can, and usually do, contribute a significant amount of noise to our stereo systems. A principal contributor of noise are the full-wave bridge rectifiers that comprise the power supplies of almost all analog amplification components. So, the answer to the question: “Can a linear power supply inject noise into its AC circuit?” is: YES.

A common misconception I find many audiophiles have is thinking that the AC “flows” only in ONE direction from their wall receptacles to their amplification components, much like water from a hose. This is simply not the case. It’s AC: Alternating Current, so the current flows INTO and back OUT of your amplification components at…60X/second (in the USA).

Caelin Gabriel of Shunyata Research has a very accurate description of what is actually happening here:

Introduction: “There are a lot of misconceptions about power transmission and power quality that make it difficult for some people to understand why a power cord makes a sonic difference. The first question is – do power cords make any difference at all? There is no sense in talking about theories of operation if we can't agree that there is an audible effect. Most of the thousands of people that use our power cords started out as skeptics and have answered that question for themselves and have found that power cables and power conditioners can have a profound impact on performance. And no - I do not care to debate with people that have not done the simplest of tests about whether power cords work or not. The only cases where a high quality cord does not have significant effects is when it is used with a poor quality power conditioner that acts as a high impedance to instantaneous current flow. “
Misconception #1: AC Power is like water coming from a large power tank, flowing through several 10s of feet of power hose into a component. This implies that the component is at the end of this system.

Answer: “Actually, the component sits between two power conductors: the hot and the neutral. AC power oscillates (alternates) back and forth at a 50-60 Hz rate. So power does not pour into the component at all. The component's power supply is within a complex network of wires and connectors. ALL of the wire and connectors can and do affect the performance of the component's power supply.”

Misconception #2: AC power can be contaminated just like water in a hose. This implies that once the water is contaminated at some point up stream, that is must be cleansed before it arrives at the audio component.

Answer: “As stated in #1, the component is not at the end of the power hose. It is between two power hoses and the current is oscillating back and forth. Further, current is not like water at all. Electrons cannot be contaminated. There are two aspects to power transmission: the electromagnetic wave and the current flow. The current itself cannot be contaminated but the electromagnetic wave can be modulated with other frequencies. We usually call these other frequencies noise or Electromagnetic Interference (EMI). Within the various parts of a power circuit there may be EMI in certain parts that is not present in others. Electromagnetic energy can be transformed or redirected to lessen their effects. "Some power cords use capacitors, inductors, or ferrites in an attempt to control the electromagnetic fields around the audio component. The success of such an approach is completely dependent upon the specific design and the reactance of the power supply of the component to which the power cable is attached.”

Misconception #3: There is up to a hundred feet of wire in the walls, so the last 6 feet of power cord can't possibly make any difference.

Answer: “The power cord is not the last 6 feet, it is the first 6 feet from the perspective of the component. As stated in #1 the local current and electromagnetic effects directly affect the sonic performance of the component.”

Misconception #4: There is a tremendous amount of electrical interference and EMI coming from outside the home that we need to protect our equipment from. This implies that we need some sort of power conditioner or filter to protect the equipment.

Answer: “Most of the EMI that affects the audio quality of a system is generated by the audio components themselves. Electromagnetic waves that traveling through space dissipate in power at the square of the distance from the source. Further, very high frequencies that propagate through the power circuit do not survive for long. Power lines present a high impedance to MHz and GHz signals due to the relatively high inductance of power lines. "A primary source of audible sonic degradation is caused by the power supplies in our audio/video components. Most components use FWBR (full wave bridge rectifier) power supplies that generate an incredible amount of transient noise when the rectifiers switch off. The design of a power cable can significantly affect the reactance of these signals within the power supply. The power cable is effectively part of the primary winding of the power transformer. The transition between the various metals used in a power cable and its connectors can cause electromagnetic reflections and diode-like rectification of the noise impulses as they propagate away from the power supply. If the power cable presents a high impedance to these signals they will be reflected back into the power supply where they will intermodulate, thus increasing the high frequency noise levels of the component. Most power supply filters are ineffective at blocking very high frequency noise components and much of it is passed through to the DC rails. The sonic effects of this include: high background noise levels, blurred or slurred transients and a general lack of clarity and purity of the sound or visual image.”

Misconception #5: There is some sort of conspiracy among audio designers that keeps them from producing a "proper" power supply that is not affect by power cable quality.

Answer: “This concept is like saying that if a speaker where properly designed, you wouldn't need to use a good quality speaker cable. PowerSnakes have been tested with the most modest of mid-fi equipment and the most exotic state of the art components. We have yet to find a component that cannot be improved by replacing the power cord.
"As long as power supply design is based upon FWBRs or switching supplies, the power cord will always be significant.”

Misconception #6: High-end power cords just increase the circuit capacitance acting as a high-frequency shunt
Answer: “There are some power cords that ARE designed this way. Some even insert capacitors within the cable to further increase capacitance. This approach has some positives and many negatives including the reactive interference with the way many power supplies are designed.
"Capacitance alone cannot account for the differences in a power cord's performance. There are some high-end power cords that are very effective that have virtually immeasurable levels of capacitance. These cables are usually designed around hollow tubes with the conductors inside. The conductors are several inches apart and cannot significantly affect the capacitance of the power circuit.”

Another class of noise that many folks are not aware of, particularly with respect to digital-content playback is low- and high-source leakage impedance current, which often is produced by most SMPS. The problem with these noise factors is they can produce threshold jitter, which results in timing errors during playback of digital content. And, our brains are sensitive to timing errors in the picosecond time-domain, which explains why we need femtoclocks in our digital music playback devices.

John Swenson, the Ethernet engineer who designed EtherREGEN discusses this in depth here:
"Understanding how perturbations on digital signals can affect sound quality without changing bits, and how these issues are addressed by the UpTone EtherREGEN."

The bottom-line here is that noise is everywhere in our systems, and as such, we need to use a systematic (no pun intended) approach to ameliorating its effects. While we will never be able to completely remove noise (this impossible from a physics perspective), the more systematically we address all the various sources of noise in our system, the more engaging and enjoyable our systems will be.

 

[DonH50]

Yes. A "linear" supply still has a bunch of nonlinear circuit elements (like diodes) inside and will create current spikes (and thus potentially voltage spikes) on the line. Of course, the power supplies inside your components do the same thing, and using a LPS on a streamer or other digital processor strikes me a pretty worthless. Any supply (linear or switching) can generate noise, and create leakage paths that add hum, but a decently designed component should be immune to such things by isolating itself from the wall source and power supply. That said, not an insignificant number of reviews show 50/60/100/120 Hz spikes due to such coupling, usually internal to the device.

One big advantage of switching supplies is that their noise is way above the audio band. The counter is that poorly designed and/or shielded components may provide a path for HF noise to be mixed or rectified back to the audio band, although that means they would also have a problem with AM radio interference and a host of other things.

As usual the argument is how much leakage is present and how much harm can be done due to a component that does not reject it as it should.

 

[Onepoint5]

An informative white paper on the differences between switchmodes and linear supplies are from Acopian.

 

[Puma Cat]

Yes. A "linear" supply still has a bunch of nonlinear circuit elements (like diodes) inside and will create current spikes (and thus potentially voltage spikes) on the line. Of course, the power supplies inside your components do the same thing, and using a LPS on a streamer or other digital processor strikes me a pretty worthless. Any supply (linear or switching) can generate noise, and create leakage paths that add hum, but a decently designed component should be immune to such things by isolating itself from the wall source and power supply. That said, not an insignificant number of reviews show 50/60/100/120 Hz spikes due to such coupling, usually internal to the device.

One big advantage of switching supplies is that their noise is way above the audio band. The counter is that poorly designed and/or shielded components may provide a path for HF noise to be mixed or rectified back to the audio band, although that means they would also have a problem with AM radio interference and a host of other things.

As usual the argument is how much leakage is present and how much harm can be done due to a component that does not reject it as it should.

and

An informative white paper on the differences between switchmodes and linear supplies are from Acopian.

Thanks, gents, excellent information. I'm going to read through this info as well. More data for the "database" is always a good thing. Cheers.

 

[Folsom]

100% of power supplies make some noise. But the average SMPS makes extensively more than an LPS. That is why people use LPS over SMPS typically. You can make SMPS more quiet but the differences are still not always a good thing, leaving LPS in favor. Both can reduce noise in several ways, but SMPS has the most egregious types.

This topic is weird because it is about LPS when it should be about SMPS. Ron are you messing with us?

 

[jeromelang]

A linear power supply placed near the power amplifier can affect the whole system's soundstaging.

 

[luca.pelliccioli]

I’m confused. Thinking about a router, that is probably the noisiest component. I feed it by a Dr.Sean Jacobs custom made linear power supply.
1. The PSU plugged into my main conditioner (being a P20 or a Puritan 156) together with other digital sources, on a dedicated line. It might pollute the system via the AC path shared.
2. The PSU plugged into a different socket (non dedicated line, separated from the system) without sharing AC with other sources should be a cleaner solution, even if it might be prone to introduce a less clean AC and it consequently might pollute the system via the Ethernet path.

So, which is the less polluting solution?

 

[Puma Cat]

I’m confused. Thinking about a router, that is probably the noisiest component. I feed it by a Dr.Sean Jacobs custom made linear power supply.
1. The PSU plugged into my main conditioner (being a P20 or a Puritan 156) together with other digital sources, on a dedicated line. It might pollute the system via the AC path shared.
2. The PSU plugged into a different socket (non dedicated line, separated from the system) without sharing AC with other sources should be a cleaner solution, even if it might be prone to introduce a less clean AC and it consequently might pollute the system via the Ethernet path.

So, which is the less polluting solution

My hypothesis is #2 would. But, why not "be like a scientist" and do a "controlled experiment" to find out?

Try both, and listen for what sounds best. Whatever you like best...go with that.

 

[bazelio]

Of course, the power supplies inside your components do the same thing, and using a LPS on a streamer or other digital processor strikes me a pretty worthless.

Yeah, I hear you. I thought the same until I heard the difference. An LPS on a SFP router, streamer, and even DDC makes an easily audible difference. Not always an improvement, mind you, but a difference nonetheless. I preferred the stock switching supply on my router. But in other positions, I preferred a LPS.

 

[luca.pelliccioli]

To be honest option #1 sounds better, plugging the PSU into the Puritan 156. More lifelike, clean and detailed. It’s not a night and day difference but audible enough to discern how I prefer between the two.

 

[sparkie]

A question for people who use linear power supplies for servers or streamers or other audio components . . .

Do you have any concern that your linear power supply is injecting or propagating noise or some other EMI nasties to the AC circuit into which the linear power supply is plugged?

Do you try to isolate in some way your linear power supply from your audio components?

Do you plug your linear power supply into a different circuit and the one into which you plug your phone in stage or DAC or preamp?

Linear power supplies solves noise issues for things that are connected to the computer, like hard drive noise an other computer noises. But it doesn't fix ground loops. Ground loops happen regardless of power supply type. My computer that runs a linear is a studio computer with DSP cards and the linear supply fixed certain jitter issues I was observing as well as mouse and hard drive noises. My home system, I run on a NAS with computers in each room that can run its each own player, but the audio is all transported by Dante instead of using usb connected DACs.

But devices like a phone, I would blue tooth into a DI box now these days because their DAC isn't that good and you have to get rid of dc offset voltages the phone jack will have. This is the DI box I use with my phone. https://www.radialeng.com/product/bt-pro-v2

 

[Adimon]

After years of using an LPS to power my switches (because I thought they were cleaner), I recently switched back to a decent SMPS.
I was shocked by the improvement in SQ.
It's like waking up from hibernation.
I was triggered by reading experiences and the comments of Jaap Veenstra of Alpha Audio.
For now, I think using an LPS for a switch or router is not a good idea, these devices need the speed of a good SMPS.

 

[audiobomber]

After years of using an LPS to power my switches (because I thought they were cleaner), I recently switched back to a decent SMPS.
I was shocked by the improvement in SQ.
It's like waking up from hibernation.
I was triggered by reading experiences and the comments of Jaap Veenstra of Alpha Audio.
For now, I think using an LPS for a switch or router is not a good idea, these devices need the speed of a good SMPS.

I know Hans Beekhuyzen holds the same opinion, that pure digital gear like routers and switches are best used with a switch mode supply, but IMO it's not that simple.

I own iPower 5v and 12V and an iPower X 9V. I have tried these with various renderers, DAC's and switches in my four network audio systems. In every setup, they were beaten by a comparably priced Zero-Zone linear supply or my custom Welborne Labs PS-REG.

The Silent Angel N8 switch ships with a medical grade SMPS. I replaced it with the Welborne Labs PS-REG and heard a major improvement. My exaSound PlayPoint came with a 12V, three prong SMPS. I was encouraged by the manufacturer to upgrade the e32 DAC with a Teddy Pardo supply, but he said the PlayPoint would not benefit from an LPS. I replaced the PlayPoint PSU anyway, with a Zero-Zone 2020 SS 4A, and heard cleaner high frequencies and improved sibilants.

I borrowed a friend's iPower Elite for a few days. I tried it versus a Teddy Pardo 12/2 on an EtherREGEN, a Zero-Zone 2020 SS on an exaSound PlayPoint, and a Pardo MiniTeddy on an exaSound e32 II DAC. In each case the LPS was better, with a larger soundstage, more bass power and improved listenability.

The friend who owns the iPower Elite uses it on his Netgear 7000 router and found it sounds better than my Zero-Zone 2020 in that situation. My conclusion is that there is no definitive answer, whether linear or switching sounds best depends on which PSU, and what it's feeding.

PS I ground two-prong switch mode supplies, as recommended by John Swenson:

SMPS and grounding

Over the last month I have been performing extensive tests on leakage currents coming from SMPS power supplies. One of the most important results from this investigations is the makeup of this leakage current. Previously my understanding was that leakage current was mainly a low frequency phenoni...

audiophilestyle.com

 

[Adimon]

I know Hans Beekhuyzen holds the same opinion, that pure digital gear like routers and switches are best used with a switch mode supply, but IMO it's not that simple.

I own iPower 5v and 12V and an iPower X 9V. I have tried these with various renderers, DAC's and switches in my four network audio systems. In every setup, they were beaten by a comparatively priced Zero-Zone linear supply or my custom Welborne Labs PS-REG.

The Silent Angel N8 switch ships with a medical grade SMPS. I replaced it with the Welborne Labs PS-REG and heard a major improvement. My exaSound PlayPoint came with a 12V, three prong SMPS. I was encouraged by the manufacturer to upgrade the e32 DAC with a Teddy Pardo supply, but he said the PlayPoint would not benefit from an LPS. I replaced the PlayPoint PSU anyway, with a Zero-Zone 2020 SS 4A, and heard cleaner high frequencies and improved sibilants.

I borrowed a friend's iPower Elite for a few days. I tried it versus a Teddy Pardo 12/2 on an EtherREGEN, a Zero-Zone 2020 SS on an exaSound PlayPoint, and a Pardo MiniTeddy on an exaSound e32 II DAC. In each case the LPS was better, with a larger soundstage, more bass power and improved listenability.

The friend who owns the iPower Elite uses it on his Netgear 7000 router and found it sounds better than my Zero-Zone 2020 in that situation. My conclusion is that there is no definitive answer, whether linear or switching sounds best depends on which PSU, and what it's feeding.

PS I ground two-prong switch mode supplies, as recommended by John Swenson:

SMPS and grounding

Over the last month I have been performing extensive tests on leakage currents coming from SMPS power supplies. One of the most important results from this investigations is the makeup of this leakage current. Previously my understanding was that leakage current was mainly a low frequency phenoni...

audiophilestyle.com

There are many variables and not everyone experiences the same thing.
That keeps the hobby alive.
The power supplies I use are DIY.
The LPS build-up is similar to the Zero Zone.
For the SMPS, I used MW IRM-30 modules with primary and secondary filtering.
In both situations, the housing is earthed and this in turn is connected to the switch housing
In my situatioin the difference is rather big.
LPS give smoothnes but i miss details.
SMPS give a more direct sound with a lot of detail.

 

[audiobomber]

I forgot to mention one of my tests. I use a miniDSP digital crossover for my subwoofers only. Powering it with an iPower2 SMPS generated tighter bass than a Topping P50 LPS, which reinforces my belief that "it depends".