balanced is inherently flawed


A recent post asking for opinions on balanced vs. single ended got me thinking once again about the inherent flaws in a balanced scheme.

A balanced signal has 2 parts called plus (+) and minus (-) that are equal in voltage but opposite in polarity. Therefore a balanced amp is really 2 single ended amps in one package, one for the + singal and the other for the - signal. So a balanced amp using the same quality parts as a single ended amp will be twice as expensive. Strike one.

That brings us to the "equal but opposite" notion. In order for this to work as planned, all of the + stages and cables connecting them must be exactly equal to all of the - stages all the way through the source, preamp, and power amp. Any deviation from the + stage being the exact mirror image of the - stage will result in an imbalance. Since perfect symmetry cannot be achieved, especially with tubes, distortions are introduced. Strike two.

Some think that balanced has to be better for various reasons that include:

1. If they hook up a balanced device using single ended cables they loose some gain.
2. They think a balanced system can achieve a lower noise floor.
3. They have balanced equipment and it sounds better when they hook it up with balanced cables vs. single ended cables.
4. It's used in recording studios by the pros so it must be better.

These arguments are flawed for the following reasons:

1. More gain does not equal better sound. Of course you need enough gain to drive your speakers to satisfactory levels, but the fact that one connection has higher gain than another has really nothing to do with sound quality.

2. This is the most misunderstood of all. A balanced amp CAN reject noise that is coming in through the interconnects. However, it can do nothing to reject or cancel the random electrical noise that comes from within the devices inside the amp. A balanced amp has no advantage over a single ended one when it comes to the major contributor of noise in the system, that which is generated inside the amp. The rejection of noise from cables relies on the fact that it is generally equal to both the + and - inputs and is therefore cancelled, but since the noise voltages generated by the devices inside the + and - stages in the amp are random and unrelated, they do not cancel and are passed on to the next stage.

Furthermore, since well designed, shielded interconnects of any type are very good at rejecting electrical noise from the outside, balanced has no advantage except in very noisy enviroments or when using very long runs, both of which apply to recording studios, not to typical home systems.

3. Since a truly balanced amp was built from the ground up to operate in a balanced mode, it makes sense that it will sound worse when fed a single ended signal. That doesn't mean that balanced is better, just that that particular amp sounds better when fed a balanced signal.

If you subscribe to the theory that more money can get you better performance, and since a single ended amp has 1/2 as many components as an equivalent balanced amp, it stands to reason that if the designer put as much money and effort into designing a single ended amp, it would sound better.

4. See 2 above.

And this brings us to our last point. ALL sound sources are single ended. Whether from a plucked string, blowing air through a horn, the human voice, or anything else; the resulting increses and decreases in air pressure that we perceive as sound are single ended. There is no "equal but opposite" waves of pressure. This is also true when the signal finally gets to a loudspeaker. There are no "equal but opposite" pressure waves coming from the speaker. It is a single ended device.

In a balanced system these pressure variations are picked up by a microphone and then some where along the line converted to balanced. A phonograph record is encoded single ended as is a digital disc. Your CD player may have a balanced output but the data that is read from the disc is single ended and then converted. In order not to introduce ditortions, this conversion from single ended to balanced has to be done perfectly. And since it can't be, strike three.
herman
Lets explore your reasons why audiophiles benefit from balanced designs:

"Lower noise
Lower distortion
cable immunity
wider bandwidth"

If lower noise and distortion are the holy grail that you make them out to be, why are you using tubes when they are inherently noisy? You can achieve even better noise and distortion figures with transistors. We went down the "lower distortion is better" road in the 70's until the end result was a bunch of high powered amplifiers with .00001% distortion that sounded like crap. I won't disagree that tubes can sound very pleasing and I would use them if I had the patience to deal with them, but you have to attribute that pleasing sound to some kind of inherent distortion, which sort of contradicts your lower distortion is better arguement.

If immune to cable differences and "sounded excellent no matter what cable you used", why do I and thousands of others hear significant differences when switching cables?

Wider bandwidth than what? The 100 kHz that your amps do is easily achieved by single ended designs.
Hi Herman

I see that you do not understand how balanced signals works. That's OK. Here's a short primer:

Any signal source (transducer: either electromechanical, inductive or light) that does not use ground as a reference is a balanced source. What this means is that as long as one of the output leads of the device will work as well as the other, and switching them only means that phase has been reversed, then the device is a balanced source.

Phono cartridges are like this, tape heads are like this, microphones are like this, light sensing pickups are like this. So *IF* you wanted to, you could use the output of a laser pickup in the balanced domain. There are CD drives that do this out in the field and in service; I read about a company doing this over ten years ago! I don't keep up on digital technology as much as analog so I don't know of current (no pun intended) examples. Now the fact that this is the case is probably not really very important, since the analog signal that comes off the light pickup is interpreted as a digital signal fairly close to the pickup itself.

Herman, to say that what happened with Mercury in 1958 has nothing to do with how your system operates is flat wrong. Here's how I know: What does you stereo sound like with nothing playing on it? Hopefully, silence, or very near that. Where does the recording come from that you play? A: The record labels, unless *ALL* the recordings you play are made by you or your friends. IF not, then the effects of balanced line technology are all around you, in every commercial recording you play. Its inescapable. Therefore it is impossible that this has no effect on your stereo, unless you just plain and simple do not play recordings on it.

That we are not talking about opinion is very simple as the benefits of balanced line are one of the few areas of audio that are both audible and quantifiable. There are plenty of ways to 'determine what sounds better', and they can be quantified as long as there is enough experience to detirmine what measurements correllate with the sounds that we hear (poor Julian...).

If this is not obvious from the equipment that you have heard, it might be that that gear has its own weaknesses that have nothing to do with blanced operation. The reason I say this is that the legacy of 45 some odd years of hifi recordings is a rather massive testemony to the success of the technology. A few high-end audio failures are not suficient to be its defeat.

My advice is to listen to balanced gear that is done right, as I am sure you will agree that it is importanat to listen to single-ended gear that is done right for the same reason!

At least we both agree about Mr. Hirsch!
Herman- Sounds like you have a nice system. Do you ever listen to it? And one more time, being that you are a Naim owner, Naim uses DIN connectors "to eliminate earthing problems and improve CMR" (go to their website or give them a call). As you have stated "Naim is not fully balanced", but conceptually the use of balanced (DIN) connectors provides much the same result. Give it up already; as I said above "no one is twisting your arm".
OK- good questions. Here we go:

I use tubes because the distortion they generate is less annoying to the human ear then the distortion of transistors. This has to do with odd-ordered generation that transistors do. Even in vanishingly small amounts, the ear is sensitive to this. Tubes don't make nearly as much of this distortion.

To get rid of the noise of tubes, I use balanced differential circuits, and acheive noise figures 5-5.5db less per stage of gain then I would have if operated SE with the same topology. In a preamp with 3 stages of gain, this is close to 18 db less noise as compared to the same topology in a single-ended embodiment. Don't tell me you can't hear that! Reducing the noise floor means more detail, and detail is a highly desirable characteristic of a good system.

You hear differences in cables due to several factors:
1) termination impedance
2) characteristic impedance of the cable.

The two are related. Since most SE cables are terminated with 100K or more in most cases, the capacitance of the cable can easily interact with the input impedance of the amplifier to result in a high frequency rolloff that is quite measureable and audible. This introduces measureable and audible phase shift. Since different cables have different capacitances, the roll-off and phase shift is different from cable to cable. You also have termination losses due to the connectors themselves, and reflections due to the characteristic impedance of the cable being poorly terminated by the input of the amp.

With balanced lines, the cable issues are swamped by the impedance of the driver circuit and also by the input impedance of the amp, which if done properly, are both quite low. With low impedances, the capacitances and other cable issues are pushed out to frequencies well outside the audio band (and the propogation delay in the cable is improved too- up to 8X better). This is easily heard and measured.

You may get 100KHz out of SE designs, although most tube based SE power amps will not do anywhere near that, but the real issue is can you get that signal *into* the amp from a distance. You might be able to over a meter or two, but not if the cable is ten meters long. Over that length, the high frequency roll-offs will be too severe. You need balanced lines to do better.

If you wonder why I brought up Mercury, the point was that a *microphone* signal went over 150 feet to get to the tape recorders. Just go ahead and see of you can get a hifi signal out of a single-ended cable that is 150 feet long! You can't. Now I know that no-one uses cables that long at home, but what happens if you are interested in that last degree of nuance that is possible? Well I can tell you quite simply that the better your cable is, even if its only 3 feet, the more nuance you will transmit. And audiophiles are all over that nuance stuff. That's what makes good things happen in good systems. If you want that last degree of nuance, you have to use a balanced sytem to get it, particularly if you have any real length in the cables.
Two things Ralph.

"Not that I'm saying that SE is bad, just that balanced (done right, which is not that hard) is better and *any* audiophile can hear it. I had a girlfriend who was deaf in one ear, and half deaf in the other, and *she* could hear it, so I am confident you can too if it give it a straight shot."

Someone that is highly respected, both within the Audiogon forums and audio industry on the whole, publicly commented that they did not like your preamps at ALL. On top of that, the preamp that they ended up running is being utilized in single ended mode. Given that they've run your amps and found them to be a quality product, how would you respond to that?

As a side note, a reduction in noise of -18 dB's is equivalent to having 1/64th the noise floor that you started out with. As Herman points out though, if the noise floor is already well below audibility, how much of a benefit is it?

There's another thing that you're overlooking Ralph. That is, ALL amplifier / speaker interphases are "single ended". Why this is, i don't know.

If you think about it, a "balanced" dynamic driver would be simple to make AND offer drastically improved transient response. By using two coils in parallel ( one positive and one negative ) and driving them with a common center tap ( ground ), you would literally have ( near ) instant acceleration and deceleration potential. When one motor was pulling, the other would be pushing, etc... They would end up sharing the load AND cancelling out / drastically reducing the effects of reflected EMF. Why hasn't anyone done something like this? While it would take an all new speaker AND amplifier design, i see no reason why it couldn't be done successfully.

Once again, the "high end" and "high tech" audio engineers / manufacturers "sleeping on the job". Sean
>

PS... I'm a "more is better" kind of guy, so i'll take any reduction in noise / increase in power that i can get. Having said that, only one of my systems is running in balanced mode.

PPS... An E-stat would lend itself to "balanced drive" also with no need for a transformer. One could use high voltage output devices in the amplifier and directly charge the two plates and mylar. By varying the positive and negative potential between the plates, the mylar, which would be grounded, would be repelled and / or attracted uniformly across the entire sheet in both directions. Only problem here is that you would have one helluva safety hazard to deal with when energized / playing music.