what happens when you put Rfb between output and R 7 and R 8 and R to ground at R7/8 junction. .
??
-RNM
Chris, it is a Theta Casablanca IV with D3 Extreme DACs. It has no single-ended outputs. The DACs are a true balanced design.
When I first asked this question I also mentioned Bruno Putzey's G-word article. It resonates loudly with this novice, particularly as I jump through all sorts of hoops (wiring permutations, shielding etc) to lower noise in my current build.
When I first asked this question I also mentioned Bruno Putzey's G-word article. It resonates loudly with this novice, particularly as I jump through all sorts of hoops (wiring permutations, shielding etc) to lower noise in my current build.
Last edited:
If output connected between R7-R8 the gain drops from -13mdB to -14.6mdB and distortion at 20kHz jumps from 0.51ppm to 0.97ppm.
It works a bit better without that connection.
Any resistor between R7-R8 and the ground will degrade it grossly.
BR Damir
Last edited:
Hi SGK,
That's my first lesson to new technicians. "When is a ground, not a ground?". This usually stumps them. It is a very valuable lesson to learn.
I'm not doubting your equipment is balanced. I'm only pointing out that the fact they are balanced didn't buy you anything 95% of the time. I think the biggest problems we have to get past as we age in any profession is to back away from ideals to put everything into perspective. This is a big idealized issue, balanced is superior, or better in some way. It isn't in most cases. Worse, in consumer audio most balanced signals do not follow the standard for balanced lines. Unequal amplitudes and departure from 180° phase isn't uncommon. If you look at how quickly CMRR drops with amount of amplitude unbalance, you quickly realise that that ideal balanced connection isn't delivering on the goods. My sources have both types of outputs, and they do agree with proper balanced audio standards. I use them in single-ended mode.
-Chris
That's my first lesson to new technicians. "When is a ground, not a ground?". This usually stumps them. It is a very valuable lesson to learn.
I'm not doubting your equipment is balanced. I'm only pointing out that the fact they are balanced didn't buy you anything 95% of the time. I think the biggest problems we have to get past as we age in any profession is to back away from ideals to put everything into perspective. This is a big idealized issue, balanced is superior, or better in some way. It isn't in most cases. Worse, in consumer audio most balanced signals do not follow the standard for balanced lines. Unequal amplitudes and departure from 180° phase isn't uncommon. If you look at how quickly CMRR drops with amount of amplitude unbalance, you quickly realise that that ideal balanced connection isn't delivering on the goods. My sources have both types of outputs, and they do agree with proper balanced audio standards. I use them in single-ended mode.
-Chris
Probably because it's a question that shouldn't really be asked. I'd characterise the issue discussed in the Bruno article differently: Why would you ever want to use something that gets all sorts of cr*p dumped into it (PSU rectifier pulses, by-pass caps etc) as one half of the audio signal? Shed the shackles of "a ground" completely and carry the complete audio signal independently.
Maybe, subject to the dark art of star ground keeping one precious half of the audio signal "clean". Bruno's article touches on the shortfalls of even the best executed star ground structures.
But that is somewhat besides the point. I'm merely surprised that a book on Audio Power Amplifier Design devotes so few pages to true balanced (Bob's terminology) designs, their benefits, issues, typologies etc etc.
I don't doubt that most consumer equipment is single-ended. But there's also a lot of source equipment out there with fully differential outputs only. Going from single-ended to fully differential is easy (connect hot to XLR pin 2 and cold to pin 3 and be done) but going the other way is a PITA, especially if you don't want to suffer a noise penalty from the input buffer. I am building diyAudio member astx's superb SA2013/14 200W 8R amp. It is single-ended. So I first had to build Self's low noise balanced to single-ended input board and then do battle with keeping this ground half of the audio signal clean. It all seems to me to be much more tortuous than it need be.
If the outputs you have are truly balanced then you can short one to ground and use the other for a single ended output. If it is differential then you can use ground and one output.
In either case there is no need to convert. However in the second case the max output level will be half.
Cris is right. Balance circuits are for long haul transmission lines and industrial instrumentation where sometimes the signal levels are in the low micro volts.
However in industrial instrumentation the long haul transmission loops are all current loops.
In the oscillator thread we are dealing with nano volt residual signals. I have found no reason to use balanced circuitry.
In either case there is no need to convert. However in the second case the max output level will be half.
Cris is right. Balance circuits are for long haul transmission lines and industrial instrumentation where sometimes the signal levels are in the low micro volts.
However in industrial instrumentation the long haul transmission loops are all current loops.
In the oscillator thread we are dealing with nano volt residual signals. I have found no reason to use balanced circuitry.
Last edited:
That would demand that the signal is symmetrical. Yet there is nothing to say that it will necessarily be. A differential signal need not be symmetrical.
Even setting transmission between components off to one side for a moment to quote Bruno again:
Anyway, I would have thought true balanced, fully differential designs would have gotten more page bandwidth.
Even setting transmission between components off to one side for a moment to quote Bruno again:
Anyway, I would have thought true balanced, fully differential designs would have gotten more page bandwidth.
Just to be clear. A balanced system has no reference to ground at all. That is part of the fundamental definition of balanced. Even Bruno neglected to mention that. At least in the G word. Best way to understand this is to visualize a transmission path with a single ended transformer at the transmit end and a single ended transformer for the load at the receiving end. There is no reference to ground. If ground is introduced it is no longer balanced. Everything else is differential.
I think that was the point he was trying to make. Don't use the rubbish tip we call ground as a reference for the audio signal. Star ground, star-on-star begone.
The amps I own are "true/fully balanced, fully differential" amps. Krell. The amps I would contemplate purchasing to replace them also just happen to be "true/fully balance, fully differential" designs Just happens to be so. And yet when I read one of the pre-eminent books on power amplifier design just 3 pages are devoted to such typologies. Struck me as odd.
The amps I own are "true/fully balanced, fully differential" amps. Krell. The amps I would contemplate purchasing to replace them also just happen to be "true/fully balance, fully differential" designs Just happens to be so. And yet when I read one of the pre-eminent books on power amplifier design just 3 pages are devoted to such typologies. Struck me as odd.
Last edited:
Hi SGK,
I'm sorry that you are disappointed with the amount of coverage I gave to true-balanced power amplifiers in my book. To be honest, in my experience, there are very few true-balanced input-to-output power amplifiers out there. There is virtually no advantage that I can think of, off-hand, to driving loudspeakers with a balanced signal, so that would leave the issues of a power amplifier receiving a balanced signal in the optimum way, and the issue of whether the use of true-balanced circuitry within the amplifier improves sound quality of the amplifier.
With respect to the balanced link between the preamp and amplifier, I completely understand your concern, especially if your preamplifier does not have a single-ended output. Apart from the possibility that balanced circuitry within the amplifier might produce better sound quality (I tend to doubt it), the matter is usually handled by a balanced-to-single-ended stage in the input of an otherwise single-ended power amplifier, as mentioned by others above.
There are several different ways to accomplish that conversion, and some are not great or optimal. Some do in fact incur an S/N penalty. Most involve the use of op amps, so op amp averse designers might be concerned about that. The single- and triple-op-amp instrumentation amplifier arrangements come to mind. Some arrangements compromise input impedance relationships between the hot and cold sides.
I know of at least one audiophile power amplifier that achieves balanced inputs without use of any op amps by configuring the entire power amplifier like a single-op-amp instrumentation amplifier (also using a discrete JFET unity gain buffer on one input in the process), but I would not go that way. I believe that approach incurs a significant noise penalty.
Do consider the THAT InGenious line receiver IC, an architecture that was patented by Bill Whitlock of Jensen Transformers and an expert on grounding and balanced interconnect issues. Also, for noise, consider using Star Quad shielded balanced cable, of which Bill is also a strong proponent. Often, the shield is left open at one end to thwart the creation of ground loops.
Hope this helps.
Cheers,
Bob
Hi David,
I'd have to say you nailed that explanation well.
Hi SGK,
The only reason you need a ground reference is because the circuits are active, not a transformer. You need to make sure the signals are within the common mode range of the circuits you are using. Then you get into topics like longitudinal voltage offset. The use of a transformer frees you from most of these problems as it naturally is balanced by every definition of the term.
-Chris
I'd have to say you nailed that explanation well.
Hi SGK,
The only reason you need a ground reference is because the circuits are active, not a transformer. You need to make sure the signals are within the common mode range of the circuits you are using. Then you get into topics like longitudinal voltage offset. The use of a transformer frees you from most of these problems as it naturally is balanced by every definition of the term.
-Chris
Hi Bob,
I agree with all your points.
The Adcom GFA-1A is another differential amplifier that people might be more familiar with. The warranty programs get hit hard due to the fact that your average person doesn't understand that the "negative" terminals are still driven. They tend to wire them together in speaker switchers. Of course, Dealers were the worst offenders!
-Chris
I agree with all your points.
The Adcom GFA-1A is another differential amplifier that people might be more familiar with. The warranty programs get hit hard due to the fact that your average person doesn't understand that the "negative" terminals are still driven. They tend to wire them together in speaker switchers. Of course, Dealers were the worst offenders!
-Chris
Hi Bob
Don't get me wrong. If there are good reasons for giving true balanced short shrift that was what I was wanting to understand. Or perhaps I'd expect your views as to such to be contained within the book. The material contains little such opinion or, if it does, apologies as I missed it.
I can accept the arguments such as balanced interconnects are largely unnecessary for consumer audio, but once in it's here to stay and in my case even more invasive in that single-ended is no longer an option. (Bruno of course expresses more strident views in favour of balanced interconnection even in the consumer environment.)
I thought the topic of whether the internal circuitry of the amp can benefit from not using a potentially polluted ground as a reference for the audio signal more interesting. I understand my Krell FPB 200 to be 'fully balanced'. Theta Digital's Citadel was also. Of course, all Bruno's Hypex NC designs are and as a result OEM implementations such as Theta's are. SAE's latest 8 channel amp is also. By no means an exhaustive list but presumably the likes of Dave Reich and Morris Kessler felt there were benefits to fully balanced designs or was it only for marketing purposes? (We know Bruno's views.)
As regards balanced to single-ended conversion, as I previously mentioned, I think I have done a pretty good stab at implementing Doug Self's ultra-low noise "quad LM4562 / quad diff amp 820R LM4562" circuit (Small Signal Audio Design 2nd Ed.) But it very much felt like a massive side-step just to be able to move forward (and hopefully not a step backwards).
It is interesting to observe the internal elegance of a product like the SAE8300 (see website for internal pics) versus the wiring spaghetti associated with a star ground approach. Each channel's PSU and amp are all on one PCB, with four channels stacked on each side.
Hi SGK,
Most home balanced designs are done for marketing value, with small companies more driven by the designer's personal feelings on the subject. Often they do not possess the talent and funds that larger firms invest in their products. Sometimes they are right, and some times they aren't. The only important thing is that it performs well for you, and the same thing goes for someone else who stays single-ended. Like me!
Sobering thought. I once took a recording studio on two floors and solved the hum problem by taking them off balanced and doing the wiring with a single ended plan. The major issue there was a close commercial radio transmission tower. I was surprised that it worked.
-Chris
The wiring doesn't have to be that way. Often times, a poorly laid out design is more at fault for noise than the ground wiring scheme used.
Most home balanced designs are done for marketing value, with small companies more driven by the designer's personal feelings on the subject. Often they do not possess the talent and funds that larger firms invest in their products. Sometimes they are right, and some times they aren't. The only important thing is that it performs well for you, and the same thing goes for someone else who stays single-ended. Like me!
Sobering thought. I once took a recording studio on two floors and solved the hum problem by taking them off balanced and doing the wiring with a single ended plan. The major issue there was a close commercial radio transmission tower. I was surprised that it worked.
-Chris
I think we are in violent agreement other than that fact that this is one of the few threads that gets down and dirty with discussions, so I feel casual consumers need to be dragged along kicking and screaming
As Bob has mentioned the Whitlock THAT designs are exceedingly good and affordable. I look on them as peace of mind but wont argue with anyone who thinks they are unecessary in domestic settings. Other things are far more important. Gain structure for one (having lived with a matching pre-power combo with way too much gain for a long time), but cakes need icing
.Secondary benefit is that going to twisted pair signal wiring cuts out 99% of the temptation to use expensive cables as plain old van Damme is as good as anything.