Single ended vs xlr balanced


I have switched backwards and forwards (going slightly mad in the process) testing (long run) interconnects.
I know some sound engineers, and they tell me I am probably persuading myself that balanced is better than single ended and THERE IS NO REASON WHY SINGLE ENDED CAN SOUND INFERIOR TO BALANCED. Sorry to use capitals but this seem to be a fair summary of the be all and end all of technical discussion. If I was to guess however, my mind would tend to follow the technical opinion not go against it, surely? In my mind the balanced is a deeper more airy sound, just better presence all round. The technical response is that I am not comparing like with like, as the balanced runs at higher voltages and subsequently higher volume (6db). I had a shock at this news and found out therefore by accident that my Bryston 28bsst2s amps have a switch upping the output from 23 to 29db to compensate. Also did I hear properly that Bryston kit is set up preferring balanced? My processor is a Bryston sp3 so maybe my preferred balanced system  is what I needed anyway. But it is odd that a reputable company like Bryston would have such a policy (if it has foundation) and not stress that on their literature. If my system can be adjusted for speaker levels then volume output is irrelevant - or is it if that higher voltage has some effect?
And don't get me started on aes/ebu vs spdif! The aes to me is noticeably superior for the same reasons as the rca vs xlr debate. Then hdmi vs spdif ... (Time for my medicine ........)
So my question is - forgetting technicalities which can get more and more complicated by the minute  - do other peoples' ears agree with mine?
tatyana69

Showing 8 responses by almarg

There is no absolute rule as to which will sound better.
It is entirely
(a) system synergy as-a-whole dependent in the first part, and
(b) the equipment and IC (both) build quality dependent.
I agree completely. 

In general, I would expect that among the most significant variables would be the specific designs of the interface circuits in the components that are being connected (i.e., the stage in each component that drives or receives the signal to or from the cable), and also the internal grounding configuration of each of the components (especially the relation between circuit ground and chassis ground), and also in the case of balanced connections whether the components connect the ground conductor in the cable (XLR pin 1) to chassis or to circuit ground.  Different designs vary in all of those respects.

Regards,
-- Al

Regarding Ralph’s contention that bdp24 referred to, for a full description of what is necessary for that benefit to be realized, as well as for what I consider to be a compelling proof of his contention, see Ralph’s (Atmasphere’s) post near the beginning of this thread. Also see the follow-up questions I presented to him later in the thread, and his response.

Regards,
-- Al

Can a well designed rca cable compensate for ground loop?
It **might** make a difference in some cases, due to differences in the resistance of the shield or other return conductor, and depending on whether the shield **is** the return conductor, or if the shield is connected only at one end and a separate return conductor is provided within it.

But in general I would expect susceptibility to ground loop issues to be more dependent on the internal grounding configurations of the components that are being connected. And in many cases also on how AC power is distributed to the components (e.g., from a single outlet vs. multiple dedicated lines, etc.), and through what kind of house wiring it is distributed (e.g., within metal conduit, or with Romex, etc.; see this paper).
... if balanced is the favoured "vehicle" for long runs ..... why are probably 98% of all xlr cable adverts for sale a seemingly average of 1 metre long? I wanted 8-10 metres but are they ubiquitous?
Interconnect lengths of 1 meter or thereabouts are of course more often needed in home audio applications than long lengths such as 8 to 10 meters. And since most cable parameters and most cable effects are proportional to length, using a greater length than necessary will generally worsen whatever sonic effects the cable may introduce, as well as increasing the cost of the cable.

I would say that the short answer to most of your questions is simply "it depends." And unfortunately many of the dependencies don’t have much if any predictability.

BTW, you may find this paper to be of interest, although it deals mainly with pro audio applications. I’ll quote two sentences from it:
So why doesn’t everyone do it [the right] way? Because life is messy, some things are hard to change, and there will always be equipment in use that was made before proper grounding practices were in effect.
Regards,
-- Al

One more point about using long interconnect cables, such as the 8 to 10 meter length you referred to: If the output impedance of the component driving the cable is on the high side at mid and high frequencies, say something like 500 ohms or more, you would want to make a point of using cables having low capacitance per unit length (say around 60 pf/meter or less). Otherwise the interaction of the output impedance of the component with the capacitance of the total length of the cable may result in a perceivable rolloff of the upper treble, undesirable phase shifts of high frequencies, and the perception of sluggish transients and compromised dynamics, to at least a slight degree if not more.

Component output impedances are usually specifed at a mid frequency such as 1 kHz, but in most cases will not be greatly different than that value at high frequencies.

All of this applies regardless of whether the cables are balanced or unbalanced.

Unfortunately, though, there are many manufacturers who don’t specify the capacitance of their cables. If (as appears to be the case) a component in your system drives an analog line-level signal into a particularly long interconnect cable, such as 8 to 10 meters, and if the component has a highish output impedance as I described above, I would contact the manufacturer if necessary to determine the cable’s capacitance per unit length. If that information cannot be determined from either the manufacturer’s website or by contacting the manufacturer directly, I would look elsewhere in selecting the cable.

Regards,
-- Al
Excellent input from Mward. There are certainly many preamps that would not be able to drive 12K with good results, especially many tube preamps, but would be very happy driving 100K (which BTW probably means that 50K is applied to each of the two signals in the balanced signal pair).

In addition, it can be inferred from the fact that the balanced input impedance of the amp is much different than twice its unbalanced input impedance that there are considerable differences in the designs of the balanced and unbalanced input stages. (In contrast, for example, to many amp designs in which the same input stage is used for both balanced and unbalanced inputs, with one of the two input signal lines simply being grounded via a switch or an external jumper when the stage is operated with an unbalanced input signal). So even if the preamp could drive 12K without issue, it is possible that the balanced input stage of the amp simply sounds better than the unbalanced input stage. Even though the amp is not "fully balanced."

Regards,
-- Al

Good comments by Erik and Kijanki, as usual.

A clarification to comments a couple of people have made, to the effect that a fully balanced amp requires essentially two separate amps on one chassis, one amp for each polarity of the balanced signal pair. The implication being that twice as many parts are necessary than would be required for a comparable single-ended signal path.

That is one approach to implementation of a fully balanced amp. But another approach, which I suspect is used more frequently, is to use a differential stage, having differential inputs and differential outputs, for each of the active stages in the signal path.

See the section on "differential amplifiers" in this paper at Ralph’s Atma-Sphere site for a simplified schematic representation of a tube-based differential stage, and also for a summary of the advantages and disadvantages of that approach. Note this statement in the disadvantages section:
Differential amplification takes more parts to execute. For a given number of stages of gain, differential amplifiers have about 50% more parts.
Regards,
-- Al

Dave_b, the belief you and Stringreen have expressed (which Stringreen has also expressed in a number of previous threads), while being shared by many audiophiles, and while being true in many cases, is simply incorrect as a general rule.

Depending on the specific designs that are involved, and also in some cases on how AC power is distributed to the system (as I mentioned in one of my posts on 8-28-2016) a balanced interface may provide performance that is better than, worse than, or similar to the performance of a single-ended interface, even if the components involved do not have balanced internal signal paths. Several reasons for that have been cited in the course of this thread. For starters, see the post above by Mward, dated 8-29-2016, and the responses thereto.

Regards,
-- Al