Shielding method for RCA cables

Hi @tedsmith

Can you explain the mechanism that this shielding method is supposed to work, described by John Swenson ?

DIY DC power cables - Page 9 - UpTone Audio (Sponsored) - Audiophile Style

His solution sounds so simple - the shield doesn’t need to be connected to any signal ground at all. You just need a silicone wire to connect the 2 ends of the shield together.

What is this silicone wire actually doing?

And his explanation about cartridge is interesting !

Galen (@rower30) is much more knowledgeable about shielding, cables, etc. than I.


Galen’s more knowledgeable about that stuff than probably most people on the planet. :slight_smile:


Here is the way a shield is SUPPOSED to work.
Each shield end needs to be GROUNDED.
The shield needs to be low resistance to the FREQUENCY it is shielding.
When a signal hits the OUTER surface (ingress) or the INNER surface (EGRESS) it tries to get to the other side of the shield. The design of the shield ATTENUATES that signal based on the designs capability.

Since both ends of the shield are GROUNDED, and if the ground is good the shield ends looks like the same “point”. We don’t generate a current in the shield in theory.

The problem is, we do see resistance. As the shields Transfer Impedance increases with frequency, transfer impedance is a description of a shields resitance to frequency, we generate a current in the shield. Better shields generate less current. The induced current trying to get through the shield inductively couples “interference” to the signal wire inside. The graphs will show a shield’s performance will get worse with frequency as the transfer impedance goes up.

The magnitude of that inner signal that isn’t completely attenuated is measured in a RATIO called dB. The higher the better and means a smaller and smaller signal level on the opposite side of the shield.

When we DO NOT ground the shield, we creat a resonance in the shield. The energy isn’t REMOVED and basically bounces around in the shield or travels along the shield, inductively coupling energy as it goes. Not good.


"SPG resonates so that the induced signal is amplified, not attenuated, at the shield’s resonant frequency. This model is a low frequency model however all of the relevant parameters are included."

With BOTH ends of the shield “open” or shorted to each other, we also have a resonance system, not an attenuator to RF. Removing a shield to break a ground loop is not right, the “broken” ground potential was already there and needs to be fixed. A ground should see the same potential at any ground point in a system such that a PROPERLY gounded shield is essentially the same potential at each end. I = E/R and if R is zero we have no current to induced into the inner pairs after the shield attenuates the signal.

Most of our audio devices are not needing a power cable to shield RF. First the RF is many, many decades smaller than the 50 or 60 Hz “signal”. Second, the power supply inductors look like an open circuit to RF and block it. Second, after the power supply we have filter caps that look like shorts to ground for RF in each circuit blocks DC supply traces.

Will adding RF “filtering” to a power cord help? Let’s say it won’t hurt and leave it at that. We use an EDPM dielectric for BAV power cords and it passively filter RF. We get it at a reasonable cost in a durable 600 V cord. It is NOT the major benefit to the cords, though. We don’t have shielded power cords as the dB relationship to the noise and signal don’t merit a shield and the device SHOULD consider RF removal right at the circuit blocks AFTER the power supply and well away from anything the power cord is trying to add. RF gets in AFTER the power supply and good RF filtering on the DC rails is a necessary design step. TED knows more about that than I do…your turn Ted.



Very interesting, Galen. And nicely explained.

1 Like

Hi Galen

If both ends of shield are tied to signal ground, does adding this silicone wire between ends help in any way? Does it do anything at all?

1 Like

I’ve re-purposed old Monster & AudioQuest RCA Cables into XLR cables. These cables come with two insulated conductors and a bare copper wire. I think originally they left the bare copper wire either connected or unconnected (I’ve never really checked before hacking the ends off) on only one end of the cable. They also had arrows presumably indicating the way the signal flows (?).

I basically wire each insulated conductor to + (pin 2) and - (pin 3) and the bare Copper wire to GND (pin 1). I’ve never had any noise issues. Is is the correct way to do it ? I’ll let you Guys tell me. I’ve saved a fortune.


1 Like

You’re good on the PIN assignments, that’s the standard. It should work as well as it is going to based on the CUB, capacitance unbalance, between the two signal leads. That CMRR depends on the physical arrangement of the wires, and how much they are the “same” to the ground.


1 Like

Well, the shield isn’t really at earth ground potential with both shield ends tied with it, and I have to assume that wire is a “conductor”. We have to ask ourselves…when is the energy dissipated? At what point in the Kirkhoff’s current law loop does the voltage (interference) go to ZERO?

If your system is working like it should, earthing the shield at both ends is always the best. This put the noise where it belongs, to earth, once and for all. At that point the signal potential (amount it is ABOVE zero) is now zero.

The shield attenuates the NOISE trying to get through, and the grounds REMOVE the signal at earth. We do have some current in the shield an this is why we try to mitigate that with a super low resistance, or Transfer Impedance, to the frequency that the noise is at. Lower is better.


1 Like

Nice reply. Thanks. I believe John Swenson was talking about DC power cablesand not AC. I believe this changes nothing in proper shielding as you described (tied to ground on both ends), but what is your take?

Also, John recommends connecting the outer rim (negative) of the DC supply to earth ground, using a specific cable. I always found this a bit dangerous. But does it make sense?

Don’t forget, a shield MOVES the ground plane, too. All the wires UNDER the shield are electrically referencing the shield based on the geometry and distance to the shield.

If we don’t have a “perfect” geometry (we never do) the electrical will not be as good as an unshielded cable. The internal NEXT (cross talk) gets about 6 dB worse, for instance and the Impedance variation that causes RL, Return Loss refelections, goes up too. All this is based on the electrical geometry.

So when do we need a shield? If the interference is LARGER than the deterioration with the shield, we get a benefit. That’s that.

Any shield that is not earthed at each end, is a resonance system. That green wire for your PHONO does indeed EARTH the system. This wire puts the turntable and the amplifier and the system at the same ground potential. Same potential means ZERO current so no noise.

The IEEE has a spec for earthing, and how to test your earth for proper performance. EIA/TIA also uses this specification to evaluate earthing.

So far I’ve NEVER seen data to support improper eathing od shields be superior to an earthed shield.

Another way we can mitigate noise is inductive field cancellation. This works, too, but is hard to implament. It ALSO brings the potential in the circuit DOWN to as close to zero as it can.

This is used in the speaker cable and IC cable to reduce inductance. The field in the cable can be what you want, or created by noise. The wire in the cable doesn’t know. This is why analog is hard, once the noise gets into the signal it is hard if not impossible to remove it in analog systems. If we knew the equal and opposite the nosie we COULD remove it but we don’t. The noise is “alien” to what we know.

Even digital systems have issues with alien noise. 10G Ethernet has cable to cable noise that is “alien” to the signal inside. It can’t be cancelled. The cable needs to control this coupled noise with a SHIELD or DISTANCE or both, between one cable’s pairs and anothers to mitigate alien NEXT. It is a random noise unlike internal signal that we KNOW what they are and create filters to cancel that noise. RL and NEXT internal to the cable can be electrically cancelled to a degree since we can define an equal and opposite circuit topology. Alien NEXT can’t, we never know what it is going to be to cancel it. We didn’t make the alien noise signal.


1 Like

Thanks Galen. So as I can best understanding from your replies, you don’t agree with anything in that post I linked, in terms of best shielding practise?

Shield should be grounded and the external silicone wire does nothing for shielding

That post I linked is talking about general shielding including RCA cables but you’re right not AC power cables.

In later posts he talks about using his favourite Belden 1804A cables for RCA cables using this shielding method.

I believe Belden 10GX ethernet cable (Cat 6A) has shielding that is not connected to the end plugs? Plastic end plugs?

Correct, 10G, the “shielded UTP” uses an isolated from ground shield called and iso-shield that shield is designed to RESONATE outside the coupling band…so it DOES resonate, but the design of the shield (we adjust the Transfer Impedance Resistance) to moves the resonance band where we know it is ineffective to the performance spec. Thye alien NEXT has a frequency band to test across and we move the resonance outside that band.

The basic principals still apply, unterminated shields RESONATE. We just used some knowledge about that to allow UTP connectivity to get around the problems with shields and ultra high performance Ethernet cables. The shield allows a smaller size cable as it mitigates the coupling between them. Special filler design needs to be used, and an aggressive pair lay to offset the internal NEXT issues all shields influence.

10GXS (S=small) is smaller than UTP 10G saving a ton of space. So thatr’s why it exists. Space is $$$ in big installs.

Good question!


1 Like

I have no data to agree or disagree. The data is the data, and none is none. Until a proper tested comparison can be made (both done “right”) it is hard to say what the comparison is.

WHY you would want to use an ungrounded shield is ONLY referenced to break a ground loop (or a single point ground)…and that is a broken system ground. Don’t blame the cabler for that. Inventing a “solution” to avoid fixing the real problem is in itself a problem.


1 Like