Learn the differences in balanced and unbalanced headphone...
As the intro to this particular post, I’d like to requote a portion of September’s hands-on review of Pioneer’s XDP-02U portable audio player(the bolded emphasis is mine):
Both unbalanced (complete with three different gain setting options for different headphones’ impedances) and balanced analog options are available, along with analog line out, and the available Bluetooth profiles include (via firmware update) higher-quality aptX.
I realized in reviewing the artwork included in my earlier coverage that none of the “stock” photos I’d sourced from Pioneer gave a clear view of the topside analog audio output connections, so I’ve snapped my own photo (in it, you’ll also note that I’ve now acquired, thanks to an overseas acquaintance, the unit’s Japan-only black leather case accessory):
Is this arrangement intended to allow two people to listen to the XDP-02U at the same time? Good guess, but no. Further clues come from the fact that the two jacks are labeled differently, as well as having slightly different dimensions. Enough suspense, here’s the answer. The one nearer to the center (and to the power button on the other side) and explicitly labeled with a headphones symbol is a standard 3.5 mm diameter (a.k.a. 1/8”, although the latter is only a dimensional approximation) TRS (tip-ring-sleeve) three-conductor female connector, intended for use with a set of “stereo” headphones.
The one next to it is the impetus for this particular post. It’s labeled “bal” (for balanced); in contrast, its next-door conventional counterpart is commonly referred to as unbalanced, along with single-ended, for reasons which Moon Audio concisely contrasts within an excellent four-part writeup on the topic (I’ll further elaborate on this difference later):
For the purposes of our discussion, from here on out we will refer to the term “unbalanced” as “single-ended.” This is because, in a balanced connection, you have a positive and negative (polarity) and a ground line. In single-ended connections you only have a single (polarity) and a ground line; hence the term “single-ended.”
The Pioneer XDP-02U’s balanced connector is 2.5 mm in diameter and has four contacts, in a TRRS (tip-ring-ring-sleeve) arrangement. It’s one of a number of common implementations of the balanced connector concepts; others you’ll frequently encounter include:
One clarification upfront: the TRRS connector you’re probably most familiar with is 3.5 mm in diameter and tacks a microphone-intended contact onto the standard left channel (tip), right-channel (ring), common ground (sleeve) stereo headphone arrangement. That’s not what we’re talking about here. In this particular case, the four contacts handle the positive and negative (i.e. “normal” and mirror-image inverted) audio drive signals for each (left and right) channel coming out of the headphone amplifier circuitry, at least in the most comprehensive implementation of the concept.
To explain, let’s follow the audio signal chain backwards, beginning with the transducer(s) in proximity to each ear (most headphones employ a single full-range speaker per side, but more advanced setups like my Shure E5c in-ear monitors (IEMs) leverage multiple speakers in combination with a crossover). As one of the online resources (from Benchmark Media Systems) I perused in researching this piece points out, “Headphone transducers are balanced devices. They have two wires. The electrons that flow into one wire must flow out of the other. The current is always balanced.”
The cabling feeding each speaker (or cluster) is where the implementation differences between various headphones (and the amps feeding them) originate (but don’t necessarily end). Quoting from another useful online resource I found, published by Headphones.com:
A normal headphone plug has three connections on it: the tip is left, the ring is right, and the sleeve is ground. The tip connects to a wire that goes to the positive (+) lead of the left headphone driver coil; the ring connects to a wire that goes to the positive lead of the right driver. The sleeve connects to a wire that goes to both negative (-) terminals of the driver coil. This wire usually has a solder joint in the “Y” or in the earpiece where ground wire from the plug splits into separate wires that are connected to the negative terminals of the driver coils.
In normal headphone amps, the most important thing to note is that as the left and right channels of the amplifier drive the left and right driver coils, the return current from the drivers gets joined together and travels some distance before returning to the amp’s audio ground.
This common pathway has some significant electrical resistance from the wiring, solder joints, contact resistance at the plug/jack etc., which causes a common signal to appear at the negative terminal of both driver coils. This common signal (a low-level summation of the left and right channel) will generate low-level crosstalk and harmonic distortion in the headphone presentation. Just how “low level” depends on the quality of the entire headphone and amplifier system audio chain, but it will always be there as a muddying factor – i.e., the background noise floor.
Conceptually, you can minimize this particular SNR- and THD-affecting phenomenon by limiting the “some distance” mentioned earlier, over which the drivers’ return current is combined i.e., by having this merging occur solely within the amplifier itself, keeping the returns distinct through the entirety of the headphones and their associated cabling. This approach, known as balanced transmission, is most commonly found with long cabling runs. Again quoting from the Headphones.com coverage:
In many professional audio applications, signals are routed from place to place via balanced cables in an effort to reduce common mode interference from radio frequency (RF) noise sources like fluorescent lights and motors. Because the voltage developed on the transmission line from RF interference is identical on both normal and inverted signal conductors, and because the balancing transformers only allow current flow when the signals are opposing, the common interference signal is canceled out.
However, the Headphones.com writeup goes on to note, “This problem is not typically an issue even with single-ended headphones because the low output impedance of the power amp prevents a significant RF interference noise voltage from ever developing.”
Hmm. So am I saying that balanced headphone setups are pretty much nothing more than “snake oil”? Not at all, because I haven’t finished the story. Significantly more notable benefits are achievable, at least theoretically, when the entire signal path for each channel is balanced, a concept often referred to as a “fully balanced” design. In the spirit of “a picture paints a thousand words,” here’s what the signal path for an up-and-coming portable DAC and headphone amp, the Qudelix-5K, looks like when run in unbalanced (i.e., single-ended) mode and in balanced mode:
In unbalanced mode, a single ESS Technology ES9218P SABRE DAC/amplifier SoC (out of two total in the design) is employed, with a single amplifier per channel and the returns of both amplifiers’ channels connected to a common ground. In balanced mode, on the other hand, the second ES9218P is pressed into service, with the two amplifiers in each SoC devoted to the positive and negative audio drive signals of each channel. This is akin to the “Twin ESS SABRE® ES9018C2M DACs” and “Twin ESS SABRE ES9601K high-output amplifiers” mentioned in in the Pioneer XDP-02U literature. And LG similarly refer to the feature as a “Quad DAC” in conjunction with the company’s smartphones.
Perhaps obviously, any common-ground crosstalk between the two channels that might be present in an unbalanced configuration, adversely affecting the noise floor in the process, is theoretically eliminated in a fully-balanced design. More notably, again quoting from the Headphones.com coverage:
Since there are two power amps driving each speaker coil, each amp effectively drives half the coil, with a virtual ground or zero voltage point halfway into the coil. Since each amp is only driving half the load, a significant improvement in control can be achieved, and because the voltages are in opposition, an effective doubling of slew rate (volts per second the amp can swing) is realized compared to the normal slew rate of either amp by itself.
The resulting benefits are implementation-dependent, of course, and come at the tradeoff of increased circuit complexity and cost. As a result, note, for example, that the Massdrop Objective 2 Headphone Amp: Desktop Edition that I mentioned back in late July 2019 doesn’t even offer a balanced headphones connection option:
Then again, the Massdrop x Grace Design Standard DAC that it’s tethered to also only supports unbalanced (RCA) analog audio interconnect, so what would be the point?
A doubling of the amplifier’s per-channel power output is, practically speaking, most valuable with high impedance headphones, but portable players typically mate with low impedance transducers, precisely to minimize the amount of battery drain needed to achieve a particular listening volume level. And your ability to audibly discern lower THD, improved SNR, and other touted advantages of a fully balanced approach depends on numerous factors, such as:
Nonetheless, probably unsurprisingly to many longtime readers, I decided to take the balanced headphones plunge anyway. The (low-impedance) Massdrop x Sennheiser HD 58X Jubilee and (high-impedance) Massdrop x Sennheiser HD 6XX headphones mentioned in my mid-July 2019 coveragecome standard with unbalanced cabling, but a user-replaceable balanced replacement cable for them is available from Drop and elsewhere. As far as IEMs go, all of my existing headphone sets from Shure (the earlier-mentioned E5c), Etymotic Research (my beloved ER-4S set), and others were unbalanced-only. So (of course) I dropped some more dollars on another set, Etymotic’s successor ER4SR (on sale at the time for $159.99 versus a $299.95 MSRP), which does support a balanced configuration option, in conjunction with a separate cable-replacement supplement:
So what’s the verdict? I certainly can’t tell a balanced-vs-unbalanced difference when I’m listening to tunes while vacuuming the house, for example. And more generally, although my increasingly-geriatric ears may be to blame, I admittedly can’t tangibly discern any consistent variation in THD, SNR, or other similar factors, even when I’m listening to the most pristine content variant available from Tidal HiFi. That all said, the balanced setup is definitely louder than its unbalanced counterpart, an unsurprising outcome when you understand the circuit implementation differences between them.
One final note in closing. Earlier I mentioned that I had a bunch of sets of unbalanced headphones. Please (as Moon Audio also eloquently states) do NOT try to use an adapter to connect them to a headphone amplifier’s balanced outputs! All you’ll be doing if you unwisely try this experiment is shorting together the two channels’ returns, sooner-or-later frying one or both amplifiers in the process. That said, the converse (harnessing an adapter to connect balanced headphones to an unbalanced audio output) works fine. Now that I’ve cable-converted my Sennheiser and Shure headphones to balanced configurations, for example, I use adapters from DD (shown below on the left) and Geekria (shown below on the right) to conveniently also listen to music coming from systems with standard 3.5 mm or (in conjunction with another adapter) ¼” TRS analog audio outputs.
Nearly 2,000 words; yikes! Time to wrap up. More on this topic in posts to come, but for now, I look forward to your thoughts in the comments.