xlr vs rca

Herman makes some excellent points. There are many folks who dabble in the "high end" who are convinced that single-ended circuits are inherently advantageous (in fact), because fewer parts are needed to construct them vs balanced circuits. I disagree with the idea that the parts count is relevant, for complex reasons, but I do agree that one should keep an open mind on the balanced vs single-ended debate, only one of many that rage among us crazies. I would only say that if the circuit is inherently single-ended (and you like it), go with that and don't try to create a pseudo-balanced interface.

As long as you and yours are beating the "it measures better so it must sound better" drum and I and mine are only concerned about what really matters, which is how it actually does sound, then there is really nothing to discuss.

No worries I get it. We have beaten this horse dead. I can respect that what concerns you is what really matters. I'll be the first to admit that accuracy in sound reproduction is what only a very small minority of hobbyists are after.

{edit} I was writing this below as you posted so I had not read your last response. I agree, we are approaching this from 2 different angles {edit}

I know I said I was bowing out but I wanted to address your comment about convincing me. Of what? That balanced circuits have some inherent advantages in the area of noise performance or that because of these advantages balanced must sound better?

The former is a given, the latter is not. It is easy to fall into the same trap that Julian Hirsch did when he proclaimed the superiority of transistor amps that had extremely low measured distortion. They measured well but sounded like crap. The same trap that some fell into when CDs were declared to have perfect sound forever. It is easy to develop tunnel vision and focus on this one tiny area (noise performance) but the picture is so much bigger. If all else was equal would it be better to have the inherent advantages that balanced offers? Of course, but since it is not all equal the point is moot.

The proof is not what is in the textbooks but in what we hear. The type of system I listen to brings me pleasure that others I have heard do not. I've heard systems that have a low level hum that is easily heard at the listening position but when the music was turned up they sounded glorious. Sure, it would be better to get rid of the hum but if you sit there tapping your toe with a smile on your face how much does it really matter?

As long as you and yours are beating the "it measures better so it must sound better" drum and I and mine are only concerned about what really matters, which is how it actually does sound, then there is really nothing to discuss.

Traditional measurements illustrate how machines hear, not how humans hear.

Yes this is a good point. We may be discussing from incompatible and quite different perspectives/objectives. My perspecitive is audio reproduction with high accuracy. Most people may only be concerned themselves with what sounds subjectively nicer to them.

For example, it has been demonstrated that audio compression with the distortion and the reduced dynamics that it brings can be pleasing to the ear - making music sound more punchy and fat. So from a "what sounds best to me" perspective then any viewpoint can be valid. (Analog tape machines and certain circuit designs are highly prized for the sound they impose on the recorded music)

This divergence in goals is analogous to the difference between trying to follow a recipe exactly (reproduce what the cookbook calls for) or adding a bit of extra or different spices (not following the recipe rigorously). I want to hear what is on the recording. I appreciate that others do not necessarily seek that at all and it make for a discussion confusing.

Herman brings up a good point. Traditional measurements illustrate how machines hear, not how humans hear. I too was convinced by all the reading I did that a balanced set-up would have advantages over single ended, and it may have, but I could not really hear a difference, especially with the critical factor I was benchmarking against which was reduced noise floor.

If your amp and preamp is true balanced then go balanced as much as possible. If it is not then I think it is a waste of time and money.

In order to affect a ground loop whatever is doing it has to change the potential of the ground. How does a power supply do that?

Through parasitic leakage through capacitors and through transformers.

FWIW: There are plenty of examples where XLR has its share of noise and hum problems (poor design choices) and Whitlock covers these issues as well. So RCA can be better than XLR and vice versa depending on the setup and specific equipment. It is only properly implemented XLR that should have an edge over RCA.

I don't care what the textbooks say. I don't care how it measures. All I care about is how it sounds. Unfortunately there is no way to make that comparison via the internet.

Ok. Then it will be impossible to convince you. Perhaps the links will be useful to the original poster.

Maybe I missed it but I don't see where it says anything about power supply leakage. What do you mean by that?

The insidious kind is when a power supply leakage affects ground on a device which adds modulated noise on your cabling as a function of power demands. The way to reduce this is to go balanced (so that the ground loop is induced equally in both positive and negative signals).

In order to affect a ground loop whatever is doing it has to change the potential of the ground. How does a power supply do that?

I think I'll bow out of this discussion. I'm told that there is no way the signals aren't separate when clearly they interact in the diff amps and then some vague references to leakage. It is really all pointless. Here's the bottom line. Those of you in the balanced camp can go on and on about all of the reasons why balanced "should" sound better than SE. You are correct; it should. There is a laundry list of reasons why CDs should sound better than vinyl. The simple truth is it does not. My ears and those of many others say that the best we've ever heard consists of SET amps coupled to high efficiency horns. I don't care what the textbooks say. I don't care how it measures. All I care about is how it sounds. Unfortunately there is no way to make that comparison via the internet.

Try the Whitlock paper directed at students - Page 9 and 10. Although as the saying goes - you can lead a horse to water but...

I understand the use of the CCS but I don't see where this makes the circuit immune to differences in the tubes. It's been awhile since I've studied them but as I recall any difference in the components will be reflected in the output.

Herman, I have never seen a tube preamp or amp that achieves balanced operation using completely discrete amplification for the plus and minus halves of the signal. I wondered if you had, and I guess not. BAT most certainly does balanced operation the same way Atma-sphere does. I think to do balanced operation without the differential topology would be a disaster because of the difficulty matching Rp, mu, and Gm for two completely disconnected sets of tubes. At least in differential topology the current can be held constant for the two halves of the circuit by using a CCS in the cathode of the dual-diff amplifier, and the current on one side will go down by the exact complement of the increase in current on the upgoing side. The only place I have seen balanced without differential is in the context of using an IC or a 2-section bipolar transistor, which can have incredibly closely matched sections.

Yes, I know an MC cartridge is an inherently balanced signal generator if hooked up to a balanced input. I guess I was thinking of a single-ended signal from a cdp or other source.

If you care to be more specific about where he states this I would be happy to read it, but the link you gave has 10 papers on it and I don't care to read all of them trying to find it.

Perhaps I was not clear in the last post.....a BIG BEEFY power supply means less fluctuations due too the varying demand from dynamic music and therefore less ground loops from leakage to ground from the PS => less noise induced into the low level analog signals between components.

I’ve found it’s very often advantageous to float the ground at one end of the signal path.

Agreed that is often the first thing to try out....it stops the microamp ground loop currents flowing between the various chassis as well as still grounding the shield to give you better RF protection.

Balanced circuits are better at rejecting any noise from the power supply but your description of it doesn't make any sense. It isn't a ground loop problem.

You need to read up on what Whitlock says about the reasons why XLR balanced is so important for ground loops. Think about how a power supply fluctuates under heavy demand and how this may affect microampere ground loops between gear (the BIGGEST problem in audio as RF noise pick up on analog audio is much less common). Then realize that ONLY ONE wire in in an RCA circuit will carry this current flow on your signal wires - IT CANNOT CANCEL (as it does in well built XLR gear and cabling).

=> this is why a BIG BEEFY power supply is NOT simply about more power...often it can be more important for how much cleaner the backgroud or canvas upon which the music plays can become...

When using truly balanced interconnects, i.e., differential signals on pins 2 and 3 and ground on pin 1, I’ve found it’s very often advantageous to float the ground at one end of the signal path. This can be easily done by disconnecting the ground on pin 1 at one end, and only one end, of the cable.

The result is that the cable is still shielded but there is no electrical connection between the chassis grounds of the equipment. This not only increases hum isolation but also prevents high frequency digital clocks and switching power supply noise from traveling between chassis.

I’ve used this floating ground scheme on large analog post production installations (with lots of patch bays and signal routers) as well as on my AES/EBU balanced digital audio cables at home.

On a separate note, while balanced interconnects have way better noise immunity than unbalanced it’s not free: they require twice as much circuitry in the signal path and both sides of a balanced circuit must be perfectly matched.

the ground loop is induced equally in both positive and negative signals

Balanced circuits are better at rejecting any noise from the power supply but your description of it doesn't make any sense. It isn't a ground loop problem.

No hum, no ground loops and I am biamping.

Ground loops do not always show up as hum with volume turned way up with no music playing (of course really bad ground loops do). The insidious kind is when a power supply leakage affects ground on a device which adds modulated noise on your cabling as a function of power demands. The way to reduce this is to go balanced (so that the ground loop is induced equally in both positive and negative signals). You really can't detect this kind of insidious effect other than to observe the improved clarity of going to a properly balanced setup where grounds from different components are much less likely to get to your signal. The effect can of course be small when comparing good equipment that is well matched... nevertheless, several DB in better noise floor can usually be expected with XLR.

Thank you all very much. I have learned a great deal from studying all of the input I have been given.

Lewn,

Also, in any "balanced" design, an early stage, have to split the signal into positive and negative halves?

A phono cartridge can output a differential signal and most if not all modern DAC chips have differential outputs so no split is needed. It can start out that way.

I am just not aware of any other way to do it with tubes

you simply build 2 amplifiers in one chassis like stated earlier and one amplifies the + signal and one amplifies the minus. I don't know how all of the balanced manufacturers approach the problem but you can certainly do it that way. I did look through BAT's website and they make no mention of differential circuits. Perhaps they use them but don't talk about it?

Shadorne,

I'm afraid I don't follow. The balanced builders are just as likely to use gold plated connectors as the SE crowd. As for ground loops, take a look at my system. The phono is balanced into the pro box but it is converted to SE by transformers at the output of that box and is SE from then on. I run 107dB efficient horns with SET tube amps and even with the phono turned up I have to my ear up near the speakers to hear any kind of noise. No hum, no ground loops and I am biamping.

Herman, I guess I should have phrased my question differently. I am aware of the Atma-sphere dual differential topology and how it works; I am just not aware of any other way to do it with tubes. Perhaps I should have asked how can you do it differently with tubes, and who (what company) uses other than a differential topology to achieve true balanced operation with tubes? Also, in any "balanced" design, an early stage, have to split the signal into positive and negative halves? It seems to me that at least the phase-splitting step is easiest to achieve with a differential topology. Downstream from the phase splitting, I get that additional differential topologies may not be necessary. Sorry that this is somewhat OT, but it's interesting, at least to me.

With the large amount of excellent SE gear it is hard to argue that balanced is inherently better in a home environment even though those that make it and use it would disagree.

SE gear is cheaper so it must be better - the savings can be used towards gold plated RCA connectors and ground loops are rare in the home... ;-) LOL

Lew, thanks for the kind comments. Herman is more knowledgeable than I am concerning the type of balanced architecture he referred to, and I will defer to his comments. Perhaps my reference to fully balanced architecture should simply have said that it maintains a balanced pair of signals throughout its internal signal path.

On a separate note, one thing I neglected to emphasize in my post is that besides providing rejection of noise (in the sense of high frequency hiss and buzz), a balanced interface is much less susceptible to low frequency hum problems. Shadorne alluded to that, and the paper he linked to explains why.

Basically, since the chassis of single-ended equipment is connected to ac safety ground, to signal ground, and to the shields of single-ended interconnects connected between components, any offset in ac safety ground potential between the two components will cause an extraneous 60Hz current to flow through the shield, in common with signal return current. As Shadorne points out, the magnitude of the resulting hum will be dependent on the resistance of the path through the shield, and therefore on the particular interconnect.

If the interface between the two components is balanced, that becomes a non-problem, since the conductors for signal current and signal return current are separate from the shield of the cable. That is true whether or not the components are "fully balanced."

Regards,
-- Al

Yes, very true. A lot of people post that don't have any idea what they are talking about.

The circuit Almarg described consists of 2 amplifiers or series of amplifiers one of which amplifies the inverted signal and the other the non-inverted. They can indeed be completely separate until applied to the speaker.

A differential circuit of the type used by Atmasphere consists of 2 tubes that share a common connection; the cathodes of the tubes. This is sometimes referred to as a long tail pair. It is difficult to explain circuit theory in a forum such as this but rest assured the 2 signals do interact in a differential amplifier. Applying a signal to either input will cause current to flow in both tubes. The circuit amplifies the difference between the signals, hence the name. It does not amplify any signal which is the same i.e. it rejects any common mode signals. If the 2 halves were separate and did not interact the circuit would not work as intended.

You can google it for more info.

It's amazing how audio myths and misinformation persist in the minds of audiophiles. I vote with Almarg. He is about the only person who got it all correct. Raul was not wrong, either. His response was just not complete. Others get part credit. Herman, please tell me how the balanced differential circuits of Atma-sphere products are not "completely separate". Atma-sphere should get all credit for pioneering the use of balanced circuits at a time (early 80s) when such topologies were practically unheard of in commercial audio gear.

Hi Herman,
I like your précis, it's about as much to the point as it will get, but let me add one small item I been missing going through the thread.

The 'common mode rejection' also has the effect of favouring odd-order harmonics and cancelling even-order.

This seems to favour a more 'dynamic' presentation for rock and pop compared to SE which in turn would sound just a tad more 'natural' or ‘right’ with classical and acoustic instruments.

I am talking nuances here, but it can be heard, and which one is preferred is a matter of taste as always.

Greetings,
Axel

Fully balanced design, which is normally found on only some very high end equipment (such as Ayre, BAT, and Atmasphere, to name a few examples) provides a separate channel from input to output for both the "hot" (pin 2) signal, and the "cold" (pin 3) signal.

Don't know about the others but Atmasphere uses differential circuits, which even though balanced are not completely separate.

Modifying an SE cable by putting XLR connectors on it or using RCA to XLR adapters does not turn it into a balanced cable. It is still SE.

I think the salient points are:

1. if it is designed to operate as balanced use it that way
2. XLR connectors are better than RCAs but if you are using SE equipment you are stuck with RCAs unless you DIY.
3. In an electrically noisy studio with a myriad of equipment balanced has distinct advantages and that is why pro gear is balanced, but your house is not a studio.
4. With the large amount of excellent SE gear it is hard to argue that balanced is inherently better in a home environment even though those that make it and use it would disagree. You will hear the argument that I tried my balanced amp both ways and balanced sounded better so balanced is better. If it was designed from the ground up to be balanced it should sound better when operated that way, but it doesn't prove anything about balanced vs. SE.

If you are connecting a balanced (xlr) output to an unbalanced (rca) input, be careful about using an adapter. Most commercially available xlr-to-rca adapters and adapter cables short pin 3 (the "cold" signal, which on a balanced output is an actual signal, despite its name) to pin 1 (ground). Nearly all pro equipment, and a lot of consumer equipment, can tolerate that, but on some consumer equipment (depending on the design of the output driver stage) the result may be compromised sonics, or degraded long-term reliability, or immediate damage. Check with the manufacturer first.

Using an adapter or adapter cable to connect an unbalanced (rca) output to a balanced (xlr) input is no problem, though. Although of course doing that does not provide most of the noise rejection that is provided by a balanced interface (if done right, it can provide some noise rejection -- see Figure 2.1 of the reference Shadorne linked to).

As Sid indicated, using an rca cable with rca-to-xlr adapters at both ends makes no sense, although it will function (apart from the possible effects that I mentioned of shorting pin 3 to ground). The noise rejection advantages of a balanced interface occur because noise picked up in a balanced cable is presumably picked up equally on both conductors, resulting in cancellation in the differential receiver stage of the equipment the signals are going into, which responds to the difference between the voltages on the two input lines. A balanced cable makes that possible by using two identical conductors, twisted together, within an overall shield. An unbalanced rca-type cable has only one inner conductor, with the shield serving as the other conductor, and so noise will obviously not be picked up equally between the two conductors, due to the physical asymmetry.

As has been alluded to, it is important to distinguish between "fully balanced" equipment (i.e., balanced internally as well as at the interfaces), and equipment which is only balanced at its interface connectors and the corresponding input and output buffer stages.

Fully balanced design, which is normally found on only some very high end equipment (such as Ayre, BAT, and Atmasphere, to name a few examples) provides a separate channel from input to output for both the "hot" (pin 2) signal, and the "cold" (pin 3) signal. (Hot may be pin 3, and cold may be pin 2, on some equipment not made in the USA). That provides cancellation of some forms of internally generated distortion, as well as cancellation of noise picked up in the cabling. Balanced cabling with xlr connectors should obviously be used with that type of equipment.

Designs which are not fully balanced, but have balanced (xlr) interfaces, and which may offer rca interfaces as well, may sound either better or worse when the balanced interface is used. Better if noise may be a problem in the particular setup; worse if that is outweighed by the sonic effects of the additional stage that the equipment introduces into the signal path to perform single-ended to balanced or balanced to single-ended conversion.

The reference linked to by Shadorne is excellent, and well worth taking the time to study.

Regards,
-- Al

I understand it is better to used balance interconnects if possible. Is this always the case?

With good compatible XLR equipment YES - this is always the case. However connecting between low quality consumer RCA to pro-quality XLR is not always straightforward - check this out Ground loops are most often the culprit - this is probably one of the main reasons consumers report so much differences with interconnects - it is just not good to have stray ground voltages flowing through signal wires, as is done with RCA.

Listen to Kr4. If the gear is all truly balanced... using xlr cables is best.

if there is a lack somewhere, using an adapter is not a terrible thing. I went all BAT power train using xlr. My source was not balanced so adapters were in use. it sounded great IMO.

Sold the preamp, kept the amp, bought a SE preamp so again, adapters were needed... best sounding rig I've owned at that time.

I used Cardas adapters, BAT adapters, and some pro adapters, in fact I use a pair now on my DAC so I can run SE/RCA cables.... the noticed diffs from one set to another was negligible. In fact I doubt in a blind test one could actually pick out which xlr to rca adapter was being used.

That's the truth as I know it... see for yourself. $50 - $100 for upscale adapters. $15 for pro. Maybe you will hear adiff. I did. It just didn't justify the extra $35 - $85. it was that slight a diff.

Enjoy.

Jamiek,

If you use XLR to RCA adaptors on each end, the XLR cable in the middle will be performing just like a RCA cable. You don't get any noise cancellation property of a XLR cable. There is absolutely no benefit to do that except higher cost.

Rauliruegas's statements confuse the issue. A completely balanced design component has some advantages over single-path designs but it is not necessary to accrue the advantages of balanced connections. Balanced connections, with their increased immunity to induced noise, require only a balanced output connected properly to a balanced input.

Kal

Dear Jamiek: What make a balanced unit is not the interconect cables but that the electronic audio link ( item ) was/is designed in balanced way.

There are units that are fully balanced ( input to output ) where you need a balanced cable at the input and output ( not because it comes with XLR ), this means that the cable internal configuration and connection to its XLR termination connectors are true balanced.
There are other units that are a only output balanced design with unbalanced( RCA ) inputs.

So it is not the cable termination the one that tell you if it is balanced or not but the audio electronic item internal design.

One advantage on balanced designs is that noise/distortions canceled between the balanced design.

So you don't have any advantage if you use XLR connectors in a unbalanced audio elewctronic item because this item accept only unbalanced audio signals.

Balanced is not better because it is "balanced" but beecause so many rca's are really bad conncetion wise. Balanced sounds worse to me in many cases, but I prefer the sound of XLR's to RCA's. A contradiction??? No, I use pin 1 gnd, and pin 2 +, and avoid pin 3 and use the XLR's. Balanced cable is good if it is shielded, and you have a long run, say more than 10M and a noise issue. No noise issue, no need for balanced cable. I like the no 3 pin setup best. Jallen

It is not always the case. It depends on the interconnected equipment and on the electromagnetic noise in the environment.

Kal

I'd stay away from the adapters..the cardas i used on my cdp put huge noise into the system...balanced is not always better anyway but some components are optimized for best performance with xlr's, like BAT gear. my raysonic cdp has rca and xlr but I use the rca since my amp is only rca.
Larry