back to Learn

published: Mar-9-2013, updated: Sep-6-2019

post separation

This could be a very short chapter but as with everything it isn’t all black and white.

IF you feel / believe / are convinced that cables DO make a noticeable audible difference and feel comfortable about spending (loads of?) money on it, feel free to do so.
When it enhances the listening experience and the audible improvement is worth your investment then mission accomplished.

You WILL hear differences between cables but you can rest assured this is not because of electrical properties of the used cables.
The fact that you can (clearly) hear differences doesn’t mean there really ARE differences. Be sure to read the anecdotal further down this page.

I have never seen any evidence, aside from anecdotal reports, that interlink and power cables do make a difference. Discussing esoteric properties of cables is therefore pointless unless you can come up with some convincing technical evidence of audible changes.
And yes… capacitance, inductance, resistance, isolation or ‘impedance’ differ measurably between various cables so it would seem only logical that will result (and explain) subjective found differences. It would be really nice if someone out there is capable of showing, unmistakably, the links between the electrical properties and subjective found differences. Note that speaker cables are purposely left out here but when lengths are short and the cables aren’t really thin the above mostly applies for those cables too.
Only when speaker impedances reach really low values (incorrectly designed crossover filters or some electrostatic speakers) the resistance will play a roll and can certainly affect the frequency response.

Measurements in the audible range, and below as well as above it, show there are no discernible differences for interlinks and mains cables. Cables physically cannot ‘limit’ the frequency range between DC (zero Hz) and 1MHz so all the perceived ‘limitations’ of a cable should be viewed from the perception side of things (hearing) IMHO.
They simply do not function as bandwidth limiting filters within and FAR outside the audible range, not even when one considers 100kHz as audible.
So basically a cable cannot ‘roll-off’ the bass or the treble nor make peaks less prominent or ‘smooth’ out sound.
Cables also cannot cause distortion or introduce noise that can reach audible levels.
That last part (distortion and noise) is true for ‘normal functioning’ cables. Dirty/corroded connectors, poor solder-joints or almost intermittent (broken) cables CAN be the cause of distortion and noise.

Cables come in all sorts, colors, thickness, sheeting and all sorts of connectors be them plugs or sockets.

Headphone cables with a single return wire (so 3 wires in the cable are splitted in the headphone itself) can have some measurable effect in stereo imaging. The question remains how audible these small measured effects are. This is covered in impedance, resistance and other issues article in the tutorial section.

There is a very thorough article here as well which is worth reading

post separation

What to look for in mains cables:

  • long enough
  • correct fitting plugs
  • Must be able to handle the drawn power

That last one is always the case, even for the flimsiest ones, that is in case of sources, pre-amplifiers, headphone amplifiers. For (high) power amplifiers a thicker conductor may be desirable, especially for countries with 110V mains voltage.

For the DIY’ers I would say making your own mains cables is only needed if you require other lengths than the normally available ones or when special plugs are required.

The mains cables supplied with equipment and the low cost cables available from electronics-shops are good enough.

Rather spend your money on things that DO matter, mains cables do NOT, no matter how badly cable vendors would like you to believe they do or your ears tell you.

Screened mains cables aren’t really needed. The screen would have to be connected to the safety ground and will not do much against common mode currents as the cables in the wall will NOT be screened and thus the ‘ground’ connection on your mains socket does not do anything at all against this.
Should you have very sensitive audio-signal carrying wires (unscreened interlinks… yes some people actually make/use/buy them) that easily picks up hum than it is not wise to run mains cables along those lines. A screened mains cable could be of slight help here when these wires run very close together over a certain distance.

Keep sensitive interlinks away (at least 10cm) from mains cables and above all from transformers. Most (heavy in weight) audio devices have big transformers in them.
When audio- and mains-wires have to cross each other try to do this at a 90 degree angle.

Despite the common wisdom that ALL audio has to come from the mains power and thus mains cables make a difference you should realize that there is absolutely NO AUDIO going though ANY mains cable, only short ‘pulses’ of a few milliseconds wide every 8 milli-seconds (60Hz) or 10 ms  (50Hz).
For more info about the signals that pass through mains cables have a look at the ‘rectifier page‘.

Also consider the fact that the mains voltage is coming from miles and miles away passes through kilometers of (mostly) aluminium wiring, large transformers and plain solid (as in not stranded) copper wiring into your home. From there it passes through a fuse, a switch and fans out to outlets again through plain solid copper wires into your wall socket.
From there it continues for 1 meter or so via the ‘audiophile grade’ cable. Enters your amplifier where ‘normal’ copper wire and (thin) copper PCB traces take over again into a cheap copper wire transformer or switch-mode power supply.
That 1 meter of cable IN SERIES with all the other wires  … would that really make a difference ?
This means it passes through ALL of these wires and while one (or more pieces) of ‘wire’ can certainly ‘degrade’ the mains power supply a short piece of wire at the end cannot ‘undo’ anything nor ‘improve’ the quality again. No matter how bad one wants this to be true and can ‘clearly’ hear improvements.

post separation

What to look for in headphone cables:

  • low in microphonics (microphonics means conducting as little as possible mechanical conducted noises such as rubbing against clothes) the cable itself does not act as an electrical microphone.
  • supple
  • nice colour scheme
  • sturdy plugs

The often mentioned capacitance, inductance and even ‘impedance’ are all nowhere near any ‘borders’ where it can have any influence.

For DIY’ers the benefits of making  headphone cables yourself is.

  • selecting the colour
  • selecting thickness (rugged)
  • selecting microphony
  • selecting the correct suppleness (not so easy to find)
  • selecting needed length
  • selecting good quality (not uber-expensive/high-end) plugs/connectors
  • using 4 wires where the common is only connected in the plug or 2 wires with a screen that can act as a low Ohmic common return wire.

Gold or Nickel-plated plugs makes no difference in anything else than aesthetics under normal ‘household’ conditions.

post separation

What to look for in interlink cables:

  • correct length
  • plugs that fit firmly yet do not clamp so tight they are hard to remove
  • Provides proper screening

For DIY’ers the benefits are:

  • selecting the correct length
  • selecting the colour
  • selecting properly fitting plug
  • selecting properly screened cable

Choosing expensive or exotic cables (silver or extreme copper purities) is pointless and only wears down your wallet.

Gold- or Nickel-plated connectors:
Gold isn’t the best conductor out there but has the pleasant property it doesn’t corrode. It still can get dirty though. Especially when used in rooms that are often filled with cigarette smoke. Also when a connector is rather loose fitting it can still make poor contact . Another material that is often used is Nickel, which looks silver-ish, and also doesn’t corrode very quickly but is not free of corrosion. There are a lot of people that ‘connect’ sonic properties to these materials (gold would sound warm, nickel would sound bright), there is no scientific basis that provides any proof in that direction and rather seems something evoked by association with the materials.

Copper- or Silver-wiring:
Silver is a (marginally) better conductor than copper. BOTH metals, however, conduct all frequencies the same way, they do NOT act as ‘filters’ nor are they physically able to ‘boost certain parts of the frequency spectrum’ in and well outside the audible band. It is often said that silver sounds ‘brighter’ than copper and it has better ‘high-frequency conductance’ and therefore sounds ‘bright’. There is no technical basis for this theory (no, also skin effect is no excuse here). It seems based on the same associative principles as with the Gold- and Nickel-plated connectors.

Skin effect

For cables like interlinks the resistance increase of a cable by the skin effect is a non issue because of the high load impedances (Well above 1,000 Ω) and even for headphone wiring.

A 3 meter long AWG28 (0.3 mm) copper wire measures 0.71217 Ω at 100Hz and 0.71578 Ω at 100 kHz which is a 0.507 % increase in ‘resistance’ for inaudible frequencies.
The same 3m cable made from pure silver measures 0.67312 Ω at 100 Hz and 0.67704 Ω at 100 kHz which is a 0.58 % increase in ‘resistance’ for inaudible frequencies.
For arguments sake lets assume a copper cable with a very thin silver coating and let’s assume all lower frequencies go through the copper and 100kHz travels through the silver.
In that case the 3m hybrid cable measures 0.71217 Ω at 100 Hz and 0.67704 at 100 kHz which is a decrease of 5% in resistance which can be considered ‘worst case’ and thus 100% skin effect.
Assuming an unrealistic low input impedance of 1 kΩ (most impedances are well above 10 kΩ)  this will yield an ‘increase’ in frequency response of + 0.0003 dB at 100 kHz compared to 100 Hz with 100% skin effect (which isn’t even close to reality). Now let’s hope your amplifier isn’t ‘down’ 0.001dB at 100 kHz already which it certainly will.

So skin effect in interlinks ? Nah.

Now lets move on to headphone cables and assume a 32 Ω resistive headphone (orthodynamic for instance).
Same 3m hybrid cable (very long for a headphone and very thin wiring).
This cable will give an ‘increase’ in frequency response of + 0.01 dB at 100 kHz compared to 100 Hz.

Dynamic headphones increase in impedance significantly above 10kHz because of the voice-coils inductance making the differences even much less dramatic

Will this theoretical maximum increase of 0.01dB  (with 100% skin effect) at 100kHz be audible ?

Then there is also the cable inductance which is the same for silver and copper cable.
This increasing inductance will ‘undo’ the possible skin effects in a much larger way as with a 3 meter cable the inductance will be about 3 μH which, at 100 kHz will have an added ‘resistance’ of 1.88 Ω.
This will cause a roll-off of about -0.5 dB (assuming a resistive load) at 100kHz vs -0.51 dB for silver plated copper wire… at an inaudibly high, and not present in recordings, 100Kz signals and providing your amplifier has a bandwidth of well above 1MHz and your headphone driver can reproduce 100kHz AND there is 100% skin effect present.
In essence I wouldn’t worry about the skin effect at all and as soon as some starts mentioning it as a ‘possible’ explanation you now know what to do ….

Another thing to consider when you live near the coast, silver and salt watery air aren’t good matches.
Salty air corrodes silver much faster.

post separation
Some headphone cable measurements.

the ‘3 wire no screen’ cable is the HD681 stock cable, the thin litze cable is the ‘Skytronics’ cable and similar to those found on many of the cheaper headphones with the thin and supply wiring.
Litze wires consists of a bunch of very, very thin wires where often each individual cable is isolated with a very thin coat of paint. The paint is really easy for identifying multiple wires in one sheath. This material is not easy to solder as one has to remove the paint before it can be soldered.
The ‘rods’ are the actual headband rods of the HD681, the ones that carry the signal to the right earcup.
It was ‘feared’ that these rods would degrade sonic quality due to them being Chinese steel.

Resistance, Inductance and Capacitance (between channels and to common) per meter cable.
Below that is the loss of power in dB/m when a 32Ω Headphone with a 200µH inductance is used.

Some people are concerned that cable capacitance and it’s inductance would create a ‘resonance’ in the audible range. With a typical 600pF capacitance (3 meter cable) this would ‘resonate‘ (if it were not damped by the amplifier and voicecoil resistance) at around 500kHz. If the voicecoil resistance would be close to 0Ω some amplifiers might become unstable but the resistance of the voicecoil prevents that.

Some people have the fear that the  cable capacitance combined with an output resistance of an amplifier causes roll-off in the audible range.
Let’s assume a 3 meter cable with 300pF/m so 1nF of cable capacitance.
Let’s assume a 120Ω output Resistance  (most will be much lower but 120Ω will give worse case scenario)
This will give a -3dB point of 1.3MHz... I doubt anyone should be worried about that…
As most amplifiers will have an output resistance of well below 10Ω the -3dB point will be well above 10MHz (10,000 kHz = 10,000,000 Hz)

Another concern is that a cable capacitance connected to a low output R amplifier (and thus in parallel to the output) .
When an amplifier has overall feedback (the ones with a very low output R usually are) that capacitance could cause the amplifier to become instable and oscillate.This is not an imaginary scenario when the amp is poorly designed, it usually oscillates at a frequency that is far above the audible range.
A worst case scenario 1nF capacitance usually isn’t a problem for most amps, most certainly not when the design is made with this load in mind.
As soon as that capacitance has a resistance in parallel the ‘possible’ problems become less severe.
The lower the impedance of the headphone the smaller the risk of instability.
Most well designed low impedance amplifiers have Boucherot filters or other countermeasures in place though and are stable regardless of the load (within reason).
All amplifiers I ever tested had no problems with 1nF, most even didn’t react to 10nF. Some amplifiers got into trouble with 100nF loads though.

Freq range of all tested cables (3m) had their – 3dB point all above 1MHz (> 1,000 kHz).

What about cable inductance ?
A 3 meter cable will give a total inductance of around 6µH compared to the 100µH – 3mH of the voice-coil itself, which is in series with the wire, this added cable inductance is negligible.

How about resistance… Well this does do something when a single return wire is used and the cable has a high resistance (thin supple wiring) while the driver impedance is low (16Ω to 50Ω) this will affect stereo imaging.
A cheap (thin) 3 meter cable can easily reach 1.5Ω and the very thin wires 3Ω, the better cables remain below 0.5Ω.

Some more info on the effect a single return wire can have (as opposed to 2 return wires or balanced operation) is found in THIS TUTORIAL near the end of this article.

Also there will be some influence in frequency range ‘flatness’ with SOME low impedance headphones, those that have wildly varying impedances.

Let’s take the Sennheiser HD558 with a nominal impedance of 50Ω rising to 300Ω at 100Hz  is connected to a 0Ω amplifier.
A cable with 3Ω is used. The frequencies around the nominal impedance will all be -0.5dB in amplitude compared to what has been put into the cable. Around the resonance frequency (100Hz) the damping would only be -0.1dB so an HD558 (which has an unusual high impedance peak) will have increased bass response of about 0.4dB. With a ‘better’ cable say 0.3Ω this difference is merely 0.04 dB.

Damping factor

The damping factor is often mentioned when people are experiencing a ‘tighter’ bass.
This term was coined in the 80-ies when manufacturers where building amplifiers with incredibly high damping factors of 10,000 + values. This looks great in spec sheets but really does say nothing. You needed 4 wires going to your speaker for this.
What the damping factor expresses is the load resistance divided by the source resistance.
It’s a bit of a nonsense story though. There is no difference in damping between an amp with a DF of 10,000 and 50 simply because the output resistance of the amp and speaker are in series. For the drawn current in a circuit the load resistance is the highest one (by far) and thus determines the actual current.
The coined idea was that a DF of 10,000 was a better ‘short’ for the speaker as it is in parallel with the driver. That way the ‘short circuit current’, which is generated by the speaker when making undesirable movements, would be completely shorted and thus damped better.
The nonsense part here is that while for voltage the amp and driver are in parallel for currents (which does the actual damping) the amplifier and driver is in series.
So the current in a circuit does not change much if the total resistance is 8Ω + 0.0001 = 8.0001Ω or when it is 8Ω + 0.1Ω = 8.1Ω. The damping current of the latter is only 0.99x that of the ‘super damping factor’ amp. (0.1dB).
Of course, with the 4-wire ‘speshul’ amplifiers the speaker cable resistance is also removed thus the cable resistance with ‘normal amps’ would have to be added to the output resistance as well reducing the damping factor a bit more.
Most speakers (woofers), because of their mass, rely on mechanical damping (speaker + cabinet) as well as on electrical damping. Lets not forget the resistance of the crossover inductor here as well which may well f-up the damping factor for the woofer more so than the speaker wiring and amp.

The same goes for headphones except that headphone drivers are higher impedance (and thus the damping factor is automatically higher when a low output resistance amplifier is used) AND above all are designed to rely mostly on mechanical damping.
For headphones the damping factor is not a ‘thing’  when it concerns the actual ‘damping’ of the driver. The varying impedance and the influence that impedes on voltage division compared to the output resistance of the source (which more often than not is higher than 0.1Ω) is explained further here.

post separation

An anecdotal (and true) story about a loudspeaker cable swap I experienced in my previous job as service engineer for a (former) high-end audio store.

As a service engineer in a small company I also had to install equipment for customers in their homes.
One fine day a customer had bought an expensive cable upgrade for his way better than average music system. The customer was in the music business is all I will say and he shall remain unnamed.

The following is what the CUSTOMER experienced. It is my free interpretation of how he was thinking based on his reactions and remarks.

Even though I wanted to see what the technician was doing he asked me to sit in my carefully positioned listening chair and not move around so I could hear for myself what the changes are in sound quality. This could be quite easy as I know this system and the music inside out and any change will be obvious. If I can NOT tell the difference the salesman had stated I could return the cable and there would be no financial loss.

First I listened to the old cable for a while, it sounded ‘as usual’. After the technician had changed the cables I listened again and the changes were not even subtle but the sound improved clearly. After enjoying the music the technician wanted to put back the old cable to verify the change that occured.
Old cable was installed again… old sound back too so the cables appear to be real upgrade.
After several seconds I already pleated to immediately put the new cables back in there.. How could I have been satisfied with this sound all this time… I should have upgraded sooner !.

The technician put in the new cables again .. Ah there is the excellent sound again and will likely have to revisit all my music again.

What to do with the old cables he asked ?… garbage can of course.

The following is what ACTUALLY happened that day:

I told the customer I wanted to show the improvements the cables made to him and asked him to sit in his chair and put on some music he knew well.
The amplifier did have an A-B speaker output which I could use to switch between cables but as the expensive cable could be returned when the customer did NOT hear any improvements it would not be wise to connect both cables and switch between A and B for obvious reasons.

I told the customer I was going to change the cable while he remained in his chair.

First listen was with the old cable… nothing unexpected came from that.

Next I routed the new cable and moved the amplifier a bit as well as sitting behind the heavy speakers pretending to connect the cable but in fact ONLY routed the cable and left the old one connected.
I didn’t even touch the connectors.

The customer put on some music and the ‘change’ was quite noticeable to him and he listened to it for a while.
He was very pleased with the results thinking (to him knowing) he was actually listening to his new cable.

I proposed to check his findings by connecting the old cable and have a listen again to make sure he heard it correctly. This time, however, I connected the new cable while the CUSTOMER was thinking he was going to listen to his old, and clearly inferior, cable.

Indeed to the customer the ‘old familiar and inferior sound’ was back and already within seconds he asked me to switch back. The change was dramatic and wanted the new cable back on there.
In reality he was now listening to the new cable for the first time.

The customer wanted his new cable back in there. So I pretended to swap cables again while doing no such thing in reality.

I told him the new cable was connected again and took away the old cable.

He listened again and immediately noticed the improvements. He was not ever going to use those old cables again and was putting on different music to listen some more !

He paid for the cable and thanked me for the demonstration…

Everybody was happy that day.

The customer for the obvious sonic improvement, the demonstration he got as well as the free delivery and installation.

I am, for the generous tip he handed me for services provided, and above all the pleasure of experimenting on unsuspecting people with excellent hearing in the comfort of their own home and familiar with the system. Of course he still doesn’t know and will gladly tell everyone about the noteworthy improvements he clearly heard when swapping cables.

My boss is happy for making the sale so he can smoke another Cuban cigar.
I have to add that returning cables was standard policy when buying a cable and there was a big sign in the shop stating cables could be returned when no improvements were heard.
In the 7 years I worked there not a single customer EVER returned a cable. And there were some pretty expensive ones amongst them. So either everyone hears it or they don’t but do not want to be laughed at behind their back or figure a better cable can’t hurt.

The importer and manufacturer of the cable are also thankful because of the big profit margin.

post separation

back to Learn

  1. solrighal says:

    Nice one! I was sceptical of sceptics but now I’m not so sure. Job done!

  2. jonas says:

    hi! i’ve got the koss porta pro and my cable broke again. I already replaced it with a cheap thing from china. It seems it’s hard to find good quality replacement cables these headphones. do you have any suggestions where to get nice spare cables?

  3. An interesting read on cables that unfortunately ignores the effects of RFI. I have spent years messing about with power amplifier design and digital kit, cables etc Having done test work on this it is clear that RFI and EMI are the key factors in system sound, cables etc. My experience is that demodulation of RFI occurs in transistors not so much FET input stages and very clearly effects sound quality wherever it manages to get into a system which is pretty much everywhere. Where RFI and EMI are concerned every thing that is not screened and even when lightly screened is a potential receiver. On the power supply and diode article on this site again EMI and RFI are ignored. If you take an analogue scope with an RFI sniffer coil and you can start to see why different diodes and snubber arrangements effect the sound on a high quality system. The clearly visible effect of diode and snubbers on sound quality relates to the effect on transformer ringing and the consequent ringing noise spectrum which tends to vary in frequency amplitude and duration depending on loading something which varies significantly in a power amplifier, diode choice and snubber arrangement. Add in some DC on the mains leading to partial core saturation of the transformer and the RFI emissions increase with the transformer ultimately becoming a low level RFI transmitter sat in the middle of the amplifier. Factor in parasitic capacitive coupling between the transformer and the case and you have total high frequency chaos with ground impedance and interconnect impedance at high frequencies coming in to play along with parasitic capacitive coupling between output transistors and heat sinks etc, etc My experience indicates that it only takes very low levels of RFI and EMI to have a large effect on sound quality especially in high gain circuits such as power amplifiers. This really is not simple stuff, my experience indicates that a methodical step by step approach to controlling RFI and EMI can reap very substantial improvement in sound quality. Having said this it can be very difficult to achieve improvements in practice, most of the standard tricks such as standard filtering are generally fairly ineffective as frequencies increase. Over time and a lot of trial and error I have come to learn some simple tricks which are highly effective in practice. The majority of equipment designers are I am sure aware of RFI and EMI as something they have to control to comply with CE approval but perhaps not it’s profound effect on sound quality. Habit dictates that they relentlessly focus on the audio band only and distortion which at sensible levels has a minimal impact on sound quality when compared to RFI and EMI, a great shame but something that will hopefully change in time. Richard George

    • Solderdude says:

      I agree that certainly in this age with many devices in GHz band RFI is more of an issue than in yesteryear. In those days the radio amateurs were a problem. Many people complained and the amateur was then asked to immunize the equipment of others they affected.

      Yes, there is also a lot of RFI floating around in switching and digital devices and they should comply to rules ensuring no influence is there on other devices. That is when connected in the prescribed way which often is not the case. Certainly not when people start connecting line level devices with (homemade) not screened interlinks.
      Power amps usually have a low output impedance which increases for HF frequencies (inductor in the output stage for instance) so very little can enter the amp.

      In power supplies RF can only enter in common mode. The power rails itself have plenty of (local) decoupling going on that works well in the 100MHz range.

      I do EMC tests now and then (for work) and the levels injected there (conducted as well as radiated) are often of insane levels.
      When some effect is registered (heard) or measured it is always some tones, sweeps or weird sounds that become audible. Some digital devices experience drop-outs or ticks.

      However, I never experienced ‘sound degradation’ because of this. Most certainly not in or caused by power supplies.

      Always use screened cables, XLR connections usually have less problems.
      More expensive does NOT mean better screened. There are plenty expensive crappy cables.
      There are also plenty of low priced cables around that are excellent.

      When one has a groundloop this also not reveals itself as ‘sound degradation’ or loss of ‘magic’ in the music but in weird sounds.
      This can happen with certain types of gear (which can work flawlessly with other people) and usually can be solved by breaking the groundloop at a specific place. That may be hard to diagnose or counter when one does not understand common mode current flows.

      So I agree that HF ‘garbage’ can influence equipment but severely doubt (have not seen any evidence) of reduced fidelity. It usually reaches a certain level and then gets rectified (becomes audible) as some sort of unwanted noise(s).

      Perhaps you can post or link to some research showing performance changes when someone just changes a snubber or diode on the output signal of equipment ?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.