Any one measured the EMI/RFI attenuation performance of audiophile power conditioners?


I was looking for a very good/robust power conditioner which will clean up reliably very noisy/dirty power supply that I have in my aprtment. While looking for one I went through catlogs of AudioQuest, Shunyata Research, Synergestic Research etc. but no one published charts showing attenuation performance over frequency range like you get for EMI filters from Schurter or Schaffner etc. which are in the industry for EMI/EMC compliance.

Since audio is very subjective, but contrary to audio reproduction Power and EMI/RFI reduction is completely objective and can be clearly demonstrated via attenuation charts.

Hence I am asking if anyone has measured the actual performance of these audiophile power conditioners. I am not denying someone saying they hear improvement after using XYZ product, but since I am talking about power conditioning or EMI reduction it's as objective as it can be.

I am not at all surprised to see all the manufacturers not publishing the performance data, else it would be used in other industries and research fields where it's far more critical and have far more stringent requirements on the performance of conditioner/EMI filters. But I am shocked to see even products ranging above 5-10K are following the same practice of not publishing the results.

Please note I am not a measurement fanatic, but I know where I can chase the measurements and where I can rely on my hearing to gauge the difference.




@tonywinga thanks for pointing it out. I wonder why they kept it in user manual where no prospective buyer is likely to look into. Though these are not exactly what I am looking for but better than not having anything.


Want to point out that series mode power conditioners are essentially low pass filters that start working ~ 3 kHz. True for ZeroSurge and Furman with SMP while AFAIK, RFI filters usually start in the 10s of kilohertz. This is why I always recommend them for noise filtering efficiency.

Parallel filtration has a difficult time filtering out lower than that due to the limitations on caps across the hot and neutral. YMMV.

Also, keep your dirty power supplies (computer, cheap wall warts) out of the clean side of a power conditioner and use shielded power cables to ensure you don't suffer re-infection of noise after the power has been cleaned up. 

Actually, I study the owners manuals online of products I am considering purchasing.  My brother says the owners manual/assembly instructions are just one engineer’s opinion.  My brother breaks things a lot.

Otherwise if they publish the data then people can easily point out the issues with their asking price and actual performance.

@audio_phool you answered your own question. While data might be useful to customers, it will not be beneficial to the manufacturers unless they have top/best measurements. In fact, it’ll create more work (accumulating data, publishing data, answering data related questions) and likely loss of sales/revenue to do so. Also, many claim to “lower noise”, but it may not be exactly/simply the same as specifically “EMF/RFI” like you are interpreting

@tonywinga Thank you for pointing out the user manual part. Now I have created a table of values for EMI/RFI Noise reduction based on available/published spec. Took me some time to gather all info. But I think its a worthwhile exercise.


Sr. No. Manufacturer Model EMI/Noise Reduction Mode Specified Load Specified Cost
1 Shunyata Everest More than 50dB, 68db @ 1 Mhz No No USD 9,900
2 Shunyata Gemini More than 40dB, 28dB @ 1 Mhz No No USD 1,998
3 Shunyata Venom PS10 More than 24dB, 28dB @ 1 Mhz No No USD 900
4 Audioquest Niagra 5000 More than 30db for CM & ore than 28 dB for DM Yes, CM & DM 10 to 50 ohms USD 5,900
5 Audioquest Niagra 7000 More than 30db for CM & ore than 28 dB for DM Yes, CM & DM 10 to 50 ohms USD 11,000
6 Audioquest PowerQuest 303 More than 22 dB for DM Yes, DM 10 to 50 ohms USD 460
7 Synergistic Research Galileo PowerCell SX No No No USD 27,995
8 Synergistic Research PowerCell SX No No No USD 9,000
9 Synergistic Research PowerCell ONE No No No USD 1,095
10 PS Audio Power Plant 20 More than 80 dB No No USD 8,000
11 PS Audio Power Plant 15 More than 80 dB No No USD 6,399
12 PS Audio Stellar Power Plant 3 No No No USD 2,399
13 IsoTek EVO3 Super Titan 85dB, No mention of frequency range No No USD 12,495
14 IsoTek V5 Titan 85dB, No mention of frequency range No 50 ohms USD 4,995
15 IsoTek V5 Elektra 55dB, No mention of frequency range No No USD 1,695
16 Puritan Audio PSM 156 No No No USD 2,400
17 Puritan Audio PSM 136 No No No USD 2,100
18 OnFilter AF Series Graph Provided for CM & DM. Max 98dB in DM, Max 62db in CM Yes, CM & DM 1:100 USD 650
19 Schaffner FN 2090 20A Graph Provided for CM & DM. Max 100dB in DM, Max 82db in CM Yes, CM & DM 50:50 USD 70


For any Power conditioner/Filter, it’s primary function is to reduce EMI/RFI noise on the mains line. Hence it’s objective performance is of high importance here. Now the performance of this noise filtering is evaluated based on the following parameters.

1. Noise reductionvalue:- The EMI Reduction value is expressed in the units of dB. Higher the value, better is the attenuation performance. Sometimes it’s also expressed as -ve dB value in that case lower the number better is the attenuation performance.
2. Frequency Range :- The performance of filter is not linear or constant over frequency range hence usually the graph of attenuation is plotted for range of frequency (Just like frequency response of speakers). For normal house hold scenario where noise is generated because of SMPS in LEDs, TV, fridges, washing machines etc, it’s frequency range usually lies around 100-150 KHz hence performance of any filter in this frequency range is crucial.

3. Mode :- EMI/RFI noise attenuation is measured in two modes Differntial Mode (DM) which is measured between live (hot) and neutral (cold) line (differential currents flow in opposite directions through the source and return path) and Common Mode (CM) which is about noise that flows in the same direction in a pair of lines (common mode currents flow in the same direction through the source and return path, completing the circuit through the ground path). Typically DM noise reduction values will be higher than CM mode values at a given frequency value & at the same load value. In both cases having higher values is desirable.
4. Load Values :- Typical Lab setting measurements are done with 50 ohms load and 50 ohms output impedance of power supply. These values often time look good (Schaffner publishes these lab values). But real world load and powerline impedances are anything but 50 ohms. Hence the realworld scenario would look like 0.1/100 where 0.1 ohm is the output impedance of powerline and 100 ohm is the load value or vice a versa 100/0.1. Schaffner removed the realworld performance data from their data sheets because it’s performance was worse in the area where it mattered the most viz. in 100Khz region.

Hence from above four points one can see what all needs to be looked at when looking for noise attenuation performance of any filter. So if any manufacturer specifies just a random number like PS audio for example, it doesn’t make any sense from the actual performance standpoint. Instead start asking for the hard facts based on these thingsfor the amount of money that we pay for these products.

@kennyc Yeah, now that you have pointed out, I can see that I had already answerd my own question. The table above puts hard facts around it.