AudioQuest NRG-X3 Review (AC Power Cord)

[Ingenieur]

I knew a lot of engineers were undercover hookers! How else do they afford all those Corvettes and such? LOL

Those that prostitute engineering are not engineers.,.they are wh@$se. Lol

 

[Spkrdctr]

A car charger or induction cooker is not the same as an amplifier. At least not for most people.

You don't say! We all know that an amp or receiver is a much harder load for the residential pole transformer. Those pole transformers just can't handle serious bass slam. They pop like popcorn from what my audiophile friends tell me. Have to have the bigger pole transformers (they have upsized versions) and the extra thick wire from there to the house panel, then step down the wire size to 8ga to the wall outlet which is then necked down to the outlet screws. Then it finally goes into the P-12 Regenerator! Now that is true audiophile wizardry right there.....

 

[CinDyment]

Their base was 0.775 V / 20 V so ~28 dB sounds right.
They should have measured the 60 Hz power, not amplified noise, lol

Notice anything strange about this one?
Conductor area?

Since R (~1/Area) the larger area 0.5 vs 0.2 gives:
~ 0.5/0.2 ~ 2.5 x the R
At 10 kHz ~ 40 vs 90, ~ 2.5x

But obviously L is at play too.


View attachment 196517

Well technically the graph says resistance, so that can only include changes in skin resistance .... though they probably mean impedance. However, I have no clue why I would want low impedance at high frequencies between line and neutral. Why do I want a big conduit for EMI? Perhaps low Neutral/GND impedance could have impact. Line/Neutral, nope.

 

[CinDyment]

We started in the street. The 300A 4-core cable was about $40/m and puts any audiophile cable to shame. No fancy braiding though and it didn't come in a fancy box. Not recommended for tonearms. And to answer your question, we did not rewire the National Grid! The good news it's copper, not aluminium that used to be used some places.

I have no issue with copper clad aluminum of a reasonable gauge to keep resistance the same as copper. Less weight.

 

[CinDyment]

. I have much longer runs to my AV and modem, for which I used Belden 19364.

Not understanding. You made your own power cords? Extension cords? .... not clear.

 

[Killingbeans]

But did you replace the wiring from your house to the substation? Did you replace the substation transformers with those wound with a high-end-approved, 99.99999999% OFC, cryogenically treated, crystal-aligned cable? Until you do all that, and also replace all the cables coming into the substation, you're not getting the best possible sound from your system, believe me!

Not good enough! I want audiophile fuel rods in the nuclear power plants! I refuse to listen to all of that distortion riddled steam pushing on the turbines.

 

[xaviescacs]

Obviously
We don't know if one is always above the other. They may have:
Taken measurements differently
Different days
Taken more than 1 and picked the one that favored them
Music at the same level
Voltage to amp was the same

The variation is random and likely within the margin of error at that low of a level.
It is a contrived measurement, extracted from a music signal.

Ok, this is going too far for what it is, but I'll try to clarify what I meant.

This is the plot.

QED's people say it's the output of an amplifier. They don't say which, they don't say anything else. This plot can be the result of many things, including being drawn by hand, who knows. But what it can't be is one random outcome of two identical distributions.

Some formalism. Usually the noise is modeled as a normal distribution with 0 expected value added to the underlying distribution or signal, to each point. This is what is called a random noise model, like in regression for instance. Furthermore, the noise of each point is independent from the others and usually it's also homoskedastic, meaning that the variance of each point is the same, although in some situations it may be convenient to add some variability on the variance. This models predicts quite well the typical situation in which we see tiny ups and downs everywhere that don't go too far in any case.

Given this setup, and if we assume that the underlying distribution of the data is the same, which is the same as saying there were measuring the same device in the same conditions, the outcome depicted in the plot, with one of the curves always above the other, is extremely improbable, it can't happen in practice.

In summary, this plot may the the result of anything we can imagine, but not a random outcome of two identical measurements where the only thing that differentiate them is an independent random noise. The differences shown are not compatible with any randomness. To be that different, the underlying distributions, if we keep modeling this as a probability problem, must be different: different gain, different set up, different amplifier, different whatever.

 

[Ingenieur]

Ok, this is going too far for what it is, but I'll try to clarify what I meant.

This is the plot.

View attachment 196532

QED's people say it's the output of an amplifier. They don't say which, they don't say anything else. This plot can be the result of many things, including being drawn by hand, who knows. But what it can't be is one random outcome of two identical distributions.

Some formalism. Usually the noise is modeled as a normal distribution with 0 expected value added to the underlying distribution or signal, to each point. This is what is called a random noise model, like in regression for instance. Furthermore, the noise of each point is independent from the others and usually it's also homoskedastic, meaning that the variance of each point is the same, although in some situations it may be convenient to add some variability on the variance. This models predicts quite well the typical situation in which we see tiny ups and downs everywhere that don't go too far in any case.

Given this setup, and if we assume that the underlying distribution of the data is the same, which is the same as saying there were measuring the same device in the same conditions, the outcome depicted in the plot, with one of the curves always above the other, is extremely improbable, it can't happen in practice.

In summary, this plot may the the result of anything we can imagine, but not a random outcome of two identical measurements where the only thing that differentiate them is an independent random noise. The differences shown are not compatible with any randomness. To be that different, the underlying distributions, if we keep modeling this as a probability problem, must be different: different gain, different set up, different amplifier, different whatever.

That is all we know. Nothing else, only what they want us to see. You assume it is done identically under controlled lab conditions.
I do not.

A 0.1 mV signal level difference in the signal could account for the difference.

Both traces are random noise generated by the system. It is generated by non-linearities, RFI, thermal noise, etc.
Random. Extracted from the signal by 'nulling' I assume.

It looks 'rigged'. A power cable (any cable) will not eliminate distortion or noise. Only absorb less. It may attenuate THD harmonics (L) but to no relevant degree.

What they are implying is:
The power cord reduced noise in the AC power to the amp prior to the xfmr, caps and rectifier. So somehow that was carried to the DC rail and thru the transistors modulated by the music.
Any line noise of that freq would be attenuated by the xfmr and caps.
If that 0.0005 uV was on a 2.83 V/8 Ohm signal (1 W) that is a SINAD of 0.0014%.
It includes THD and noise.
It is nonsense.

Please stop explaining signal theory and measurement methods. I am well versed in the subjects.

In the R testing they compared a 0.2 mm^2 vs a 0.5 mm^2 and said theirs was better because it had lower resistance! Lol

 

[DWI]

Not understanding. You made your own power cords? Extension cords? .... not clear.

Use it from the consumer unit instead of standard 3-core.

 

[Ingenieur]

Well technically the graph says resistance, so that can only include changes in skin resistance .... though they probably mean impedance. However, I have no clue why I would want low impedance at high frequencies between line and neutral. Why do I want a big conduit for EMI? Perhaps low Neutral/GND impedance could have impact. Line/Neutral, nope.

They said impedance in the text.

The skin effect factor is ~ k sqrt(freq/Ro) and is linear with frequency.

In this case with a ratio of R ~ 2.5 that is a 1.6 delta or change at 20,000 Hz. ... with the smaller being LESS impacted.

Rough numbers

Why do we care about R/Z above 1000 Hz on a 50 or 60 Hz system?

 

[pras1011]

Because of these reviews I have sold off my power conditioner and I am selling my exotic power cables.

 

[egellings]

The peakiness shows up in the same places in both waveforms--doesn't look completely independently random to me.

 

[Prana Ferox]

I would really like a link to a video showing the manufacturing process by which you select longer copper grains. And what do you do with the short ones? Just throw them away or do they get, like, donated to charity?

 

[theREALdotnet]

You assume it is done identically under controlled lab conditions.

I wonder where you got that from? In fact, quite the opposite was stated:

This plot can be the result of many things, including being drawn by hand, who knows.

 

[Ingenieur]

The peakiness shows up in the same places in both waveforms--doesn't look completely independently random to me.

Each signal is random.
Generated from the same source (so they say).
Noise is random. The only known is the 60 Hz and harmonics likely 50 Hz UK)

The signal goes thru too many conversions.
They play music, I assume digital
Player
DAC
preamp
Amp
Sampled / digitized
Software manipulated to separate the music from the signal leaving the noise.
Not at the same time.

We have seen tests $$$ cords vs 50' extension cords. No difference.

The amp power cord can't reduce noise, only absorb less. RFI is uV or nV.
It will go to the lower potential, not the 'hot'. It is in the MHz range. It is not getting thru the xfmr or caps. The noise is from amplification not the AC power, it is not a signal.

Their R tests are absurd. I plotted them on a linear f scale. It makes no sense, not linear with L or skin effect. In addition the 'test' starts at 1000 Hz! Power is 50/60.
They don't show that because at that point no difference. At 1000 Hz ~ the same.
It's called baffling with BS.

Wiki
In communication systems, noise is an error or undesired random disturbance of a useful information signal. The noise is a summation of unwanted or disturbing energy from natural and sometimes man-made sources.

 

[Ingenieur]

I wonder where you got that from? In fact, quite the opposite was stated:

He said identical distributions.
Same sample or method.

But we don't know that. No FFT

 

[DonR]

I would really like a link to a video showing the manufacturing process by which you select longer copper grains. And what do you do with the short ones? Just throw them away or do they get, like, donated to charity?

They are dupes because the short ones go into First Run Copper cables and are regarded as the Premier Cru Superieur of copper grains and thus they are able to charge VASTLY higher amounts to only the most gullible discerning audiophile.

 

[amirm]

Given this setup, and if we assume that the underlying distribution of the data is the same, which is the same as saying there were measuring the same device in the same conditions, the outcome depicted in the plot, with one of the curves always above the other, is extremely improbable, it can't happen in practice.

It happens all the time when I measure equipment and I vary for example, grounding. Sometimes this happens with which USB port I use to measure a DAC for example. Noise floor simply jumps up or down. In these situations, I don't use the setup for comparisons like cables. You have to first find a repeatable, stable platform and then vary one element, i.e., the cable.

Regardless, can we reach out to the company and try to get more detail? I could try to repeat their test. What is the cable they used in that test and how much does it cost?

 

[theREALdotnet]

He said identical distributions.
Same sample or method.

But we don't know that. No FFT

Yes, the argument was that the two graphs can not be the result of noise affecting the measurement of identical things. It would be impossible (or at least highly improbable) for randomness to conspire to plot these two curves based on measurments of identical test subjects. The difference cannot be explained away as noise. Since we don’t know anything about the test setup we simply have no idea what caused the difference.

 

[JSmith]

That's why once you buy that power cord you should take good care of it and then pass it on from generation to generation.

Yes like the "watch" scene in Pulp Fiction, except it's a cable... ouch.


JSmith