• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Differential, Single-Ended, Balanced, Unbalanced, and all that jazz...

DonH56

Master Contributor
Technical Expert
Forum Donor
Joined
Mar 15, 2016
Messages
8,457
Likes
18,256
Location
Monument, CO
I posted this picture over on WBF in the middle of a long semi-fruitful discussion on the benefits and drawbacks of differential operation. Ken Newton suggested sticking with the accepted technical definitions and distinction among differential vs. single-ended and balanced vs. unbalanced that I do think is appropriate even though I did not follow it here (I called the unbalanced interconnection single-ended, which it is, but may not be as clear to some).

Anyway, HTH - Don

p.s. I was going to post in the "Reference" subforum but only King Amir has that power... ;)
p.p.s. How do I edit the title? I left one period off my ellipsis...

20170210_wbf_diff_se_pix_large.png
 
Thanks Amir, for the kind words, the move, and fixing my typo. :)

I didn't post all the background behind the picture, but by itself it may help folk understand things a bit. I may tweak it to use "unbalanced" for the interconnects, and of course could expand to a big article that describes things in more detail. At some point I should probably just copy all those WBF technical articles over here but it's an insane amount of time I simply don't have; too many other things in Life and Work going on. I've had more time at work to post due to long test runs, but far less time at home, due to both long work hours and some health issues in the family that have led to a lot of phone/email time and several last-minute trips recently.
 
Special case. The output can be considered balanced but most cables route them single-endedly (there are, or were, those TTs that actually provide balanced outputs).
 
But the short take-away is that differential, balanced provides the best sound-quality, right?
 
But the short take-away is that differential, balanced provides the best sound-quality, right?

At audio frequencies there are far too many variables (for me) to make a positive blanket statement. In most cases I would not expect a significant (audible) difference, except when noise coupling is significant and/or there is a ground loop in the system. Then differential/balanced should provide better sound.
 
Special case. The output can be considered balanced but most cables route them single-endedly (there are, or were, those TTs that actually provide balanced outputs).
Audiophilia in a nutshell: a rational discussion about engineering, noise immunity etc. but applied to a system that is, in any way you like to look at it, noisy, inferior, expensive, fragile, high-maintenance, coloured, distorted, crosstalk-ridden.
 
Audiophilia in a nutshell: a rational discussion about engineering, noise immunity etc. but applied to a system that is, in any way you like to look at it, noisy, inferior, expensive, fragile, high-maintenance, coloured, distorted, crosstalk-ridden.

Well, yeah, but when did inferior tech specs ever stop anything in the audiophile world?

Tubes, reel to reel tape resurgence, single-driver speakers, etc.
 
Audiophilia in a nutshell: a rational discussion about engineering, noise immunity etc. but applied to a system that is, in any way you like to look at it, noisy, inferior, expensive, fragile, high-maintenance, coloured, distorted, crosstalk-ridden.

Now you're making me nervous.
 
Well, yeah, but when did inferior tech specs ever stop anything in the audiophile world?

Tubes, reel to reel tape resurgence, single-driver speakers, etc.
You make my point for me. :)
 
Now you're making me nervous.
Don't worry - I have no desire to talk vinyl versus digital; those days are long behind me. I'll not mention it again...
 
Where does a pseudo-differential cable from single-ended output to balanced input fall? @tomchr doesn't go into much detail in the link. Putzeys describes this a bit more in the "legacy pin 1" paper at Hypex and adds an RC network. Finally, Rane's tech note prefers this for such connections (fig. 5e) but notes that floating one end of the shield might help. (Whitlock's papers focus on balanced connections.)

These sources recommend how to connect SE to balanced, but none quantify the scale of the differences. The mountain of variables will make concrete answers difficult, but can we reach some approximations? (If answering this is far more complicated than I realize, let's first assume a sensibly engineered SE source running to an AES48 receiver via shielded cable.)

Specifically, how much rejection can pseudo-diff provide when compared to an RCA/RCA cable running to an XLR adapter? How much additional rejection could a truly balanced connection provide? Are the differences mainly seen in ground-loop hum or random noise?

Because many users will upgrade 1 component at a time, running from SE to balanced is common. When would ASR's EEs bother with these cables? Does simply upgrading to a balanced source ASAP make more sense?
 
A balanced connection can reject common-mode interference to something around 60dB, depending on the accuracy of the balance, and the frequencies involved, but 60dB is about right at audio frequencies.
Pseudo balanced is marketing speak for unbalanced, and often worse than normal unbalanced as the shield may or may not be connected at both ends.

Differential is usually (but doesn't have to be ) balanced, whilst balanced doesn't have to be, and often isn't differential, especially in things like USB bus powered sound cards, although some are.

S
 
Pseudo-balanced, what I call quasi-differential, comes in several flavors. You can use a resistor to match (roughly) the load impedance, use independent buffers with one inverting and one non-inverting, use a balun (transformer), etc. The amount of common-mode rejection depends upon how well matched the (+) and (-) sides are and how well the receiver at the other end works.

Shorting the shield (ground) of a differential (balanced) cable turns it into a single-ended cable as @sergeauckland said. How much that degrades the common-mode rejection depends somewhat upon the transmitter and receiver in the system but it will cut it at least in half (and often by much more).
 
Sounds like the consensus is it's not worth bothering in most cases. Thanks, folks.
 
Err, no. Tons of lower-end (home) studio gear uses merely impedance-balanced outputs with only one side driven. It works perfectly fine for keeping interference from common-mode sources at bay, since obviously small signal wise the impedance balance is all that matters. (Even the simple version with a direct output ground to cold connection has a decent amount of CMRR still.) It does confront the input stage with half the signal amplitude in common mode, and input stage common-mode distortion may be significantly worse than differential mode. Hence it's probably not the best option if you're after best distortion performance.
 
Rookie question : why is there a difference in the harmonic distortion between differential and SE?
 
Back
Top Bottom