|
|
Post by Boomzilla on Jan 22, 2020 16:09:04 GMT -5
Thanks, KeithL - I was previously under the misconception that "balanced" meant that the positive part of the waveform was on the + lead and the negative swing on the - lead. But actually, the entire waveform is on both leads - just inverted in phase. Makes sense now - thanks again. |
|
|
|
Post by dimitryz on Jan 22, 2020 18:01:44 GMT -5
Also thank you for reminding me that bridge XPA-2 has TWO balanced channels! Hence its rather massive 1kw output in bridged mode...and it runs cool to boot!
|
|
|
|
Post by Gary Cook on Jan 22, 2020 22:05:35 GMT -5
Thanks, KeithL - I was previously under the misconception that "balanced" meant that the positive part of the waveform was on the + lead and the negative swing on the - lead. But actually, the entire waveform is on both leads - just inverted in phase. Makes sense now - thanks again. I think of an audio signal as AC, so there's a "normal" signal and an "inverted" signal, both of which contain the complete audio signal. One is simply 180 degrees out of phase with the other. So one could hear all of the music while only using one of them. Of course one would be in phase and the other would be 180 degrees out of phase. One day when I see keith face to face I'll ask him to help me understand why the we lose the discrete/differential ability when we use the high pass filter. It seems to me that it would be relatively simple to have 2 high pass filters (per channel), one for the in phase signal and one for the out of phase signal. That would maintain the discrete/differential ability. Cheers Gary |
|
|
|
Post by dimitryz on Jan 22, 2020 22:30:41 GMT -5
Thanks, KeithL - I was previously under the misconception that "balanced" meant that the positive part of the waveform was on the + lead and the negative swing on the - lead. But actually, the entire waveform is on both leads - just inverted in phase. Makes sense now - thanks again. I think of an audio signal as AC, so there's a "normal" signal and an "inverted" signal, both of which contain the complete audio signal. One is simply 180 degrees out of phase with the other. So one could hear all of the music while only using one of them. Of course one would be in phase and the other would be 180 degrees out of phase. One day when I see keith face to face I'll ask him to help me understand why the we lose the discrete/differential ability when we use the high pass filter. It seems to me that it would be relatively simple to have 2 high pass filters (per channel), one for the in phase signal and one for the out of phase signal. That would maintain the discrete/differential ability. Cheers Gary I thought the same thing. However, for very low distortion, both filter circuits have to be identical and very stable over temperature and time. I think this represents a technical challenge and expense. The ICs for converting unbalanced signal to balanced are both good and cheap. So it is cheaper and easier to drop out of balanced, filter with an unbalanced circuit, where imprecision is just a small difference in filter frequency performance and convert back to balanced with an IC. I am sure Keith can explain the exact reasoning. |
|
|
|
Post by Gary Cook on Jan 23, 2020 1:03:42 GMT -5
I think of an audio signal as AC, so there's a "normal" signal and an "inverted" signal, both of which contain the complete audio signal. One is simply 180 degrees out of phase with the other. So one could hear all of the music while only using one of them. Of course one would be in phase and the other would be 180 degrees out of phase. One day when I see keith face to face I'll ask him to help me understand why the we lose the discrete/differential ability when we use the high pass filter. It seems to me that it would be relatively simple to have 2 high pass filters (per channel), one for the in phase signal and one for the out of phase signal. That would maintain the discrete/differential ability. I thought the same thing. However, for very low distortion, both filter circuits have to be identical and very stable over temperature and time. I think this represents a technical challenge and expense. The ICs for converting unbalanced signal to balanced are both good and cheap. So it is cheaper and easier to drop out of balanced, filter with an unbalanced circuit, where imprecision is just a small difference in filter frequency performance and convert back to balanced with an IC. I am sure Keith can explain the exact reasoning. Would it be that much different to the requirement for the left and right channels to be identical and stable? Cheers Gary |
|
|
|
Post by dimitryz on Jan 23, 2020 9:19:07 GMT -5
I thought the same thing. However, for very low distortion, both filter circuits have to be identical and very stable over temperature and time. I think this represents a technical challenge and expense. The ICs for converting unbalanced signal to balanced are both good and cheap. So it is cheaper and easier to drop out of balanced, filter with an unbalanced circuit, where imprecision is just a small difference in filter frequency performance and convert back to balanced with an IC. I am sure Keith can explain the exact reasoning. Would it be that much different to the requirement for the left and right channels to be identical and stable? Cheers Gary I think it is. Left right channel balance is fine at ~0.25db, but a similar filter discrepancy on a balanced circuit will produce a noticeable amount of distortion, at least in measurements. I think a proffessional should weigh in here. |
|
KeithL
Administrator 
Posts: 10,261
|
Post by KeithL on Jan 23, 2020 10:01:50 GMT -5
There are two "reasons" why the tone controls and filters in the XSP-1 aren't fully balanced.
The first reason is that it adds complexity - in several ways. First, since a balanced signal is actually a pair of signals, each passing independently through the system, that about doubles the total number of parts and connections required. (A "two channel fully balanced amplifier or preamp" requires FOUR well matched channels rather than two.)
Second, in many cases, in addition to that, it imposes stricter requirements for an exact match between channels. That isn't that big a deal for some types of circuits, including simple "gain blocks", but it increases the parts cost and design complexity considerably for others.
(For one thing.., very precise capacitors, which are required for tone controls and analog filters, are more costly than equally precise resistors.)
The second reason is more of a context. The main reason for using fully balanced circuitry is to reduce certain types of distortion - and so to deliver "a more exact perfect reproduction of the original signal". And, to be blunt, tone controls and other filters are for people who WANT to alter the signal. So, "purist audiophiles", of the sort who care the most about balanced circuitry, are also the least likely to ever use tone controls and filters. This is why, until a few decades ago, most "true audiophile preamps" didn't include tone controls or bass management. (The logic was that: "good quality tone controls would be expensive to include - and real audiophiles wouldn't use them anyway".)
We've chosen what we believe is a good compromise... The Tone Trims and bass management in the XSP-1 are actually very precise, very low in distortion and noise, and sound just fine... and they don't force you to sacrifice the benefits of balanced inputs and outputs. But we didn't take the final expensive step of designing them with fully balanced internal filter circuitry.
(We just didn't think most of our customers would want to pay several hundred dollars more for "fully balanced tone controls and bass management filter circuitry".
I thought the same thing. However, for very low distortion, both filter circuits have to be identical and very stable over temperature and time. I think this represents a technical challenge and expense. The ICs for converting unbalanced signal to balanced are both good and cheap. So it is cheaper and easier to drop out of balanced, filter with an unbalanced circuit, where imprecision is just a small difference in filter frequency performance and convert back to balanced with an IC. I am sure Keith can explain the exact reasoning. Would it be that much different to the requirement for the left and right channels to be identical and stable? Cheers Gary |
|
KeithL
Administrator
Posts: 10,261
|
Post by KeithL on Jan 23, 2020 10:20:31 GMT -5
That thought had occurred to me - and I'll look into it the next time I have some spare time.
Text descriptions can also be even more confusing because... - some terms like "dual differential" and "quad differential" mean slightly different things to different people - some terms have different meanings in different contexts (for example "balanced amplifiers" and "balanced connections")
I should also point out that "discrete" specifically refers to "using separate parts rather than things like integrated circuits and multi-pack transistors". Before the invention of integrated circuits discrete circuitry was the only choice. And, in the early days of ICs, there were many situations where ICs were cheaper or more convenient to use, but discrete parts delivered better performance. (So, in those situations, "discrete" was the better choice, and "ICs" was the compromise.)
However, these days, that is no longer anywhere near universally true... Discrete circuitry still often offers better heat dissipation (which is why many amps still use discrete components for power applications). However, integrated circuits and component packages often offer much better matching and thermal tracking between components, which is far more critical in other applications.
Therefore a good engineer will always choose the best part for each individual application.
I've also got to say one more thing here ... In my opinion, many audiophiles today are FAR too obsessed with details like circuit topology and component choices... The reality is that there is no specific type of circuitry, or type of part, or type of connection, that will guarantee good performance or great sound...
And, in many cases, a well designed example of a very basic circuit will perform better, and sound better, than a poorly designed example of a "more elegant" circuit. And, after all, in the end tally, the important thing is how it performs, and how it sounds... This type of question come up a lot (a simple balanced output versus fully balanced discrete differential circuitry) and if I may be so bold as to make a suggestion, KeithL I think a simple drawing of the signal path would be useful in explaining the differences. Not a detailed circuit but just a very basic schematic. That way it can be copied whenever the question pops up. Explanations in words are always long winded and people’s understanding of the exact meaning can be quite different. Cheers Gary |
|
KeithL
Administrator
Posts: 10,261
|
Post by KeithL on Jan 23, 2020 10:27:25 GMT -5
Let me clarify something there....
The XPA-2 has two channels... Each channel has a fully balanced fully differential input... But the output of each channel is not fully balanced (each channel has one "hot" speaker terminal and one "ground")... However, AS A PAIR OF CHANNELS, OPERATING IN BRIDGED MODE, you have a fully balanced output as well...
For everyone else, note that we're talking about the old XPA-2... the XPA-2 Gen3 does not have a bridge option. The modern equivalent of "an XPA-2 in bridge mode" would be a Gen3 XPA-DR1.
Also thank you for reminding me that bridge XPA-2 has TWO balanced channels! Hence its rather massive 1kw output in bridged mode...and it runs cool to boot! |
|
|
|
Post by garbulky on Jan 23, 2020 13:32:20 GMT -5
Thanks, KeithL - I was previously under the misconception that "balanced" meant that the positive part of the waveform was on the + lead and the negative swing on the - lead. But actually, the entire waveform is on both leads - just inverted in phase. Makes sense now - thanks again. My DC-1 also has a balanced conversion for its unbalanced analog input. Therefore the output is balanced. However, keep in mind that though going from unbalanced to balanced reduces noise in the circuit, it does not provide ALL the advantage of a fully balanced chain. The advantage of a fully balanced chain is that every step of the way the signal is balanced through all the different gear - DAC - preamp - amp. If you break the chain by having an unbalanced gear (for instance an unbalanced preamp but a balanced dac and balanced amp) you lose the advantage of noise reduction throughout the chain, though you still have the noise reduction present in the individual components that are balanced. So yes there is distortion reduction but no it is not the same distortion reduction as having all balanced gear when you look at the system as a whole. |
|
KeithL
Administrator
Posts: 10,261
|
Post by KeithL on Jan 23, 2020 14:43:49 GMT -5
I should also add to that the fact that neither balanced connections nor balanced circuitry ALWAYS yield an improvement.
This is one of those situations where things aren't always quite as simple as they seem.
A balanced connection has better immunity to noise being picked up by the connection between components.
HOWEVER, this is only relevant if there is noise there to be picked up. So, if you have a long unbalanced connection, and it is picking up hum or stray noise from the environment, then replacing it with a balanced connection may well reduce or eliminate the problem. And, if you plan to run long cables, or run cables in walls, or run interconnects near power lines, a balanced connection is good "insurance" against this sort of problem being discovered after the wire has been sealed in the wall. (This is why balanced interconnects are almost always used in professional applications.)
But, if you are currently using an unbalanced connection, and it is already quiet, or the only noise present is a tiny bit of background noise coming from the source, the a balanced connection is NOT likely to improve anything.
Likewise, while balanced circuitry, when properly designed, can deliver lower levels of specific types of distortion and noise, this doesn't always happen either. Because balanced circuitry requires more parts, there are more parts to potentially contribute other types of noise. And, depending on the design of the source device, and the design of the destination device, extra stages may be required in one or both to convert between unbalanced and balanced. And, on a given piece of equipment, the circuitry on one or the other type of input or output may simply be more well designed. (All of the generalities we've been talking about start with an unstated "all else being equal" - which isn't always the case.)
Another interesting factor is cost... A balanced amplifier or preamplifier requires more parts than an unbalanced version (all else being equal). This means that, assuming the quality and cost of the individual parts used to be equal, the balanced unit is going to cost more. And, assuming equal total costs, since the balanced unit requires more parts, lower cost parts will have to be used, which might result in other performance sacrifices.
The bottom line is that, while we tend to generalize about the benefits of balanced connections and balanced equipment, those benefits aren't always realized in real life.
(And, when comparing equipment that is not otherwise equal, it's really not a good idea to assume that the balanced one will always be better.)
Thanks, KeithL - I was previously under the misconception that "balanced" meant that the positive part of the waveform was on the + lead and the negative swing on the - lead. But actually, the entire waveform is on both leads - just inverted in phase. Makes sense now - thanks again. My DC-1 also has a balanced conversion for its unbalanced analog input. Therefore the output is balanced. However, keep in mind that though going from unbalanced to balanced reduces noise in the circuit, it does not provide ALL the advantage of a fully balanced chain. The advantage of a fully balanced chain is that every step of the way the signal is balanced through all the different gear - DAC - preamp - amp. If you break the chain by having an unbalanced gear (for instance an unbalanced preamp but a balanced dac and balanced amp) you lose the advantage of noise reduction throughout the chain, though you still have the noise reduction present in the individual components that are balanced. So yes there is distortion reduction but no it is not the same distortion reduction as having all balanced gear when you look at the system as a whole. |
|
|
|
Post by Gary Cook on Jan 23, 2020 14:53:23 GMT -5
My stereo 2.1 music system is balanced all the way from the ERC-3 though the XSP-1 to the XPA-1L’s. I’m not sure I could differentiate between the distortion present in my set up if if I “broke the chain” of balancing all the way through by, say, utilising the high pass filter on the FL and FR. What I can absolutely claim to notice is the total absence of noise, the music just speaks to me out of a deep black hole of silence. Once heard it’s quite addictive and listening to music with any noise present, no matter how small, is no longer satisfying.
FWIW I haven't used tone controls for decades, they don’t exist in my lexicon.
Cheers
Gary
|
|
|
|
Post by Boomzilla on Jan 23, 2020 17:36:11 GMT -5
Curiously, my audio amigo, Russell (who, next to garbulky and novisnick) has one of the very best systems I've heard, uses absolutely NO balanced wiring or equipment. He thinks he gets better sound from all unbalanced. And I must say, not only is his system great sounding, but it's also absolutely quiet (despite being fairly complex). |
|
|
|
Post by garbulky on Jan 23, 2020 18:43:41 GMT -5
Curiously, my audio amigo, Russell (who, next to garbulky and novisnick) has one of the very best systems I've heard, uses absolutely NO balanced wiring or equipment. He thinks he gets better sound from all unbalanced. And I must say, not only is his system great sounding, but it's also absolutely quiet (despite being fairly complex). I think it's totally possible to get great sound from unbalanced. There's no reason there can't be! |
|
|
|
Post by Gary Cook on Jan 23, 2020 18:51:34 GMT -5
Curiously, my audio amigo, Russell (who, next to garbulky and novisnick) has one of the very best systems I've heard, uses absolutely NO balanced wiring or equipment. He thinks he gets better sound from all unbalanced. And I must say, not only is his system great sounding, but it's also absolutely quiet (despite being fairly complex). He's just lucky, his house is "quiet", he lives in "quiet" area and his neighbours don't fire up the MIG welder while he's listening to music. And, unlike me, the ice maker in his fridge isn't an EMF generator. Cheers Gary |
|
|
|
Post by novisnick on Jan 24, 2020 1:08:07 GMT -5
Curiously, my audio amigo, Russell (who, next to garbulky and novisnick ) has one of the very best systems I've heard, uses absolutely NO balanced wiring or equipment. He thinks he gets better sound from all unbalanced. And I must say, not only is his system great sounding, but it's also absolutely quiet (despite being fairly complex). He's just lucky, his house is "quiet", he lives in "quiet" area and his neighbours don't fire up the MIG welder while he's listening to music. And, unlike me, the ice maker in his fridge isn't an EMF generator. Cheers Gary Ice maker? Grandma died years ago with the recipe in her head! 😢 Lol |
|
|
|
Post by vcautokid on Jan 24, 2020 2:20:47 GMT -5
Supposedly the XSP and I really believe this to be true remains truly Differential to its core. Unlike many others that use summers and phase splitters at least in it's pure direct form remains fully Differential. Kind of the holy grail in audio. Lonnie talks about the XSP-1 and what Emotiva's goal was in making it especially Gen2. Check out his YouTube video in the Emotiva search. Fun to watch. Lonnie is awesome.
|
|
|
|
Post by selind40 on Jan 24, 2020 6:25:14 GMT -5
I was told if you use Monoblocks, nothing else matters and everyone will like and respect you whether balanced or unbalanced..  |
|
