Are headphone balanced cables snakeoil?

[Sawdust123]

I think I need to invent balanced cable-lifters for the audiophile crowd. The typical ones are clearly single ended.

 

[abm0]

I know this is an old post but I agree 100%. I used to love head-fi (and still do for several reasons) but the near-total ban on ABX chat means that so much goes unchallenged now.

I suppose they might've banned it from product threads, but there's a science section where it seems perfectly fine to discuss audibility limits and stuff. I was able to challenge the notion of "Sabre glare" in a dedicated thread for example, and the preliminary consensus turned out to be that there's no such thing.

 

[trl]

From a technical perspective, balanced amplifier are much more "potent", delivering like 3 times more power when using the same output buffers, so this can achieve lower THD numbers and ability to power hard to drive cans more easily.

Also, a highly resistive cable (long and thin) will have a bigger impact on single-ended cans vs. balanced cans, although I doubt the differences will be audible.

Also, Jan Meier has a good representation of the magnetic field into headphones cables here: http://www.meier-audio.homepage.t-online.de/grounds.htm.

 

[abm0]

From a technical perspective, balanced amplifier are much more "potent", delivering like 3 times more power when using the same output buffers, so this can achieve lower THD numbers and ability to power hard to drive cans more easily.

From a price perspective though, they are not much more powerful, not even close to 2x, so telling the final consumer "balanced amplifiers are more powerful" is essentially a lie.

 

[trl]

From a price perspective though, they are not much more powerful, not even close to 2x, so telling the final consumer "balanced amplifiers are more powerful" is essentially a lie.

Not quite true! A balanced amplifier has 4 amplifiers inside instead of 2 and will provide up to 4 times more power because amplification in balanced operation is not related to GND (which is zero volts), instead it's a sum between the original and the inverted audio signal. My Matrix HPA-3B has over 1.5W in single-ended operation and over 4.5W in balanced mode. Schiit Ragnarok has 6W/ch. in single ended and 24W when balanced, so you need to stop thinking to balanced amps as something that sucks and shouldn't be on the market, because it's not true. As for their higher price, you can always find balanced amps to a decent price or just get to diyaudio forum, find a schematic and build yourself one.

Have a look at the main page and check DAC SINAD graph: price difference between 120dB and 119dB is 3000 USD - 500 USD =2500 USD! So yes, only 1dB may cost 2500 USD! Sometimes a tiny little increase in performance may cost you way more. Although, you can anytime build cheap 4 amplifiers with OPA1611 + LME49600 and get about 3W/channel when used in balanced mode and all of that with less than 100 bucks for the modules and probably a bit more for the PSUs.. While getting the same amount of power in single-ended will definitely cost you more because you will need to use more powerful transistors.

I feel an aversion on balanced amplifiers and I don't know why, because it has been already proved that balanced amplifiers are providing more power and a better SNR and they have a very good rejection of mains hum, basically because of the way balanced operation works.

 

[Julf]

I feel an aversion on balanced amplifiers and I don't know why, because it has been already proved that balanced amplifiers are providing more power and a better SNR and they have a very good rejection of mains hum, basically because of the way balanced operation works.

But that is mainly due to the differential input circuitry.

 

[abm0]

My Matrix HPA-3B has over 1.5W in single-ended operation and over 4.5W in balanced mode. Schiit Ragnarok has 6W/ch. in single ended and 24W when balanced

That is about as relevant as the marketing trick of showing a high price with a strikethrough line over it and a lower price next to it, which is how much you were going to ask in the first place anyway. You don't compare an amp's outputs with eachother, you compare them with the outputs of other (SE) amps available on the market for the same price. That's what I meant when I said "from a price perspective". Balanced amps being about massive power is just marketing. The only thing you will consistently get out of them is lower crosstalk, that's all. Anything else is exceptional/situational/synergy/luck, and promising it to everyone is misleading the consumers (a.k.a. marketing).

 

[eboleyn]

I'm glad I found this thread. I'd recently looked for an article on balanced headphone cables and generally found nothing actually useful until this thread.

So, I'll preface the rest by saying I'd long found speaker systems hugely superior to various supposedly high-quality headphones I'd been recommended to try by various people I knew. About 3 years ago I was got a set of higher quality over-ear headphones (Hifiman HE-500) with balanced cables and a decent headphone amp/DAC combo (5.5W max power output, it happened to be balanced amp output) for fun. It was definitely a bit of a jump over the best I'd heard before but not enough to change my mind about how much better speakers are than headphones. I also didn't really consider the balanced vs. unbalanced nature and if that is important for the sound quality, but thought it might be snake-oil and just the better quality components were giving me the better results.

Recently after having read yet more reviews about the "best headphones", I thought "what the heck it'll be fun" and took the plunge on a set of Hifiman HE-6se headphones, then built a speaker-amp to 4-pin XLR cable to factor out the differences compared to my fancy living room setup - so identical except for headphones vs. speakers. WOW it's now nearly spot-on (lovely tonality at all frequencies, super-engaging, clean imaging, the sense of the singers being "right there in front of you", etc.), with the only real difference left being the effect when you move your head around (imaging issues). Maybe the speakers are still a tad better but it's now a very small difference if not just all in my head so to speak. OK, I think I still prefer a speaker setup when I can but it means I can really enjoy music at work, for example.

In the process of this I'd been starting to use my HE-500's for listening quietly when the family is asleep, and, never really having though about the balanced vs. not headphone cable issue got a "standard" headphone extension 25 foot cable. Bleah I really noted the quality got worse (tonality still generally there but the imaging and sense of "right there" blurred) when I used the extension vs the balanced cable I'd had even with the 4-pin adapter to standard headphone 6.3" TRS. I kind of doubted it was due to environment noise, and after pondering "what could this be?" I was stumped because all the effects I could compute were vanishingly small and I didn't think human ears are nearly that good, until I finally discovered the shared return conductor effect that "solderdude" mentioned and on the cable I'd bought it would lead to around -35db crosstalk on my HE-500's!! Yikes!

So I built a cheap 4-pin XLR extension cable using 25 feet of CAT6 wire, using the 2 highest twisting pairs for signal and the 2 other pairs connected for ground. Using that and connecting again with my 4-pin XLR to 6.3" TRS adapter at the amp made the results essentially the same as no extension cable.

I do indeed still think that most of the "balanced headphone" effects are snake-oil, but the crosstalk thing is no joke and had a major effect for me when testing at audiophile quality levels. At this point I'm now deeply irritated that manufacturers use 3-wire for headphones as the default when they could use 4 wire and cheaply get better quality.

Anyway, thanks again for this thread for confirming my results!

 

[Sergei]

My Matrix HPA-3B has over 1.5W in single-ended operation and over 4.5W in balanced mode. Schiit Ragnarok has 6W/ch. in single ended and 24W when balanced

It's when I test my amplifier, usually 0.0001x% thd with 8ohm load 100mw SINGLE CHANNEL DRIVEN, in DUAL CHANNEL DRIVEN test, the distortion rise up to 0.002x% basically 26db+ degradation in performance.

Perhaps it pays to consider a wider signal delivery chain. The quoted posts seem to indicate that an amp's power supply unit may have something to do with the elevated levels of distortions in one channel when a signal is present on another channel. This effect may be also measuring as elevated crosstalk.

The first post says that Ragnarok (weighting 32 pounds, consuming up to 400 watts) delivers 4x more power with doubling of effective rails voltage, which is as expected. The HPA-3B (3.75 pounds, 25 watts) does not - it delivers 3x instead of expected 4x - which gives a reason to suspect that HPA-3B output is limited by its power supply.

The second post describes the behavior of Topping DX3 Pro. Out of the box, it is equipped with a little wall wart 15W switching power supply. Interestingly, one can buy a popular linear power supply compatible with DX3 Pro, which, while nominally delivering the same power, has much beefier capacitors. When only DAC's line outputs are considered, the type of the power supply doesn't appear to make a difference.

Yet when a DAC drives headphones dissipating significant power, my hypothesis is that other effects may kick in. The precise magnitude of undesirable effects rising when a power supply is driven close to its capacity may differ dramatically between circuit designs. Still, from general principles, we can expect two major effects: (A) increase in output voltage ripple and (B) output voltage occasional sagging. The beefier capacitors help combating both effects.

Single-ended headphone amplifiers have to effectively use a power supply unit shared among the channels (even if there are technically two power supplies, their mid-voltage outputs have to be galvanically connected). If the power circuits are not "overbuilt" relative to the power demands, the minuscule ripple and sagging may effectively manifest as perceived crosstalk.

If the power supply is well-designed - like pro-audio power supplies tend to be - than these effects are much less pronounced, and even amplifying a dozen headphone channels off one power supply unit is fine. Proper power supplies and no-tricks, conductive enough, symmetrical enough, pro cables perhaps explain why pro audio users and designers don't find it necessary to use balanced headphone cables.

Balanced headphone amplifiers may effectively separate the power supplies, at the very least those parts of them that feed the output stages. Either double-mono approach, or intermediate voltage regulators, may reduce the "power supply crosstalk", and then cheaper and more economical power supplies could be used without noticeably degrading the sound stage.

Once again, this is a hypothesis, which I regrettably don't have time to test myself. If valid, it would explain the controversy, and why some people feel so strongly about it: it is not just balanced cables per se, it is also decoupling of amplifier channels from an insufficiently beefy shared power supply unit that may help reducing the crosstalk.

 

[trl]

Perhaps it pays to consider a wider signal delivery chain. The quoted posts seem to indicate that an amp's power supply unit may have something to do with the elevated levels of distortions in one channel when a signal is present on another channel. This effect may be also measuring as elevated crosstalk.

The first post says that Ragnarok (weighting 32 pounds, consuming up to 400 watts) delivers 4x more power with doubling of effective rails voltage, which is as expected. The HPA-3B (3.75 pounds, 25 watts) does not - it delivers 3x instead of expected 4x - which gives a reason to suspect that HPA-3B output is limited by its power supply.

The second post describes the behavior of Topping DX3 Pro. Out of the box, it is equipped with a little wall wart 15W switching power supply. Interestingly, one can buy a popular linear power supply compatible with DX3 Pro, which, while nominally delivering the same power, has much beefier capacitors. When only DAC's line outputs are considered, the type of the power supply doesn't appear to make a difference.

[...]

In HPA-3B the PCB traces coming from the XLR plugs and going to the 4-gangs ALPS Blue Velvet pot are in the proximity of the transformer and power rails. Same applies to the 4 big input caps. Shorting to GND the XLR inputs will cancel any mains hum. Shielding the transformer on that specific side (with the analogue input PCB traces) will also lower the mains hum.

The HPA-3B starts clipping most likely due to the 16 SMD transistors from the PCB (pre-amps). Also, the internal SMD parts are not so beefy like inside the Audeze Deckard (TTH parts), so after I did some tests I found couple of resistors being "browny" and I swapped those with TTH ones. I also swapped the 16 SMD transistors for TTH ones, but I honestly don't feel like re-doing the "max. output power" test again.

 

[EB1000]

Hi

Are these genuine Canare starquad XLR cables??

http://www.ghentaudio.com/part/a03.html

The info is confusing. It says Canare, and maked "made in Japan" on the cable, but then it says manufactured in China. Are Ghent cables legit or fake?

Thanks

 

[GokieKS]

Hi

Are these genuine Canare starquad XLR cables??

http://www.ghentaudio.com/part/a03.html

The info is confusing. It says Canare, and maked "made in Japan" on the cable, but then it says manufactured in China. Are Ghent cables legit or fake?

Thanks

Ghent Audio is big and well known enough that it's very unlikely that they will be using counterfeits, so most likely it means the raw/unterminated cable is Canare L-4E6S (and made in Japan), but the manufacturing process (of cutting them to length and then terminating to XLR-M / XLR-F) is done in China.

 

[Veri]

Ghent Audio is big and well known enough that it's very unlikely that they will be using counterfeits, so most likely it means the raw/unterminated cable is Canare L-4E6S (and made in Japan), but the manufacturing process (of cutting them to length and then terminating to XLR-M / XLR-F) is done in China.

Exactly Ghent makes great DIY amplifier kits, DC cables and interconnects. Never had any problem.
(Considering Chinese new year though they might be not very responsive at this time).

 

[Koeitje]

I got a pair of Ghent RCA cables, very solid and a good price. Sure you can probably go even cheaper, but $20 for a meter or so is not bad at all.

 

[Paperdragons]

Sorry, my bad...but the topic, as written, was broad. Didn't know it was only headphones.

BTW, does the typical audiophile listen in front of his PC, over cans, these days?

Guess I do

 

[eboleyn]

I'm glad I found this thread. I'd recently looked for an article on balanced headphone cables and generally found nothing actually useful until this thread.

...

I do indeed still think that most of the "balanced headphone" effects are snake-oil, but the crosstalk thing is no joke and had a major effect for me when testing at audiophile quality levels. At this point I'm now deeply irritated that manufacturers use 3-wire for headphones as the default when they could use 4 wire and cheaply get better quality.

Anyway, thanks again for this thread for confirming my results!

Picking up this thread again to mention yet another effect related to the crosstalk issue and the superiority of using 4-wire "balanced" vs. 3-wire "unbalanced" headphone cables.

Specifically, if you use a resistor-based "headphone volume attenuator" which is the standard way to do attenuation for a headphone, the balanced version works as intuitively expected while the unbalanced version exacerbates the crosstalk problem mentioned above, depending on where it's placed in the cabling chain.

A quick review of how impedance-neutral (well, mostly neutral!) headphone volume attenuators work (excuse the bad ASCII art):

Balanced version (right side, but both are the same):

Amp +R ------(resistor1)--------+---------(to headphone +R)----> | (resistor2) | Amp -R ------(resistor1)--------+---------(to headphone -R)---->

Note that both sides use the same value of "resistor1", but the actual values of resistor1 of resistor2 are related as follows:

• resistor2 roughly ends up being the apparent output impedance value seen by the headphones, so you keep it small to work well with most headphones, say 1-2 ohms.
• resistor1 would want to be a larger value, as 2x resistor1 + resistor2 is roughly the input impedance of the headphone to the amp, say 8-ish ohms for resistor1.
• The value of 2x resistor1 divided by resistor2 roughly gives you the level of attenuation, so these example values would give 16x

The unbalanced version is identical to the above but for the negative leg it is just the ground wire. The lower side of resistor2 then just connects to the ground wire, like so:

Amp +R ------(resistor1)--------+---------(to headphone +R)----> | (resistor2) | Amp ground ---------------------+-----(to headphone ground)---->

The resistor values follow similar rules for the balanced version, but the apparent headphone input impedance would be resistor1 + resistor2, so you'd want a larger value for resistor1, say 16-ish ohms.

OK, given the above, why would the crosstalk potentially increase when using this and connecting a headphone to a headphone amp?

The crosstalk effect is normally based on the headphone input impedance vs the resistance of the shared ground wire. In lower impedance headphones this can become a real factor, especially when using headphone extension cables.

However, in the presence of a headphone attenuator...

First, as the current drawn through the ground network increases, it increases the crosstalk effect. The headphone amp "attenuation" works by dumping extra current through the resistor network between + and ground on both sides of the stereo pair at much higher levels than it would normally have, so the voltage of that point where the resistor2 touches ground is lifted over the level it would have been without the attenuator, distorting the audio.

Second, if the resistor1 + resistor2 value gives an apparent headphone impedance value much lower than the actual headphones you're using, then again the crosstalk effect is larger. The generic "impedance neutral" kind of headphone attenuator you might purchase would try have a low value of resistor2 to be most compatible with a wide range of headphones, so therefore have a lower overall input impedance such as 16 ohms or so.

When you have an unbalanced amp but have 4-wire "balanced cabling", treat the balanced version above as tying the right and left "Amp -" together, which has some crosstalk but it's typically tied much closer to the amp and the "shared ground resistance" is minimized.

In summary, again the 4-wire solution gives a superior result to the 3-wire solution, regardless of the presence of a balanced amp or not, though in this case an actual balanced amp can in principle completely factor out the crosstalk effect, which is nice!