• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

PS Audio Noise Harvester AC Cleaner Review

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
28,181
Likes
72,606
Location
Seattle Area
#1
This is a review and detailed measurements of the PS Audio Noise Harvester and AC Cleaner. It is on kind loan from a member and sells for $99.

The harvester is a bit bigger and uglier than typical phone charger:

PS Audio Noise Harvester Distortion and Noise AC Cleaner Review.jpg


Claim to fame is that AC noise is captured gradually and then discharged all of a sudden into the LED on the right to get rid of it:

PS Audio Noise Harvester Distortion and Noise AC Cleaner Advertising.png


One wonders why a sudden discharge wouldn't be bad for your sound versus a gradual/constant thing!

As to discharging "all the unwanted noise" company shows no specifications or measurements to demonstrate that. It is trivial to measure noise so not do that?

Also, I see no indication of safety regulation on the device and on the website. It is one thing to lack such marks for low voltage devices but an entirely different matter when you plug something in to mains and leave it there. A safety flaws could lead to fire or worse. What would this thing do for example if a powerful external surge arrives? Does it open up gracefully or short and cause smoke?

Does It Do Anything?
A lot of tweaks sold to audiophiles make claims that follow the laws of another universe so we can't use our tools in this universe to test them. Here though, we can. AC mains is at 50 or 60 Hz. A filter better not touch these frequencies or you would loose the very energy you need for our device to function! As a result, filtering will be at some higher frequency, often well above hearing range making them useless for improving audio.

To determine the filter range, I set my Audio Precision to produce 21 volts output and captured and plotted that using very wideband response (1 MHz). I then inserted the Harvester in the middle to see the effect:

PS Audio Noise Harvester AC cleaner frequency Response transfer function Measurement.png


Hey! There is some good news here. There is filtering that occurs in the audio band and indeed, we have reduction of frequencies above 500 Hz with best response (naturally) at higher frequencies. There is a resonance causing that peaking but otherwise, we have about 20 dB of reduction. The harvester was blinking away happily saying it was doing its job.

Would this do us some good? Well, let's look at the spectrum of AC in the power strip that I use to power my analyzer, the PC and anything I review. To not kill myself or the analyzer, I used a simple AC transformer to bring the voltage down. Otherwise everything is the same:

PS Audio Noise Harvester AC Cleaner Time Domain Mains FFT spectrum Measurement.png


In blue is the spectrum of the AC mains. Instead of the ideal single 60 Hz tone, we have a myriad of other tones. Highest peak is nearly -30 dB. This result in a THD+N of nearly 3%.

Once I plug in the Harvester, we get the graph in red. Yes, it almost looks identical because as I showed before, it doesn't do anything below 500 Hz. Unfortunately that spectrum between 60 and 500 Hz has most of our distortion and noise. What is above that has far lower amplitude.

Even in high frequencies we don't get much attenuation because the amount of energy harvested and dissipated by this box is too little. Imagine what it takes to power an LED for a fraction of a second. It is probably in microamps which would mean nothing in grand scheme of things. No wonder then that THD+N doesn't change (the number varies fair bit so don't focus on decimal places).

In addition to frequency domain, we can also look at the waveform in "time domain:"

PS Audio Noise Harvester AC Cleaner Time Domain Mains Scope Measurement.png


We can easily see the effects of 3% distortion in how the tops of the sine wave are distorted (flattened mostly). Sadly, the Harvester does nothing to turn our ugly sine waves into pretty ones. With or without, the shape is the same.

At this point we have fully characterized the device. It is a low pass filter with very negligible effectiveness. But maybe it performs some magic that escapes these measurements but shows up in audio devices so let's test them with and without the Harvester.

AC Cleaner Effect on Audio Performance
Let's start with the most sensitive and highest performance "audio" device I have which is my Audio Precision APx555 analyzer. I put it in the loopback mode and measured its own distortion+noise over time. After a bit of time, I plugged in the Harvester and then took it out. This is the results:

PS Audio Noise Harvester Distortion and Noise SINAD Measurement.png


As you see, it made no difference whatsoever even though we are measuring distortion and noise that is whopping -123.5 dB below full level! Why doesn't it care? Because it designed with the assumption that AC mains is never clean or remotely so. All the necessary filtering is performed internally which is a lot easier when dealing with low voltage DC, than high voltage mains AC.

OK, so that is a $28,000 analyzer so let's step down to a much more reasonable $500 Topping A90 which I recently reviewed:

PS Audio Noise Harvester Distortion and Noise SINAD Topping A90 Measurement.png


The harvester blinking away saying it is doing something. But again, in super sensitive measurements representing far lower threshold than our hearing, there is no effect on noise+distortion at all.

Let's go even cheaper to $99 JDS Labs Atom DAC:
PS Audio Noise Harvester Distortion and Noise JDS Labs Atom SINAD Measurement.png


We can see how sensitive our measurement system is in detecting very small variations in performance of Atom DAC. Clearly there is some "noise" there to clean up but the Harvester does nothing for us.

Maybe we shouldn't use high-performance audio devices. For that, let's use the Schiit Modi 2 Uber DAC:

PS Audio Noise Harvester Distortion and Noise Schiit Modi 2 Uber SINAD Measurement.png


Not... one... thing... different! Surely the Harvester should have gotten lucky by now and did something. But it did not.

Listening Tests
Folks always come back and say "but I hear a difference, you should use your ears." So that is what I did. On the same computer used for above, I use an RME ADI-2 DAC V2. Into that I use an Ether CX headphone. I plugged in and unplugged the Harvester multiple times but could not detect any difference whatsoever.

If you are hearing a difference, turn away from your power strip, have a loved one plug the device in and out at random times and have you raise your hand when you think it is plugged in. After just a minute or two, you will get the right answer which is there is no audible difference. Your brain concludes that there is less noise because you are told there is less noise with this device and hence you "hear it." Lack of change in audio waveform reaching your ears be damned.

Conclusions
In some sense it is a relief to find an audio tweak that we can analyzer and characterize what it does. It indeed filters "some noise." That filter is too small to begin with. But importantly is of no use because your good audio devices already filter that noise. Imagine how bad my audio system would sound if the 3% harmonic distortion and noise was let through! Which designer you cherish would sell you an audio device that doesn't do such filtering when it is so easy and cheap to do???

Remember, the company could easily measure noise and show the improvement with their own or other companies' products. But they have not. You wouldn't take a drug with no verification of its efficacy. Why would you trust a company to sell you something to "clean your AC" where no such evidence is not at all?

On top of that, you have a mains operated device which no third-party verification that is safe to be plugged in. As such, it is illegal to sell in probably every western countries if not others.

Be a smart audiophile. Don't put your guard down just because there are positive testimonials for such devices. I don't care how useless an audio device is. Some people will say it performs miracle. These people don't know how to perform a simple, bias controlled test to verify the truth. Don't listen to them. Don't put your guard down.

And before you say everyone who owns one likes it, watch this:


-----------
As always, questions, comments, recommendations, etc. are welcome.

June is turning out to be an expensive month. I will probably be close to shipping back 500 pounds of gear and have spent nearly $1,500 on gear. I can afford it but will feel entirely more positive if you all donate it! https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
 
Last edited:
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
28,181
Likes
72,606
Location
Seattle Area
Thread Starter #3
and

”At this point we have fully characterized the device. It is a high pass filter with very negligible effectiveness.“
A low pass?
Oops. Corrected.
 

GXAlan

Addicted to Fun and Learning
Forum Donor
Joined
Jan 15, 2020
Messages
506
Likes
701
#5
Just as I suspected. Piggy bank panther indeed. Thanks for testing it!

Here's the PS Audio quote about the product:
"A simple product designed to reduce power line noise above 10kHz. On its input there is a capacitor and a choke (inductor or magnetic device) placed across the power line, forming a simple bandpass filter. At the junction of these two elements energy is singled out at a specific frequency range. We then siphon off the noise energy and use it to fill up a capacitor. When the capacitor gets enough of the energy stored, it triggers an SCR which sends that energy to an LED and the light blinks; thus converting the noise energy to light. Bingo, the noise is removed from the line. Simple and effective removing about 10dB.

We could have made this product even simpler. Instead of all the fancy capacitor and LED arrangement, we could have just converted the noise energy into heat and it would be just as gone. But then, it makes it harder for people to see the results. Every time the Harvester’s light blinks you know noise is being removed. Positive feedback. Plus, it also helps find the best spot to place the Harvester by finding the point where the blinking is greatest. But in the end, we went with the light to help sell Harvesters knowing full well they would have been just as effective without the light."
1592285356787.png
 

rebbiputzmaker

Addicted to Fun and Learning
Joined
Jan 28, 2018
Messages
915
Likes
260
#6
This is a review and detailed measurements of the PS Audio Noise Harvester and AC Cleaner. It is on kind loan from a member and sells for $99.

The harvester is a bit bigger and uglier than typical phone charger:

View attachment 69147

Claim to fame is that AC noise is captured gradually and then discharged all of a sudden into the LED on the right to get rid of it:

View attachment 69148

One wonders why a sudden discharge wouldn't be bad for your sound versus a gradual/constant thing!

As to discharging "all the unwanted noise" company shows no specifications or measurements to demonstrate that. It is trivial to measure noise so not do that?

Also, I see no indication of safety regulation on the device and on the website. It is one thing to lack such marks for low voltage devices but an entirely different matter when you plug something in to mains and leave it there. A safety flaws could lead to fire or worse. What would this thing do for example if a powerful external surge arrives? Does it open up gracefully or short and cause smoke?

Does It Do Anything?
A lot of tweaks sold to audiophiles make claims that follow the laws of another universe so we can't use our tools in this universe to test them. Here though, we can. AC mains is at 50 or 60 Hz. A filter better not touch these frequencies or you would loose the very energy you need for our device to function! As a result, filtering will be at some higher frequency, often well above hearing range making them useless for improving audio.

To determine the filter range, I set my Audio Precision to produce 21 volts output and captured and plotted that using very wideband response (1 MHz). I then inserted the Harvester in the middle to see the effect:

View attachment 69149

Hey! There is some good news here. There is filtering that occurs in the audio band and indeed, we have reduction of frequencies above 500 Hz with best response (naturally) at higher frequencies. There is a resonance causing that peaking but otherwise, we have about 20 dB of reduction. The harvester was blinking away happily saying it was doing its job.

Would this do us some good? Well, let's look at the spectrum of AC in the power strip that I use to power my analyzer, the PC and anything I review. To not kill myself or the analyzer, I used a simple AC transformer to bring the voltage down. Otherwise everything is the same:

View attachment 69150

In blue is the spectrum of the AC mains. Instead of the ideal single 60 Hz tone, we have a myriad of other tones. Highest peak is nearly -30 dB. This result in a THD+N of nearly 3%.

Once I plug in the Harvester, we get the graph in red. Yes, it almost looks identical because as I showed before, it doesn't do anything below 500 Hz. Unfortunately that spectrum between 60 and 500 Hz has most of our distortion and noise. What is above that has far lower amplitude.

Even in high frequencies we don't get much attenuation because the amount of energy harvested and dissipated by this box is too little. Imagine what it takes to power an LED for a fraction of a second. It is probably in microamps which would mean nothing in grand scheme of things. No wonder then that THD+N doesn't change (the number varies fair bit so don't focus on decimal places).

In addition to frequency domain, we can also look at the waveform in "time domain:"

View attachment 69155

We can easily see the effects of 3% distortion in how the tops of the sine wave are distorted (flattened mostly). Sadly, the Harvester does nothing to turn our ugly sine waves into pretty ones. With or without, the shape is the same.

At this point we have fully characterized the device. It is a low pass filter with very negligible effectiveness. But maybe it performs some magic that escapes these measurements but shows up in audio devices so let's test them with and without the Harvester.

AC Cleaner Effect on Audio Performance
Let's start with the most sensitive and highest performance "audio" device I have which is my Audio Precision APx555 analyzer. I put it in the loopback mode and measured its own distortion+noise over time. After a bit of time, I plugged in the Harvester and then took it out. This is the results:

View attachment 69151

As you see, it made no difference whatsoever even though we are measuring distortion and noise that is whopping -123.5 dB below full level! Why doesn't it care? Because it designed with the assumption that AC mains is never clean or remotely so. All the necessary filtering is performed internally which is a lot easier when dealing with low voltage DC, than high voltage mains AC.

OK, so that is a $28,000 analyzer so let's step down to a much more reasonable $500 Topping A90 which I recently reviewed:

View attachment 69152

The harvester blinking away saying it is doing something. But again, in super sensitive measurements representing far lower threshold than our hearing, there is no effect on noise+distortion at all.

Let's go even cheaper to $99 JDS Labs Atom DAC:
View attachment 69153

We can see how sensitive our measurement system is in detecting very small variations in performance of Atom DAC. Clearly there is some "noise" there to clean up but the Harvester does nothing for us.

Maybe we shouldn't use high-performance audio devices. For that, let's use the Schiit Modi 2 Uber DAC:

View attachment 69154

Not... one... thing... different! Surely the Harvester should have gotten lucky by now and did something. But it did not.

Listening Tests
Folks always come back and say "but I hear a difference, you should use your ears." So that is what I did. On the same computer used for above, I use an RME ADI-2 DAC V2. Into that I use an Ether CX headphone. I plugged in and unplugged the Harvester multiple times but could not detect any difference whatsoever.

If you are hearing a difference, turn away from your power strip, have a loved one plug the device in and out at random times and have you raise your hand when you think it is plugged in. After just a minute or two, you will get the right answer which is there is no audible difference. Your brain concludes that there is less noise because you are told there is less noise with this device and hence you "hear it." Lack of change in audio waveform reaching your ears be damned.

Conclusions
In some sense it is a relief to find an audio tweak that we can analyzer and characterize what it does. It indeed filters "some noise." That filter is too small to begin with. But importantly is of no use because your good audio devices already filter that noise. Imagine how bad my audio system would sound if the 3% harmonic distortion and noise was let through! Which designer you cherish would sell you an audio device that doesn't do such filtering when it is so easy and cheap to do???

Remember, the company could easily measure noise and show the improvement with their own or other companies' products. But they have not. You wouldn't take a drug with no verification of its efficacy. Why would you trust a company to sell you something to "clean your AC" where no such evidence is not at all?

On top of that, you have a mains operated device which no third-party verification that is safe to be plugged in. As such, it is illegal to sell in probably every western countries if not others.

Be a smart audiophile. Don't put your guard down just because there are positive testimonials for such devices. I don't care how useless an audio device is. Some people will say it performs miracle. These people don't know how to perform a simple, bias controlled test to verify the truth. Don't listen to them. Don't put your guard down.

And before you say everyone who owns one likes it, watch this:


-----------
As always, questions, comments, recommendations, etc. are welcome.

June is turning out to be an expensive month. I will probably be close to shipping back 500 pounds of gear and have spent nearly $1,500 on gear. I can afford it but will feel entirely more positive if you all donate it! https://www.audiosciencereview.com/forum/index.php?threads/how-to-support-audio-science-review.8150/
Maybe you could try playing some songs with explicit language and see if it cleans them up?
:eek:
 

P_M

Member
Joined
May 31, 2019
Messages
57
Likes
16
Location
San Francisco Bay Area, CA
#10
What might be the big spike around 65Khz ? I am also curious to see what might be out there in the 100Khz+ range. Thats where most of the smps noise is located. What is the max sampling frequency you can use ?

Also, I'd suggest a drop down resistor rather than a transformer - for two reasons. One is that the transformer itself would introduce its distortions and secondly it will filter out the high frequency content. A pure resistor based voltage drop down network would be best. Be careful though when working directly with mains.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
28,181
Likes
72,606
Location
Seattle Area
Thread Starter #13
Also, I'd suggest a drop down resistor rather than a transformer - for two reasons. One is that the transformer itself would introduce its distortions and secondly it will filter out the high frequency content. A pure resistor based voltage drop down network would be best.
A transformer exists in most audio gear so what I am measuring is more representative. I have a 20 Mhz differential probe but have misplaced it. If i find it, I will pos spectrum using that.

Be careful though when working directly with mains.
:( I am an electrical engineer and have worked on thousands of pieces of electronics. I have been shocked more ways than I can count. Please don't give me advice like this. It is hard to not be offended.
 
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
28,181
Likes
72,606
Location
Seattle Area
Thread Starter #15
Fascinating how low the fidelity of the mains electricity is... Would be interesting to see if there are differences with other regions.
To keep ground loops low, a bunch of stuff is plugged into this one strip. There is also stray pick up as I am not trying to perform a high-precision measurement.

But yes, I have actually seen glitches in the waveform in addition to deformity!

Despite all this noise, we are measuring audio devices that are instrument grade, showing how there is no problem to be solved by this box.
 

solderdude

Major Contributor
Joined
Jul 21, 2018
Messages
6,177
Likes
11,316
Location
The Neverlands
#18
When do you guys learn... you have to listen and hear the improvements... not look at graphs !

We can say many things about it but their claim is it lowers noise above 10kHz and it seems to do just that.
In fact, assuming the intenal resistance of the mains outlet it is connected to is high enough for it to actually work (depends on this as it is a parallel device) it may even exceed the claimed 10dB improvement. It stands to reason that for frequencies above 10kHz the mains 'impedance' is high enough given the long wires involved and street transformers limited frequenct range.

That it doesn't improve things in real life with the vast majority of audio products is another matter. It won't 'improve' SQ though. That part is the snake-oil BS. Claimed technical functionality not so it seems.

question to @amirm did you, during the mains tests, see the LED blink ?
It could be a useful device to see if mains is polluted with lots of 'spikes'.

Good to see it measured !

Edit:
Scope plots of injected >10khz spike (using cable clamp or coupling capacitor to mains) may have been useful as well to show effects it has and may have even shown up on averaged (over time) noise analysis plots.
Of course .. this device only does something for differential mode peaks. Most 'spikes' however, are common mode and it does absolutely nothing against those spikes.
 
Last edited:
OP
amirm

amirm

Founder/Admin
Staff Member
CFO (Chief Fun Officer)
Joined
Feb 13, 2016
Messages
28,181
Likes
72,606
Location
Seattle Area
Thread Starter #19
question to @amirm did you, during the mains tests, see the LED blink ?
Yes. It blinks like mad. Did so during my AP sweeps as well.
 

digicidal

Major Contributor
Joined
Jul 6, 2019
Messages
1,305
Likes
1,912
Location
Sin City, NV
#20
Of course .. this device only does something for differential mode peaks. Most 'spikes' however, are common mode and it does absolutely nothing against those spikes.
I disagree... even with common mode noise/aberrations it still deposits nearly $100 of profit into Paul's bank account... that's something. :p
 
Top Bottom