Measuring line noise and power conditioners


I recently purchased a Trifield EMI (Dirty Electricity) Line Meter to measure noise coming from my outlets. To my surprise, my $500 power conditioner (name withheld to protect the potentially innocent) appears to not filter any noise per the Trifield readings. In fact, with some of my outlets the measures are higher through the conditioner’s outlets, than the measures coming straight out of the wall. The manufacturer denies anything is wrong with their conditioner, claiming the Trifield is measuring the wrong frequencies. Can anyone explain?

output555
This keeps being repeated as "factual", when it never has been. The output of a rectifier includes DC and high frequency components. Simple transfer function analysis. The more of those high frequencies you eliminate, the greater the ratio of DC to AC. Those high frequencies can extend dynamic range of the amplifier by providing a higher minimum voltage on the capacitors, but they will never contribute to lower noise. They will always contribute to more noise.

The circuits in the amplifier or other equipment may prefer the higher voltage being designed that way (bias values, etc), but the noise is never going to be less.

For emphasis, this is a simple one way transfer function, it is not loop analysis where high bandwidth can be used to cancel noise.
The spectral composition of that current draw therefore includes frequencies that are much higher than 60 Hz, and filtering those higher frequencies out will tend to adversely affect perceived dynamics and other sonic characteristics. Which is no doubt a major reason why many audiophiles prefer to plug power amps directly into the wall outlet.
In fact Ralph (Atmasphere) has stated in past threads that in many cases these desirable frequency components can range up to several tens of kHz, and I have no reason to doubt that. And if I recall correctly Shunyata has a paper in which measurements are presented supporting that conclusion.

Generally people don't think of frequencies in the audio band (<=20KHz) as EMI, but as line harmonics. 20Khz has a wavelength of 9 miles (15km), so you need a pretty big wire for a good antenna. These EMI toys don't measure harmonics. You need a good power meter for that. They of course don't give you the spectrum either.



Sorry for all the posts, but this thread is just a wealth of misinformation. Take his 1/2 truth from onfilter.com who is trying to do what? Sell you filters

Some points:  Most people add filters to their designed products not to reduce noise getting into them, but to reduce noise getting out of them to pass EMI. his part they got right.

The "standard" line impedance used for testing varies from 0.1 ohm to 0.4ohms, at low frequencies, which is not where EMI is tested. IEC60669 defines common line impedance for 230V as 0.2ohms and 400uH. North American standards for 120V are typically 0.2-0.4ohm, and 100uH - 400uh.  FCC conducted emissions testing begins at 150KHz. At 150KHz, 400uH = 60 ohms, at 100uH = 15ohms. That transformer outside your house is not presenting a 0.1ohm source impedance at 150KHz.  Now I am not complaining that OnFilter products don't do anything, likely they do, but they aren't being very honest with the information they are presenting. Now there are likely things in your house, i.e. motors, dimmers, etc. that because they are close have more impedance, but even 30 feet of Romex is 3-4 ohms at 150Khz.

The 50 ohms used in EMC measurement is more of a parasitic impedance, and unavoidable due to the input impedance of test equipment, and the passive setup. Technically in EMI/EMC measurement, the equipment is fully isolated from the AC and the measurement is volts.


Conventional filters will do little in actual power lines or may even amplify EMI - it is in their own specification (that is, for those manufacturers who bother to publish such specification). For a brief explanation please see this link: https://www.onfilter.com/real-life-filtering    In short, regular filters are designed to perform at 50 Ohms termination (in and out) for EMC Compliance - CE and FCC. I personally haven't met a power line with 50 Ohms impedance. In real-life applications a better impedance ratio is 1/100 or 0.1/100 (not a critical difference in reality) where 1 or 0.1 is output impedance of AC power and 100 is rough number for a load, i.e. your amplifier. It is imprecise but much more realistic than 50/50 Ohms. Since a filter is a combination of inductors and capacitors, when designed with one goal in mind to work in a 50/50 Ohms environment, this is where it "tuned" to. In actual use it either does nothing or amplifies noise. Our company (tooting my own horn here) designed filters for actual installations that we provide to the factories around the world, NASA, governments, hospitals, etc. - they are impedance-independent and essentially kill emissions anywhere they are plugged in -

The audience here simply doesn't know much about noise, including the people fronting noise reduction technology. The subject is codified in many ways by many international organizations. It is based on a knowledge of mathematics and of many other areas of engineering. Because we are supposed to be an organization of music lovers, add our group to the subject of "noise." It isn't nearly as codified and regulated, but still covers a wide area of human reaction to sound.
I am not going to say anything other than that "stupid is as stupid does." If you would like to really understand noise, it's measurement and appreciation, start with yourself. As a human being, interested in music... learn to appreciate the music, not the noise. The best filter I know is your mind. "If you don't mind, it doesn't matter."
While I don't agree with caycol in totality, a good portion of what he is stating is true at least for audio. In a well designed product, high frequency external RF noise is much easier to get rid of than low frequency noise, i.e. in the audio band on power lines, and guess what, low inductance, low capacitance AC cords are not going to help the situation, though consistency of the chassis ground reference between equipment can be critical for low noise.

High frequency RF as measured by these EMI toys does not represent well the much harder to filter frequencies on the AC line that are either in the audio band and/or close it and can sub-modulate down into the audio band.

Where the problems come in is with low level, high impedance signals such as interconnects, whose signal path may include the chassis ground due to intentional (or unintentional) capacitance between the signal ground and chassis ground (usually intentional).

piaudiol, most of the papers you linked have nothing to do with the topic specifically, which really doesn't address the effectiveness of line conditioners either where it matters, nor do they really refute caykol, though they do introduce topics such as ground loops, and hum, that my have nothing to do with EMI filtering at all.