harmonics and single-ended versus balanced audio cables

[Ken Newton]

http://sanderssoundsystems.com/tech...-balanced-vs-unbalanced-operation-white-paper

I'm surprised to find that professional component vendors are still confused on this issue. In that paper, I found the common conflation of balanced interconnection with differential amplification. Along with personal bias offered as technical rationale. As I've before mentioned, it is fully possible to have a single-ended amplification stage driving an balanced interconnection. For single-ended gain stage afficianados, this could provide the best of both worlds - single-ended sound character plus the ground noise rejection of balanced interconnection.

 

[Ron Resnick]

I suggest that we all make an effort to only use the term 'differential' instead of 'balanced', and 'single-ended' instead of 'unbalanced' when referring to amplification circuits. And likewise, to only use the term 'balanced' instead of 'differential', and 'unbalanced' instead of 'single-ended' when referring to signal interconnections.

Do the technical people agree that "balanced" signal interconnection alone properly describes, from a technical point of view, the cancelation of common-mode noise?

 

[Ron Resnick]

I'm surprised to find that professional component vendors are still confused on this issue. In that paper, I found the common conflation of balanced interconnection with differential amplification. Along with personal bias offered as technical rationale. As I've before mentioned, it is fully possible to have a single-ended amplification stage driving an balanced interconnection. For single-ended gain stage afficianados, this could provide the best of both worlds - single-ended sound character plus the ground noise rejection of balanced interconnection.

I found the Sanders piece confusing!

 

[DaveC]

I'm surprised to find that professional component vendors are still confused on this issue. In that paper, I found the common conflation of balanced interconnection with differential amplification. Along with personal bias offered as technical rationale. As I've before mentioned, it is fully possible to have a single-ended amplification stage driving an balanced interconnection. For single-ended gain stage afficianados, this could provide the best of both worlds - single-ended sound character plus the ground noise rejection of balanced interconnection.

That's fair but overall I agree with him.

The one area I don't agree is the subjective results, IMO often balanced systems do sound dry and mechanical as a result of distortion spectra. For that reason I think balanced only works out well when the overall distortion is kept extremely low, which rules out simple no-feedback tube circuits, which (also IMO) sound better. So for me, there's a basic incompatibility between my philosophy and implementation that prevents any consideration of the use of balanced/differential anything. I also feel the same about balanced signal transmission, when 99.9% of the time in home systems simple unbalanced, unshielded, short interconnect cables work just fine. Why make things far more complicated than they need to be when the result is going to be inferior anyways?

 

[DaveC]

Just out of curiosity is there any reason an SET designer doesn't fit an input transformer to convert balanced input to single-ended? I have a headphone amp where I'm doing the opposite - an input transformer converts a single-ended signal (from a phone) to balanced to drive two complementary SE stages.

Because you should really be using a good SET preamplifier anyways.

Also, good transformers are expensive. The only place I use input transformers is for my active bass section, the DSP takes balanced input so I use a SE > BAL transformer from Jensen, this also galvanically isolates the DSP unit. It has XLR out to a Hypex amp with balanced inputs.

 

[Yuri Korzunov]

Thank you, Ken.

My goal for this thread was to achieve an agreed, consensus technical explanation of the subjective observation posted by the sound of Tao.

I think we have achieved a consensus that this subjective observation is explained technically by the cancellation by a differential circuit of

(i) even order distortion or

(ii) common-mode noise or

(iii) both​

(without necessarily achieving agreement on exactly which one is the answer).

1. Balance circuit suppress external common-mode noise (ii) only.

2. Balance circuit increase level at 6 dB. It can be perceived as sound improvement.

 

[Ron Resnick]

. . . I also feel the same about balanced signal transmission, when 99.9% of the time in home systems simple unbalanced, unshielded, short interconnect cables work just fine. Why make things far more complicated than they need to be when the result is going to be inferior anyways?

In your opinion how short is "short"?

(I fully appreciate the answer depends on the EMI/RFI nature of the environment. The answer for an apartment-dweller in a big city is likely to be a shorter length than the answer for someone living in the country with no interference-generating neighbors.)

 

[Ron Resnick]

What about mixing balanced and unbalanced signal interconnections? Could that ever make sense objectively?

I have always used a balanced phono cable into my (truly balanced connection) Aesthetix Io (Jim White prefers the balanced inputs on the Io). But then I use unbalanced interconnects to my VTL amplifiers (which are not true balanced, and Luke Manley prefers the unbalanced inputs).

 

[DaveC]

In your opinion how short is "short"?

(I fully appreciate the answer depends on the EMI/RFI nature of the environment. The answer for an apartment-dweller in a big city is likely to be a shorter length than the answer for someone living in the country with no interference-generating neighbors.)

You got it! Often longer runs (in the context of a home system) are no problem, like 20-30 ft or so, but it depends.

 

[Ken Newton]

Do the technical people agree that "balanced" signal interconnection alone properly describes, from a technical point of view, the cancelation of common-mode noise?

No. There are other ways to cancel common-mode noise appearing across a signal interface. For example, by utilizing a common-mode choke located in series with an unbalanced interconnection. Common-mode chokes are often utilized in power supply circuits, but may also be utilized directly in the signal interface itself. Such chokes would pass signal current relatively unimpeded, which is in normal-mode, while choking common-mode noise current.

In addition, ground loop type common-mode noise can be broken at the power supply by various means, such as via power common-mode chokes, or via devices which impede noise current flow between a component's signal ground and it's safety (chassis) ground, so long as there isn't a circuit fault. For example, via thermistor, or diode bridge. Upon a fault condition, those devices become a low impedance between the grounds, ensuring that the shock hazard fault protection is activated.

Even with those measures taken, ground loop noise could still couple between the power supplies of components connected in a chain unless all employ effective common-mode power supply rejection methods. For example, your preamp may have effective P.S. ground loop noise rejection, yet, ground loop noise could still couple between an DAC and power amp through that preamp's signal interfaces. The most effective points of protection are at the signal interconnections between components. The most effective common-mode rejection via the signal interconnection points is by utilizing balanced interfaces with transformer coupled component inputs located at each receiving end throughout the chain.

 

[Yuri Korzunov]

In music production using of balanced cables for microphones and guitars (via direct-box) is inevitably. Otherwise there will too noticeble noise

For home audio 50/60Hz with harmonics and other noise may be noticeble for high volume control levels.

If in pause mode we don't listen noticeble noise, using of additional efforts with balanced/unbalanced cables have no sense.

As rule, the noise that appear during playback is not balanced/unbalanced matter.

First need use ferrite rings on cables for decreasing of noticeble noise. It is experimental work: set the ring on cable - check the noise. Better way with panoramic spectrum analyzer.

 

[Ken Newton]

That's fair but overall I agree with him.

The one area I don't agree is the subjective results, IMO often balanced systems do sound dry and mechanical as a result of distortion spectra. For that reason I think balanced only works out well when the overall distortion is kept extremely low, which rules out simple no-feedback tube circuits, which (also IMO) sound better. So for me, there's a basic incompatibility between my philosophy and implementation that prevents any consideration of the use of balanced/differential anything. I also feel the same about balanced signal transmission, when 99.9% of the time in home systems simple unbalanced, unshielded, short interconnect cables work just fine. Why make things far more complicated than they need to be when the result is going to be inferior anyways?

Dave, you can have your cake and eat it too. It can be a quite simple matter to convert a single-ended circuit to correctly drive a balanced interface. For example, say, that your favorite single-ended tube stage features an 1k ohm output impedance that is A.C. coupled through a 1uF capacitor. All you would need do is connect that output to pin-2 of an XLR connector. Then, connect pin-3 of the XLR to the local signal ground via a 1k ohm resistor in series with an 1uF capacitor. Pin-1 of the XLR should be connected to the local chassis (safety) ground. Done.

Now, you have the sound of your single-ended circuit with the ground loop noise rejection of an true balanced interconnection.

P.S. Of course, this assumes an balanced input at the receiving end.

 

[Folsom]

What about mixing balanced and unbalanced signal interconnections? Could that ever make sense objectively?

I have always used a balanced phono cable into my (truly balanced connection) Aesthetix Io (Jim White prefers the balanced inputs on the Io). But then I use unbalanced interconnects to my VTL amplifiers (which are not true balanced, and Luke Manley prefers the unbalanced inputs).

It sounds like you already have your answer, here, Ron. BTW I believe it's at 50ft where people expect SE to take a dive.



Ken, I'm not so sure about subjective performance of using the XLR balanced interface with SE, as I've found shields are rarely in favor of preference on IC's that are SE. The only exception I really have is my TT that has them built it, and let me tell you, I don't have a clue as to why they work on that. Before my current TT I ran unshielded with another TT because it sounded better despite a little hum when it wasn't playing.

 

[Ken Newton]

It sounds like you already have your answer, here, Ron. BTW I believe it's at 50ft where people expect SE to take a dive.



Ken, I'm not so sure about subjective performance of using the XLR balanced interface with SE, as I've found shields are rarely in favor of preference on IC's that are SE. The only exception I really have is my TT that has them built it, and let me tell you, I don't have a clue as to why they work on that. Before my current TT I ran unshielded with another TT because it sounded better despite a little hum when it wasn't playing.

Folsom, I may not have been clear enough in my quickly worded description. I mean, an standard 3-conductor balanced interconnect, with 2 signal conductors and 1 shield conductor. No signal current should flow on the shield conductor. Of course, this assumes an balanced input at the receiving end.

 

[Folsom]

My basic observation remains the same.

 

[Ron Resnick]

I calculate my interconnect length in the listening room at 45 feet.

 

[Folsom]

I suspect that should be ok, the environment isn't very dirty from other electronics and such. And sometime in the near future I'll have a finished product that can be used to reduce the noise picked up down to balanced or lower levels, so that'll become an option. (it exists already, just hasn't been made into the final version)

But I have to say, wow, that's a long run. Mike L's amazing room only takes 28ft~, for example.

 

[DaveC]

Dave, you can have your cake and eat it too. It can be a quite simple matter to convert a single-ended circuit to correctly drive a balanced interface. For example, say, that your favorite single-ended tube stage features an 1k ohm output impedance that is A.C. coupled through a 1uF capacitor. All you would need do is connect that output to pin-2 of an XLR connector. Then, connect pin-3 of the XLR to the local signal ground via a 1k ohm resistor in series with an 1uF capacitor. Pin-1 of the XLR should be connected to the local chassis (safety) ground. Done.

Now, you have the sound of your single-ended circuit with the ground loop noise rejection of an true balanced interconnection.

P.S. Of course, this assumes an balanced input at the receiving end.

Thanks! I was going between something like what you describe vs the Jensen input transformer to get a signal from my tube pre to my DSP. I went with the trafo as I can use either XLR or RCA without needing a switch.

 

[DonH50]

There are a lot of pros and cons among using simple impedance matching, a transformer, or active device to perform conversion between balanced and unbalanced (and the other way if required). Since this thread is about cables, and my test at work is almost done, I'll let that one lie despite the fact that the concepts are commingled in audio (and many other systems). I'll just suggest trying different schemes if you can to decide which is best.

 

[DaveC]

It sounds like you already have your answer, here, Ron. BTW I believe it's at 50ft where people expect SE to take a dive.



Ken, I'm not so sure about subjective performance of using the XLR balanced interface with SE, as I've found shields are rarely in favor of preference on IC's that are SE. The only exception I really have is my TT that has them built it, and let me tell you, I don't have a clue as to why they work on that. Before my current TT I ran unshielded with another TT because it sounded better despite a little hum when it wasn't playing.

TT's are hit or miss, often it seems the cable runs right through the motor's em field and you get hum with unshielded ICs. I actually built identical shielded and unshielded phono cables and measured noise + gathered subjective preferences. It does turn out that if hum isn't an issue the additional capacitance from the shield is more of a problem than noise. Despite that, I shield phono cables as you can never tell if different tt's hum or not, I don't want a customer to buy a new tt and then have a cable that causes hum.

Also, it is possible to build XLR cables without shields... I can make them either way, but for short cables in home systems it's almost never necessary so my standard XLR cables are not shielded. I know some will find that to be questionable, but results are what matter and unshielded cables sound better.