Question for Galen about power cables

I’d like to ask about the why’s of power cables and their effect on the sound of our equipment. The case for speaker cable geometry has been clearly laid out. Outside of shielding, are there any parameters that should be optimized in a power cord? Propagation speed of different frequencies isn’t an issue with 60Hz. I read of different conductor bundles for low vs. high frequencies, but it’s 60Hz only being supplied at the wall. Seems to me that sufficient current capacity should be enough. Are there other issues, such as L & C that should be controlled in a power cord?
Thanks in advance if you are kind enough to respond. I understand if you don’t want to open this can of worms.

1 Like

I’d like to think in addition to low DCR providing for unimpeded power delivery better power cords would provide for significant noise filtering.

1 Like

Galen must be napping or started celebrating early. He will provide a detailed response soon.

I sure like that Trawler!!

Happy New Year and the very best of everything in 2022 to all!

Bob

3 Likes

Thanks! Albin 40. Good old girl, we spend a lot of the summer on her.

1 Like

OK, OK I’m here!

I’ve commented on power cables here and there, but the BIG attribute is to hold a low impedance between all your equiment and the earth ground. This holds the voltage constant going into our stuff.

WY down the line, the equipment takes the AC coming in, and converts it to DC to feed all the gain stages. The DC bias determines the GAIN and thus how linear the analog signal is going into and out of our amplifiers or driver circuits. If the AC voltage changes up or down, it is harder for the power supply to keep the DC “exactly” the same voltage and thus the linearity of the amplification.

OK, what does that MEAN buying a power cord? The voltage drop across the cord is the current in the cord squared times the cords DC reesistance. A preamplifier that draws near no current has no need for a 10 AWG power cord where an current hungry amplifier does.

A power cord really has no business being a noise mitigation strategy as much as you are told it is. No, that is the job of the equipment’s power supply to suppress RF INGRESS and EGRESS, both. The use of RF bypass capacitors and other RF impedance traps stops RF moving in or out far better than a power cord can, or should.

Sure, we can use dielectric and materials that are high dielectric resistances to RF (absorb the energy verses send it down the cable) RF. But is more even necessary?

Most equipment has “open” IEC receptacle and has passed UL RF properties such that the device is clean enough to be used with even a worst case cord. There are TWO levels, class I and II where the consumer grade is stricter than the industrial standard (see the chart).

What if a device fails the UL RF emissions test? That will require a CAPTIVE cord that is SHIELDED and / or has RF supression chokes installed. This makes sure the device is always “quiet”.

Can you use better and better shielded cords than UL requires to pass the emissions ingress / egress test (shields works the same both ways)? Sure you can.

Where is the beef in the RF noise? changing your amplifier and driver circuits gain properties? Is there noise or non linear voltage gain under test? If not, the signal isn’t going to change it sound. The EM signal HAS TO CHANGE somehow to be “different”.

Where we hear a difference is if we have a poor earth ground differential between components so fixing that is job one as they say. That differential if it approaches a ohm or so will be heard as HUM. This means we need a better AWG to the box and/ or a heavier power cord AWG. Technically the ground point everywhere should be the exact same “potential” (none). We add wire between everything so all systems have hum at some dB level. It is impossible to have zero hum, as wire has a resistance so we have a ground loop, even a tiny one.

I would worry a lot more about using 10 AWG ROMEX to the box than wasting money on a power cord that costs $$$ to solve RF noise as the AWG of the cord is all that solves the ground loop noise and that is far worse than RF. Yep, it isn;t exotic noise like magical RF. it is way worse, it is real and tangible noise we all have. The equipment can’t fix that problemn, it can and DOES fix RF noise. So upgrade the ROMEX way before you spend money on $$$ power cords.

Once the circuit ROMEX if fixed, THEN consider what current each device draws and buy accordingly. A phono head amp doesn’t need a 10 AWG cord. YES, a 10 AWG cord can be “different” and has to be to some place to the right of the decimal point but should you spend way out there? Just be aware of what is happening and to what degree. If Resistance is 0.00001 verses 0.000001 between power cords and you are drawing 0.001 amps what cord would you chose? A power supply can tolerate the voltage fluctuations more than the ground loop induced noise superimposed on top of it. BOTH are acting at the same time. At idle we have the ground loop noise, under load we add the voltage drop fluctuations. Lower DC grounds differential improves both problems.

System respond to removing the “system” noise as I call it, or the inherint ground loop noise. Power supplies resist voltage fluctuations (those capacitor banks) and filter RF really well. So we are left with the obvious choice to kill ground potential noise above all else. A system component can’t fix ground noise to the electrical box. Even a P20 is just a reservoir in a power chain and the better the input, the better the volt*amp (VA) output over a longer period of time.

I have a quiet system 35’ from the box in the basement. If I had noise, I would NOT go buy a $$$ power cord with magic RF suppression that isn’t the problem for most of us. I’d take that money and pay an electrician to replace the 14 (usually lights only) or 12 AWG (min for wall outlets) ROMEX with 10 AWG (optional). Some even add another fuse box closer to the system location. FIX THE GROUND.

If you do want a shield, DO NOT buy foil shields on high flex cables like power cords. They make them, don’t buy them. True, a foils shield is a low transfer impedance to ground for RF but…they don’t stay that way for long under flex. Aluminum cracks and few use high-draw / high flex alloy tapes. Use a double copper braid, 36 AWG is fine as we have little current. A double 95% braid is equal to a foil tape for RF. Sure, double braids cost a lot, but they will not degrade over time. A double braid will attenuate about 100dB of RF noise. That is a reference, though. If you have no RF it will be 100 dB down of nothing. If you have a zillion volts it will be down 100 dB and that’s probably not enough! Few will ever see RF such that a shield is even needed on a stereo but if you use one, use one that remains a shield.

We can get PHONO section noise that is usually diode rectified type AM radio noise. Yep, I could listen to stations states away with one phono stage I used years ago in my old PS Audio IV pre-amplifier! That’s a different RF noise circuit. The PS Audio experts can comment on that noise as they design around it all the time.

Best next year to all,
Galen

10 Likes

Thank you!

Thanks Galen. That was excellent as usual.
Can power cords, assuming well engineered and matched to the application, sound different?

Galen,

What are your thoughts on DC power cables using multiple RF Shields and microphone quad wires? It seemed to make a difference I could hear.

Also adding shields raising capacitance in AC cables with the dielectric layer. When is the capacitance an issue or is it fairly negligible?

You are spot on with Aluminum shields and AC cables as well a 10 gauge ground. I did recently wrap my amplifier cables first in Aluminum then added more ductile copper foil layer with conductive adhesive draining to the ground in the P20. It did wonders in dropping the noise floor adding copper layer and preventing shield breaks that were happening in the Al wrap. The aluminum foil was a good enough trial to help me realize RF was an issue in my system. I am about to do the same Al followed by copper wrap to my DS dac AC power cord, Though it no longer powers the analog board in my DAC since I went to a separate LPS with a direct shielded DC power cord. Hopefully it makes a difference it is only about a 30 minute DYI job.

The technology says we “hear” the nonlinearity of the DC transistor bias as “distortion”. There is no “AC” present at the transistor DC bias stage. The analog SIGNAL is controled by the gain properties and those are sdjuysted with the applied DC bias. Yes, we have TWO kinds of voltages on a transistor or tube, the AC (signal) and the bias (DC).

If the DC bias is dead stable, the “sound” of the amplified analog can’t be different. Well, it can if we have a terribly poor power supply and that is it’s main job, to rectify AC to stable DC. Since RF filtered out, what is left if any is rectified that to DC as well. All those capacitors take any AC voltage and slowly reduces the AC ripple to zip.

It has to measure different to sound different. If we measure RF at the transistor bias, we can change the sound. True. We no longer have pure DC. If we don’t see RF or any sign of AC, then how can we even postulate what level of RF interference we hear? We can’t say we hear what isn’t even measured or we can but how honest is that?

OK, that goes back to metals in cable. But we CAN measure and define the alloys or a metals grains and resistivity. We know there are differences in metals. But to look at a DC bias voltage we are darn good at spotting AC ripple in the DC that COULD modulate the DC.

When a BIG amplifier drops the line voltage from 120V to 110V under high power draw (look at amp reviews under max load!) we can possibly start to see changes in the supplied DC bias. This is where a P20 can help. It fills in the deficits with a higher VA delivery for a brief moment to mitigate that problem. A P20 isn’t magic, and a LONG sustained draw will depletes the capacitor banks, eventually. Music isn’t a sustained test signal. And yes, that AC line drop will be seen on the DC bias stages and thus be non-linear gain to some extent . How much I don’t know. The PS Audio experts can chime in on how unstable a gains stage is with Vcc DC bias sags or ripple from severe AC line drops.

There is a lots of stuff that will make every component different under severe load. How much of the transistors load line where it is “FLAT” does a design use? Stretch it out to the very limit and it is near non linear in the best of circumstances. One advantage to bigger amps is the power supply is near bullet proof to dynamics at lower power draws. A circuit has to be balanced between the power supply and the gain stages. If the transitor’s load line is maxed out, better have a stout power supply. Use a smaller section of the load line and the circuit MAY be more robust to nonlinear gain from DC bias voltage sag.

The next situation is if the RF is in the ANALOG side. A power supply won’t know about that side of the circuit. A high BW stage will see anything you send it as the signal. Some older high BW amps would “clip” due to amplified RF that is inefficient to amplifiey. Designers got smarter and kept the highlights wide BTW properties provided in the analog frequencies by making sure RF was filtered out. How does the RF get in the amp? Some sources are our cables!

Balanced cables and circuits eliminate common mode RF, and is one reason designers like balanced circuits. Most XLR even use an outer SHIELD (about 50 dB isolation) that attenuator the RF BEFORE the CMRR (another 65 dB isolation) stage and further improve the “residual” CUB (Capacitance UnBalanced) noise.

A balance circuit isn’t ever perfect and the smaller the common mode signal is entering the CMRR stage the more perfect it can look. The remainder voltage is thus way smaller. The RF is removed in two stages with quality XLR balanced cables. A cable is far harder to make perfectly balanced (we get to about 1%-5% CUB or so) than a op-amp circuit, so we like to add another layer of defense. We use a single ended shield (XLR PIN 1 is a unbalanced outer shield GROUND) and a balanced circuit (PIN 2 signal + and PIN 3 signal -), both.

RCA cables need at least 100 dB shield attenuation to remove RF. This means a durable low DCR high optical coverage shield. We have TWO resistances to worry about; the RF transfer impedance, and the DC resistance to ground. Each are independent of the other. A thin foil can shield RF with a low 100 dB transfer impedance factor but it can’t provide a low ground loop mitigation for ground differentials. We need a shield that does BOTH at the same time and that’s a double low DCR braid. This is a 100 dB RF shield and a low DC ground.

So there you have it looking bat the problems from two sides of the circuit, the signal input and the bias voltage source (power supply).

Best,
Galen

3 Likes

Capacitance isn’t an issue on a power cable since we have steady state 60 Hz, and the cord acts like a super small capacitor bank, storing energy to be delivered under instantaneous load. Don’t get carried away, the power supply does this millions of time better.

The dielectric choice is more for HIGH RF transfer impedance properties to the load, and route it to ground with a low transfer impedance. Or turn RF to energy (attenuation). We don’t really worry as much about capacitance. Inductance isn’t determined by the dielectric constant but physical proximity of the EM fields. We have those physics facts to work with. The dielectric is a low frequency volts per mil determined material.

Since we are at essentially DC with power cords, we can also use alternative wire materials with low DCR. If the exotic materials lower DC is worth the price over just getting a larger copper cord all else the same…may as well entertain yourself!

What a star quad can do is lower INDUCTANCE through magnetic field cancellation. This makes a power cord a lower impedance path for CURRENT draw. We don’t really send “voltage” so much as a flow of electrons at a set potential. Thus, we consider inductance. The EM fields cancellation are not conductor determined but field line orientation.

Best,
Galen

4 Likes

Wow. Thanks Galen. I’ll read this 10 or 50 more times until something fires a signal in my poor addled brain.
But, so far, I’ll take your answer to my question to be “no”.

3 Likes

Wow is right! Good to know my 12 AWG BAV cords are more than adequate for my pre-amp and source components (have 10AWG on my P20 and BHK250). Thanks Galen and Happy New Year to you and everyone else too!

1 Like

Thanks again for further analysis and comment. Jives with what I learned in college physics and hobbyist electronics years ago. I just couldn’t see where a power chord in front of a power supply could have much effect as long as sufficient current is present. It’s the power supply that has to provide the stable DC, and the chord is not an active element of the power supply circuit to any significant extent.
All the best to you for the New Year. Your contributions here are highly appreciated.

1 Like

For the sake of argument, communication and civility, let’s assume everything you wrote in the post I am replying to is absolutely correct.

Where, pray tell, are the citations to support your inerrant, absolutely correct statements?

You may not give a whit, but your credibility will benefit greatly if you were to hold yourself to the same, high rhetorical and scientific standards you use to bash the assertions and claims of others.

Just a thought…

Carry on.

3 Likes

Let me make sure, are you saying Galen is a) a liar, and b) not a scientist?

2 Likes

Just go tell a bumble bee they cant’ fly in the same tone
you use here…

The bumble bee will sting you with multiple stings which you
will certainly try to explain away with your highly evolved
intellect…

If you don’t like our home then do like a tree and leave…
Don’t pollute…

Best wishes…

1 Like

4 Likes

The DC bias is the life blood of an analog stage. Yes, there are different ways to manage the DC bias ripple. The argument is that it is “best” if it is ZERO, not the methods used to try to get there. The amps gain across frequency is DC BIAS controlled. Regulation improves the ripple, it doesn’t remove it. Feedback works but is always out of step to the correction. It has to see it before it reacts to it. But, does a power cord do anything to make it better or worse? I say no, it doesn’t as that’s the power supplies job and so far I’ve seen little data to show the DC supply sags under use. If we do have DC sag, we can maybe now say “better” is to get rid of it to some essentially “zero” value. Ripple can never be zero, in theory.

Until the DC bias, and any method, is improved with a power cord it is hard for me to justify the “sound” is changed and as you state, the ground imbalances caused by the wire we add between devices (ground pumping under load) may be worse and I say so, too. So look to keep differential DCR unbalance low based on the current. Current draw is causes the ground pumping and is proportional to the DCR. If you are worried about RF, well, use a passive dielectric that is poor for RF. The swept DCR “impedance” tests illustrate the dielectric effect on various cords and how an industrial design power cords works as well as $$$ power cords. Ideally the AC response needs to be a “short” at 60 Hz and be infinity above that. All we need is the 50-60 Hz power.

There is no such thing as “reflected resistance” unless you use the wrong term for Return Loss (fixed termination value) or SRL (variable termination to match the structural cable load) at RF. Analog does not really have RL, but simple reflections. My guess is you want to reference simple low frequency reflections, not a high frequency transmission line reflection.

If you are worried about RF in a coaxial lead, two braids is 100 dB. More is better for the worry wart. But the power supply itself really matters over the power cord if you are talking about power cords RF properties. Unbalance RCA, where the loop DCR is most important, the RF is attenuated by the braid as well and to 100 dB with two braids that also address the loop DCR requirement. You address both at the same time.

No, this RF shield effectiveness value is not out of thin air. For good RF attenuation, use a proper shield design.

For SEED test, Shield Effectiveness Evaluation Device, for RF go HERE;

https://www.researchgate.net/publication/3057334_Cable_Shielding_Effectiveness_Testing

Belden’s tested data corresponds to the independent IEEE SEED tests.

All referenced @ 1 MHz

A braid is ~80 dB

Dual Braid ~ 100 dB

Foil + braid ~100 dB

Foil + braid + foil + braid ~ 110 dB

Adding a BRAID lowers the shields effectiveness frequency range and is good for wide band problems like ESD.

A foil by itself is about 100 dB but at 10 MHz or higher frequencies. Once a foil cracks, it’s RF shield properties erode significantly.

MIL-C-17 recommends double braids for the overall systems worst case base response that reason.

Attacks with no technical justification to support your accusations is unwarranted and really, really, poor technical responsiveness. Saying someone doesn’t know what they are talking about with no proof seems to say you don’t know either. Better, bring support so the argument is based on proper measurement and the discussion is based on APPLICATION and if it is an improvement to the SYSTEM being discussed. I clearly state the ratio of what to expect based on the systems we use for audio. All the numbers are there. Once all the numbers are correct, we can argue what we hear and no one hears the same. I understand your desire to place a line in the sand for improvements being “meaningless” but everyone will draw a different line. To say the data is wrong is incorrect. The effect on the system is what we need to talk about once the product’s tested data is verified. I just gave you my data and will stand by it until it is proven to be inaccurate, It is not.

Best,

Galen Gareis

(Attachment BAV Power Cable Performance.pdf is missing)

5 Likes

Hear, hear!

5 Likes

So, I take it no data, citations to scientific papers, links or other references supporting your technical criticism of @rower30’s statements in the subject post are forthcoming…?

3 Likes