True Balanced outputs
- Status
1 - 15 of 15 Posts
Joined
·
1,028 Posts
The difference lies in the internal circuitry. True balanced circuitry divides the signal into positive and negative phases and carries it on seperate pathways relative to ground. Lots of "fully balanced" equipment does not truely have differential circuitry but instead uses a transformer or opamp or other active circuit to add balanced outputs/inputs onto single ended internal circuitry.
If you want to learn more, look at www.passdiy.com and read the tutorials on the "Son of Zen" and "Bride of Son of Zen". Great examples of truely differential circuitry.
If you want to learn more, look at www.passdiy.com and read the tutorials on the "Son of Zen" and "Bride of Son of Zen". Great examples of truely differential circuitry.
Joined
·
3,947 Posts
Pick me! I can! Pick me!
True balanced outputs have two non-grounded signal conductors (signal + and signal –), which contain the same signal, but of opposite polarity (phase). The signal voltage is twice that of either conductor to ground, very much like a bridged amplifier pair.
A (usually shielded) twisted-pair interconnect is used to transfer the signal to the destination component, usually an amplifier with a differential amp input stage, which allows for common-mode noise rejection.
A 'non-true-balanced' output has a signal (signal +) that is referenced only to ground, even if transferred via a shielded twisted pair. While the shield may be grounded at only one end, the conductor that is tied to ground is connected at both ends, and doubles as the signal –.
A side advantage of the true balanced signal path is that the higher circuit voltage provides a 3dB signal-to-noise ratio improvement. Ideally, both the source and destination components contain truly symmetrical circuitry, which can be carried all the way to the amp outputs.
True balanced outputs have two non-grounded signal conductors (signal + and signal –), which contain the same signal, but of opposite polarity (phase). The signal voltage is twice that of either conductor to ground, very much like a bridged amplifier pair.
A (usually shielded) twisted-pair interconnect is used to transfer the signal to the destination component, usually an amplifier with a differential amp input stage, which allows for common-mode noise rejection.
A 'non-true-balanced' output has a signal (signal +) that is referenced only to ground, even if transferred via a shielded twisted pair. While the shield may be grounded at only one end, the conductor that is tied to ground is connected at both ends, and doubles as the signal –.
A side advantage of the true balanced signal path is that the higher circuit voltage provides a 3dB signal-to-noise ratio improvement. Ideally, both the source and destination components contain truly symmetrical circuitry, which can be carried all the way to the amp outputs.
So-called "true balanced" circuitry is a misnomer and a red herring. It's more properly called "differential". Given that, it's about as useful to balanced interconnection as it is to unbalanced interconnection. In other words, it really doesn't matter. The reason is that balanced interconnection depends on the common-mode impedances at the inputs or outputs to be matched, or balanced, in order to do its job. That is it --- it doesn't depend on your component's internal circuit topology, nor should it ever.
The reason it should not depend on your internal topology is because the job of a balanced input is to cancel common mode noise at the input, before it gets into the rest of the component --- this is well-known in any electronic design field that's trying to shield against outside noise. One of the worst designs and persistent myths in high-end audio is that you need separate amplification of both balanced signals, because, along with allowing common-mode noise into your component, which is difficult to remove once it's inside, you also end up taking up headroom with the common mode signals that should have been canceled at the input. In effect, you end up paying for twice the components and the design becomes worse.
There are many ways of getting balanced or matched input or output impedances --- active balanced circuitry, transformers, or passive matching. There are advantages and disadvantages to these methods, and which one you use depends on your application. The extra +6dB of voltage gain for some kinds of balanced interconnection is a side effect, and not a primary effect of balanced interconnection.
There are also good reasons to use an internal differential topology, but accommodating balanced inputs or outputs isn't one of them.
The white papers at the Jensen Transformers website gives good explanations, technical and otherwise, of balanced interconnection:
http://www.jensentransformers.com
Especially relevant is application note AN-003 .
--Andre
The reason it should not depend on your internal topology is because the job of a balanced input is to cancel common mode noise at the input, before it gets into the rest of the component --- this is well-known in any electronic design field that's trying to shield against outside noise. One of the worst designs and persistent myths in high-end audio is that you need separate amplification of both balanced signals, because, along with allowing common-mode noise into your component, which is difficult to remove once it's inside, you also end up taking up headroom with the common mode signals that should have been canceled at the input. In effect, you end up paying for twice the components and the design becomes worse.
There are many ways of getting balanced or matched input or output impedances --- active balanced circuitry, transformers, or passive matching. There are advantages and disadvantages to these methods, and which one you use depends on your application. The extra +6dB of voltage gain for some kinds of balanced interconnection is a side effect, and not a primary effect of balanced interconnection.
There are also good reasons to use an internal differential topology, but accommodating balanced inputs or outputs isn't one of them.
The white papers at the Jensen Transformers website gives good explanations, technical and otherwise, of balanced interconnection:
http://www.jensentransformers.com
Especially relevant is application note AN-003 .
--Andre
Joined
·
1,028 Posts
When discussing audio components, the difference between the true balanced outputs and not truely balanced is the way the balanced signal is derived. The addition of a transformer or opamp to the signal path does convert a single-ended signal a balanced signal, but does not create a truely balanced component. This commonly referred to as "pseudobalanced".
I've never heard anybody argue that using transformers on the input/output signal is better than true differental topology. I do think that if you use really good transformers, like Jensen, that the resultant sound is as good as true differential topology. I think this is what Jeff Rowland does with his equipment. I don't know anybody who really likes opamp-based converters too much, there's just too much noisy active circuitry there- unfortunately it is very cheap and very commonly implemented.
Anyway, I'm not trying to start an unwinnable argument about which is better. Only poining out to the original poster what the variables are.
Also see http://www.rane.com/note110.html for a little more light reading!
I've never heard anybody argue that using transformers on the input/output signal is better than true differental topology. I do think that if you use really good transformers, like Jensen, that the resultant sound is as good as true differential topology. I think this is what Jeff Rowland does with his equipment. I don't know anybody who really likes opamp-based converters too much, there's just too much noisy active circuitry there- unfortunately it is very cheap and very commonly implemented.
Anyway, I'm not trying to start an unwinnable argument about which is better. Only poining out to the original poster what the variables are.
Also see http://www.rane.com/note110.html for a little more light reading!
Quote:
Whether you use op-amps or discreet components you still need active components in a differential design. All amplifiers have active circuitry.
| Originally posted by Bill Lummus I don't know anybody who really likes opamp-based converters too much, there's just too much noisy active circuitry there- unfortunately it is very cheap and very commonly implemented. |
As I've understood it: There are two basic kind of audio signals, single-ended (ground reference and signal) and differential (ground, signal- and signal+). Further, a differential signal can also be balanced, meaning the + and - signals are always of equal amplitude. In most cases, differential and balanced signals are just as good and not easy to tell apart, which has resulted that often people talk about "balanced", when they really mean "the signal in the XLR connector" - very seldom you do or need to know if it is really balanced or just differential.
For example, most devices are single-ended internally. To derive a differential output, the - line is the +line signal driven by an inverting op-amp (or the same thing in one IC called a line driver). If you load only one signal line, the other signal is not affected - the signal is not balanced. If the circuitry keeps both sides the same amplitude no matter what, it is really balanced.
For example, most devices are single-ended internally. To derive a differential output, the - line is the +line signal driven by an inverting op-amp (or the same thing in one IC called a line driver). If you load only one signal line, the other signal is not affected - the signal is not balanced. If the circuitry keeps both sides the same amplitude no matter what, it is really balanced.
Bill,
For whatever reason, audiophiles seem to have a perjorative view of "pseudo balanced" circuitry (even the name implies that it's not really balanced), when in fact, it doesn't really matter how it was derived as far as the functionality of a balanced interface goes. I was trying to dispel the notion that somehow fully-differential circuit designs are essential to getting full use out of a balanced connection, when they are not.
--Andre
For whatever reason, audiophiles seem to have a perjorative view of "pseudo balanced" circuitry (even the name implies that it's not really balanced), when in fact, it doesn't really matter how it was derived as far as the functionality of a balanced interface goes. I was trying to dispel the notion that somehow fully-differential circuit designs are essential to getting full use out of a balanced connection, when they are not.
--Andre
Joined
·
1,028 Posts
I don't have enough information to judge which is better, so I don't intend on reinforceing or dispelling any notions either way. It's easy for me to see the advantages to true balanced circuitry. I also see that there are several good transformers available with very flat frequency responses. Opamps- well they are a bit more of a wild card. I've listened to a lot of opamps in a lot of implementations with quite variable results. I think I would rather just have single-ended connections than an opamp output stage.
Maybe someday I'll build some and listen for myself...
Maybe someday I'll build some and listen for myself...
Quote:
I think there is also some misinformation on op-amps. Back in the early 80s op-amps where new, not really tweaked for audio, and sounded bad. Nowadays, it seems there are some really good low noise, low gain op-amps specifically designed for audio. Its been a very long time since I played with op-amps, but I find it hard to believe that these modern audio op-amps are not vastly better than what people think. Opinions in audio tend to persist well beyond their usefulness. JMHO
| Originally posted by Bill Lummus Opamps- well they are a bit more of a wild card. I've listened to a lot of opamps in a lot of implementations with quite variable results. I think I would rather just have single-ended connections than an opamp output stage. Maybe someday I'll build some and listen for myself... |
Joined
·
1,028 Posts
I've played with op amps quite a bit. Some are pretty good, particularly in low gain settings (like preamps and headphone amps). But in terms of active circuitry, the opamp is MUCH more active than the transformer and the potential for distortion is greater.
You are right, however, about the progressing quality of op amps.
You are right, however, about the progressing quality of op amps.
Poor design and mediocre construction techniques are the hallmarks of inferior equipment. When you are lazy, sloppy, and cheap, op amps can provide the easy way out. This does not mean well designed equipment, with high quality parts and construction techniques should not contain any op amps. To the contrary, sometimes a well selected op amp will out perform all other practical solutions.
Poor design and mediocre construction techniques are the hallmarks of inferior equipment. When you are lazy, sloppy, and cheap, op amps can provide the easy way out. This does not mean well designed equipment, with high quality parts and construction techniques should not contain any op amps. To the contrary, sometimes a well selected op amp will out perform all other practical solutions.
Nice thread! I was thinking of a Rane or digital Behringer or even BSS active crossover and the balanced input output compatibility issue. I wanted 4th order crossovers, phase and volume attenuation with subwoofer and mains/satellites.
I want to use an RME sound card with 2 stereo pairs for 2.1 (it’s even expandable/switchable for 5.1 w/ 3 channel pairs). I can adjust volume from the sound card (digitally and set switches). No preamp, the sound card IS the preamp and attenuation rivals any active or passive preamp IMO.
I decided to run one channel pair straight to the mains, use sealed enclosures (18" corner placed sub and full range mains overlapped octave 40-50Hz crossover) and a Jensen cap for 6+12=18dB/octave passive crossovers w/sub and mains. Identical amps (maybe bridging to run each main channel), and keep the short unbalanced cable runs.
Since now there is much less crossover distortion, phase and latency delay (speaker time alignment) vs. active crossovers and bass reflex speakers I won’t be using that “stageâ€. I can correct for relative phase with a simple all pass filter (180 degrees).
I want to use an RME sound card with 2 stereo pairs for 2.1 (it’s even expandable/switchable for 5.1 w/ 3 channel pairs). I can adjust volume from the sound card (digitally and set switches). No preamp, the sound card IS the preamp and attenuation rivals any active or passive preamp IMO.
I decided to run one channel pair straight to the mains, use sealed enclosures (18" corner placed sub and full range mains overlapped octave 40-50Hz crossover) and a Jensen cap for 6+12=18dB/octave passive crossovers w/sub and mains. Identical amps (maybe bridging to run each main channel), and keep the short unbalanced cable runs.
Since now there is much less crossover distortion, phase and latency delay (speaker time alignment) vs. active crossovers and bass reflex speakers I won’t be using that “stageâ€. I can correct for relative phase with a simple all pass filter (180 degrees).
1 - 15 of 15 Posts
- Status
Join the discussion
