Should the whole system be balanced to make a difference?

A DAC I might buy has a balanced out. A line pre in the system where that DAC might go does have a balanced in. But, the outs from the pre to the power amp are RCA, and so are the ins on the amp.

Question to those who know for sure (i.e - have balanced components in their systems): will I hear any discernible difference by running balanced DAC out to pre ins only? The amp connection will have to stay RCA.

I have read somewhere that to make a real difference the whole chain needs to be balanced. Educated answers are appreciated.

 

It is certainly going to be more expensive, that is if you have a fetish for expensive interconnects. I have a bit of a mixture like you. For instance in my Stax valve amp for the headphones it hardly makes any difference. My power amp runs on XLR and I would say it is better, how much is hard to say because I don't have a simple way to make a proper comparison. I would go for the XLR but sorry I can't give you the answer you want.

 

I've bounced between fully balanced systems from input to output and single ended ones and have some takeaways, at least as they apply to my experiences.

Whether a component is balanced or not doesn’t determine its overall sound quality. I would not choose solely on that. Also, many products may have balanced connections, but the circuit may not be true fully balanced, so be careful. First and foremost, choose on sound quality alone. If it happened to be balanced, then that a bonus. Also, the true advantages of balances are a dedicated ground and shielding. So, if you have grounding issues in your home, balanced is a good choice. Or, if you have really long distance between components. Also, it adds 3db of gain. If you don’t need the gain, don’t have long cable runs or grounding issues, the advantages of balanced may not be fully realized.

For a circuit to be true balanced it requires much more circuitry compared to non-balanced. So, if it’s not implemented properly, all that extra circuitry could be a negative and certainly increase the cost. Many times, simple circuits yield the best sound quality.

 

Exactly the answer I was looking for.

No long runs - everything is very close within the same rack, no interconnect fetish, no grounding issues either. The DAC in question is a Schiit Modius - it has balanced outs and very positive reviews for sound. The Modius apparently is not a true balanced design, just has balanced outs.

On top of that, I already own a Modi 2 Uber, and pulled the trigger on a Modi Multibit today. - Sounds like my "need" for a Modius is nearly non-existent...

 

There is precisely one reason for a balanced system: immunity to external noise. It works because if noise is introduced to a cable, it enters both signal wires of a balanced circuit equally. The receiving end amplifies only the difference between the wires, not signals in common, a process called “common-mode rejection”. Once that’s done, there’s really no point. Fully balanced amplifiers and DACs literally use twice the circuitry to pull that off, with no real advantage. There’s been an argument floating around that certain types of distortion are also cancelled by fully balanced systems, and while that may be true, those distortion products are already below audibility, so it’s of questionable benefit.

Something to consider, though. The word “balanced” carries with it a payload of positivity, while the word “unbalanced” carries a possibly greater payload of negativity. Balanced...who wouldn’t want everything perfectly balanced? And who wants anything that is unbalanced? Well, the reality is, the unbalanced topology and interconnect system is very, very simple. In most home installations, it’s perfectly adequate because there’s no real noise issue in the first place. So, what does double the circuitry get you? More chances for noise, distortion, and cost of course doubles.

Balanced systems, when you need them, there is no substitute. When you really don’t, they run from no benefit, no harm, to no benefit and a little harm. It’s also important to note that the common-mode rejection aspect is not a single-figure parameter, it should be stated across a full range of noise frequencies, up to and beyond the highest noise component of concern. When stated that way, it’s becomes clear that not all balanced systems are created equally. There’s the degree of rejection and the frequency of rejection, which sets differing types of balancing circuits apart. Even transformers can and should also be characterized this way, as their CMRR can be huge over the audio band, but not so good up in the low RF spectrum where much of today’s noise resides. There have been cases where even though balancing was applied, the noise problems didn’t vanish.

This gets to the question “Do balanced systems sound better?” In some cases, absolutely. When they are called on to reduce noise that has made its way into the system, the absence of that noise “darkens the blacks”, if I may us that terminology. It’s quieter, and that’s very good. But, if there was no appreciable noise (that’s induced noise that entered the system from outside, not internal thermal noise), the no, they don’t sound better. The one aspect of a balanced system that always sounds better is it’s name, “balanced”.

 

Paging @Ralph Karsten

 

To that I would say potentially, especially on the XLR front - my take is if your spending more than the relatively inexpensive Mogami 2549 (or the Canares equivalent) on XLR for most home audio applications, you're probably throwing money away - especially if you have a soldering iron and build your own.

Don't want to turn this into a cable thread, but anyone spending 100's on home audio XLR should try and compare the Mogami or Canares. (even the Belden 1800).

 

FWIW, I’ve used fully balanced systems, fully single ended systems and a combination of the two and never heard a difference. It should be noted that I never had long runs to contend with.

 

For all intents and purposes, and on many home audio implementations of XLR this is true. But for designs where the entire connection chain is a "true balanced" instrumentation architecture, there are cases where balanced could see some advantage.

But typically you have to spend up to realize this benefit, just because a piece of gear has XLR ports doesn't mean it's a true balanced design.

Especially if you have long runs in addition, and / or you have issues where additional noise rejection might be a requirement.

 

Lets not elaborate a lot then because I don't want to either. What I would just like to say is that getting hold of something for testing is not always that easy. Then I have have other people telling saying exactly the same as you with completely different recommendations, the list is pretty long. Sometimes is just better to pay the price for the things you can actually get hold of as long as you can test without buying first.

 

I didn't mention "long runs" in Post #5, and I have several reasons for that.
1. There are no long runs at all in the typical, home system
2. The definition of "long run" changes with signal level, and driving impedance
3. There are certain conditions in which the cable length makes no difference to noise, because it's a ground loop problem. Balanced inputs can help with that.

Nearly every output circuit has a relatively low output impedance, or it would have trouble driving a piece of cable without HF rolloff. That driving, or source impedance, also serves as a "sink" for signals entering the wire via capacitive or electromagnetic coupling. Both of those coupling methods have an effective series impedance that's fairly high. When a noise signal gets coupled that way, the low source impedance of the output circuit creates an effective voltage divider, sinking a large portion of the coupled signal to ground. Capacitive coupling is the easy one, it's generally a very high Z source, and shielding is effective. Electromagnetic is a bit tougher because a driving source impedance may get higher with frequency...or it may not, so a little inconsistent there, but again, a good shield can be very effective.

Inductive coupling is a different animal. For starters, it goes right through all cable shielding like it's not there. The series impedance of inductive coupling is affected in several ways, proximity to the AC noise field source, cross-sectional area of the AC noise field, and effectively the "turns ratio", like a transformer, which can range from 1:1 with just two single wires, to a much higher ratio. Regardless, there is generally only lower frequency noise inductively coupled, and the source Z of most outputs is low at those frequencies.

So, even with unbalanced interconnects, we do have a substantial degree of noise immunity in most cases. But it all goes wrong if the ground potential of the two connected components isn't the same. When one device is, say, a few volts above earth-ground, and another is at earth-ground, two things happen. One, there's an AC current flow through the shield, which then provides a coupling path to the inner conductor. Two, since the signals are ground-referenced, having one device with an AC ground reference impresses that signal on the inner signal wire. This is what goes on with noisy PC power supplies, and makes connecting PCs to audio systems a challenge.

None of the above is specifically tied to cable length, though. And all would likely benefit from using a balanced interconnect. Fortunately, balanced is usually just fine, and if we don't plug in components on different AC power phases without the same ground reference, hum problems are mostly gone too.

As to fully balanced topology all the way through the system...well, you won't need it because you only need CMRR at the point where the interface problem (with noise) occurs. After that, unless components are spread out 50' apart from each other and on different power phases, it's no longer required but for that one point.

Yes, XLR doesn't always mean "balanced", only mostly. At very least it should mean "differential", but even that's not always true. There is a wiring convention for an unbalanced XLR, for example.

 

Benzion, you got some great advice there from jaddie.

What I learned a long time ago is if you run balanced keep the whole thing balanced. Same is true for un-balanced. Cross connecting the two types can led to less than optimum results. Also, after 18 and a half feet, a guitar cord acts like a radio antenna. That means keep the unbalanced stuff as short as possible.

 

Nice if you can, but no problem if you can't. Intermixing the two does not inherently degrade anything, if done correctly.

Guitar pickups and amps are a special case of unbalanced and high impedances. Just a bad combination. And can be RF sensitive if the designer didn't do his homework, and for some reason in that case, they usually don't.

That combination does not exist in HiFi, other than between a phono cart and preamp. We don't do 18' cables there. Outputs are low Z, inputs high Z. Just works.

 

Solid advice that I found to be true, make sure you use good connectors such as Neutrik or WBT.

 

I tried the full balanced once, no difference in sound in my case. No long signal cable runs. No high RFI environment. Back to RCA for me. Current rig doesn't support balanced.

 

I recently compared the RCA outs, buffered RCA outs, and XLR outs of an Arcam C30 preamp and could hear no differences.

 

This statement is false with qualifications. Many people in audio (I'm one of them) hear differences in interconnect cable. Its annoying- when that happens it really means that neither cable is correct (because sooner or later someone will have a cable to beat the one you think sounds better...). When you go balanced, and the equipment supports the balanced standard (which is the qualification) then you no longer hear differences in cables and the benefit is there regardless of length. Most of the time you have to do this using transformers, unless you own equipment that supports the standard (AES48). We were the first manufacturer anywhere to produce a balanced line preamp for home use and it didn't occur to us to not support the standard. As others got into the game (ARC, Jeff Roland, Sonic Frontiers, BAT and so on) we found ourselves nearly alone in this support. So as a result many people don't get all the benefit of balanced operation because their equipment, even though balanced, doesn't support the standard.

This needs clarification. First, there are long cables in home installations. I run 30 feet at home as this allows me to keep my monoblock amps right by the speakers.

Lots of output circuits have embarrassing high output impedances. A lot of passive volume controls suffer this issue. I've seen tube preamps with the volume control at the output, which to me is insane. And also limited to really short runs.

Balanced line is meant to eliminate ground loop issues among other things. That is one of the main issues with support of the balanced standard- and the biggest one that most companies ignore. The other is low output impedance- you really should be able to drive a 600 ohm load if you want to be successful driving a balanced line, even if the load impedance is higher. Your low output impedance swamps noise.

The bottom line is if all these things are not swept under the carpet, balanced line does not have to be more expensive (its often cheaper, since people pay stupid dollars for RCA cables) and when you hear it there will be no going back to single-ended. If you hear no difference, its likely because something is wrong in your comparison. One way this happens is some solid state amps have a separate input and circuitry to accommodate balanced operation- which adds noise and distortion- not an even playing field. Others put signal current in the shield (IOW the signal is referenced to ground rather than its opposite) and this negates the benefits of balanced operation as well. You really do need to compare the apples to the oranges if you know what I mean

 

From what I understand, a lot of balanced gear is not truly balanced and that is the problem. Manufacturers tend to cut corners.

 

I agree.

One short answer also might be if you have extra long cable runs in some spots (like some do, me included) or as you mentioned might have noise rejection issues (which most don't at home). Otherwise, if you're not running true balanced, have short runs there probably isn't a whole lot of justification. Oh, there is one other thing - on many XLR implementations, the XLR output can give a 6db boost over RCA, some take advantage of that when maybe pairing an ultra low MC, like 0.1 - 0.2mv to make up gain deficiency for their phono pre or line stage, and don't have configuration options on the pre's.

Can't speak to how many, but as a ratio most in the mid-fi range probably don't have full balanced designs. As one who owns both true balanced design and unbalanced designs that still have XLR, I tend to run the XLR where it makes sense and I can benefit from the true balanced design or due to cable length.

If one has long runs, or issues with noise rejection - that might be a reason to go XLR, even on a system that's not a true balanced architecture, or where there's a long run and RCA might have roll off due to cable length.

For instance I run multiple turntables in my main room (yes, many, don't ask how many). For one of my TT setups that has three TT's and a phono pre with 3 inputs, it's a 18+ foot run to the line stage. For that, I run XLR due to the distance. Another setup is 10 foot runs, but I employ mostly low capacitance RCA for that, but do use XLR at times too as the gear can use both simultaneously. The setup with the 18ft run has a phono preamp with inputs for 3 TT's, and offers both RCA and XLR out (Accuphase C-27). Although I don't think the C-27 is true balanced, given the distance there is a benefit for me choosing XLR first due to distance.

But, again, for those with short runs, no noise issues, and have equipment that might offer both XLR / RCA but not a true balanced design - I agree with your earlier comment that it can be a waste of time. The XLR is really just an option to RCA but shouldn't show any improvement because the XLR is really just an extension of the RCA connection internally, and not an independent true balanced design.

It is interesting to read in many manuals for gear - like for instance the Parasound A21 I've owned for a few years now - that they recommend using balanced connections over RCA for "best sound quality". They even do it for their P5 preamp - which is not a true balanced architecture. A lot of manuals state this in addition to Parasound - even when they're not a true balanced architecture. I am not even certain if the A21 is true balanced or not, can't recall. I will also say I ran the A21 balanced with both the P5 and then later with a higher end Accuphase line stage - and although I really couldn't tell the difference with the P5, I did notice a minor difference with the A21. I know it's subjective, but it just felt more dynamic, better separation. So in that case I stuck with balanced. (no, no double blind tests to verify). But others have stated the same with the A21, lots of testimonials.... I'm sure there are more cases.

For both my Parasound JC3+ and Accuphase C-27 phono preamps, I can have both RCA and XLR hooked up at the same time, and choose which one I want to hear at the line stage. That's nice to compare, but also at times when I have a cartridge mounted that might be a 0.1-0.2mv LOMC, or a 3mv MM - it's nice to choose XLR because on Parasound it gives a 6db boost on the output to compensate. To be honest I am not even certain the JC3+ is true balanced or not!

Same with my Accuphase line stage ---> Accuphase power amps (I own more than one) - I leave both inputs (XLR / RCA) hooked up and can quickly select / compare. Most of the time I leave the balanced connection active, as my newer power amp is true balanced design.

Anyway, got long winded as usual. But for someone with long runs, might be in situation with noise issues.... or someone with true balanced design or has gear with both and wants a db boost when needed, I can see XLR as a potential option.

Also, realize I am one that likes choices, likes to experiment, overkills it many times, so I may not be the best one to ask what's best, many times I say "get both!"

 

I read these posts after posting my diatribe above and responding to Steve many good points, we crossed over in some of the points made.... good read.

 

As far as I'm aware the modius IS true balanced, you only need one 4493 to do that if memory serves.
Also there's 2 output stages. Balanced has the benefit of noise rejection and ground noise killing.
Analog Stage: Based on LME49724 for balanced output, OPA1662 for SE output, both independent, with precision thin-film resistors and film capacitors, DC coupled.

I would use balanced, short cables can be had for less than super pricey RCA, they don't have to be super expensive be as good as you can get. I bought a pair of 3 footers for $42 shipped....

"As far as I know, Modius is the only DAC that has two separate output stages, one for balanced and one for SE. The balanced output stage is built specifically to the strengths of balanced—it uses the very good-sounding LME47924 balanced, differential op-amp, plus film caps and 0.5% precision thin-film resistors, and is DC-coupled. The AK4493 is inherently balanced output, so it’s balanced from start to finish."

This is from the horses mouth Jason Stoddard on another forum.

 

I saw that too - and as a potential Modius customer I was glad to read that....

 

It's a pretty fabulous dac IMO. It's been compared favorably to many more expensive DS dacs. It doesn't have the velvet akm sound which is nice. It's really resolving, although I favor multibit/r2r sound wise usually I liked it more over modi multibit. The SE is a litter warmer sounding, I use it with 2 SE amps and a speaker amp. One of my headphone amps has XLR and SE in and uses a phase splitter so SE is balanced as far as the amp is concerned. 4v from either input. makes comparing easy. Prefer the balanced output in that situation. It has better clarity and low end seems more clean background or "blackground". SE pairs nicely with colder analytical amps.