Diff, diff

December 3, 2016
 by Paul McGowan

I had mentioned in yesterday’s post the standard way to build a balanced output stage was to simply tack on another amplifier that flips the phase around. That’s pretty much how it’s done, but it isn’t best because it’s “unbalanced”. A purposefully misleading statement.

Adding an extra stage for one signal that is not present for the other is more accurately described as asymmetrical—and asymmetries raise the hair on the necks of audio perfectionists.

A symmetrical way of making a balanced output is to start with a balanced amplifier. You’ll recall that most amplifiers aren’t actually balanced at all – they may enjoy a balanced input, or they may sport a balanced output – but that does not a balanced amplifier make them.

A true balanced amplifier (see why I hate that term?) is balanced from input to output and symmetrical for both signal phases. In this type of amplifier all that’s need for the balanced output is the proper XLR connector, not another stage tacked onto one side and not the other.


This is one of our designs and I hesitate to show it because it might be intimidating. There are a lot of squiggly things! For simplicity sake, we colored the inputs and outputs red. Input 1 and 2 are where the XLR cable would connect for the input—a simple diff pair (J1 and J2) as I described earlier. This feeds another diff pair (J3 and J4) that becomes the output. (for those technically inclined, ignore R11, it is not supposed to be there).

This is an entire true balanced amplifier with very few parts. Completely symmetrical from input to output, zero global feedback, and cheap. Less than $5 in parts. If we really wanted to use this in a product, we’d want to add some buffers to the outputs so it could drive complex loads, but basically, this isn’t all that hard.

Tomorrow I’ll cover balanced phono cartridges then we’ll wrap up the series.

Subscribe to Paul's Posts

23 comments on “Diff, diff”

  1. The deceitful thing in the industry is the fact that many products have XLR outputs that are not actually balanced outputs.When you open up the product you find that they are nothing more than an XLR connector with one Hot connector ( jumped over from the SE output) and one ground.. They are merely built in XLR “adpaters” built into the chassis and not balanced outputs..
    So to the unknowing customer, who he thinks he is getting the CMR benefit of a balanced connector…he is not.

        1. Now in all fairness I will point out that they do not claim they have balanced ins and outs…..they say they have XLR ins and outs..
          My point is that you need to do your research and you can’t assume that XLR connectors are an indication that you will get the benefits of a balanced circuit…. The Dog

  2. One more of the advantages of a truly balanced amplification is (as can be seen in the diagram) that the signals are “floating” with respect to ground ie they are free from contamination of minute leakage currents from the AC line transformers That flow in the signal ground and that exists in any system of audio electronics.

    People who have adequate systems of music reproduction at home are a minority compared to the rest, and within this minority, we who use truly balanced systems from the AES / EBU output of the transport, to the terminals of the amp, we are a A minority; But by current trends it is expected that this condition will change sooner rather than later in the class of audio that today is known as Hi-end.

    I consider that a truly balanced system is more rational, although in honor to the truth, I have heard single ended, that sound very pleasant to the ears. Like everything, this is a matter of taste.

    1. Single ended designs, and they are the majority of the designs out there, can and do sound excellent. There are many fine designs.

      True balanced like this, or a true differential amplifier is one of my favorite circuit types, but many, many PS designs have been single ended as well. You have to look at each application and see where it makes sense to do what, then go for it.

      I don’t know that we’ve ever made a design without a diff pair on the input, however.

      1. @ Paul

        As I said before, there are many single ended designs that sound nice, for example the PS Audio 250 Delta mono pair, which has a rich tonal balance and an interesting transient response (which can be appreciated with symphonic material) may be due to its Main multicap design, I do not know, because I’m not a designer, I’m just a man who can notice when an amplifier sounds good, but my preference is for true balanced designs.

  3. Despite being fully differentially balanced, many high end preamps, phono preamps and DACs have an external power supply, meaning another “box”, preumably to reduce there noise effect on the circuit by physically separating it.

    Why do you never see that concept applied to amplifiers?

      1. Yes, we invented that concept. The very first application was a preamplifier back in the 1970s. It allowed us to place bigger than normal transformers in external boxes. Those transformers were far too big to fit into a slim chassis.

        1. Ahem, with all due respect, i think what you meant to say is that your earliest experimental preamplifier designs were influenced by this man’s first product to market.


          In my mind, the LNP-2 established the concept of high-end audio. Jonathan Valin will disagree and mention Bill Johnson’s early design, the ARC SP-3 from 1972. Ironically we were both exploring the same Chicago based dealers in the early 1970s,
          Basil, Paul Heath, Victor and Simon.

          1. Actually, the above sentence should have read, “In my mind, the LNP-2 established the concept of the ‘modern’ high-end audio movement.

            The acknowledgement of who innovated high-end audio electronics belongs to an old friend, Saul Marantz. The 7C preamp and Model 9 mono amplifiers started the revolution IMHO, all respect to Frank McIntosh not withstanding.

            I had the pleasure of working and getting to know Saul at a CES show in Chicago in the 70s.

            Check out page 5;

  4. I’m not taking a position pro or con about balanced versus unbalanced amplifiers or whatever name you choose to call them. However, I’ll offer an observation (It’s been nearly 50 years since I studied this folks so bear with me if it isn’t exactly right.) All vacuum tubes work the same way more or less. Electrons are boiled off a cathode, controlled by a grid, and collected on a plate. The plate is connected to the positive side of a power supply, the cathode to the negative side, and the grid usually to the negative side. Always the same. Transistors can do something tubes can’t. They can go the opposite way, in fact either way. And so for example in addition to NPN transistors you can have PNP transistors. In addition to N channel FETs and MOSFETS you can have P channel counterparts. When carefully chosen for an exact mirror image match, their operations can be paired so that one half phase amplifier is the exact mirror image of the other. With a class A amplifier, you are stuck with the fact that wherever you bias it, the characteristic curve above the bias point will not be the same as the portion below it. They have little symmetry at all. Even in a so called balanced design using class AB as soon as you use a class A stage or a phase inverter, one side will not be exactly the same as the other. The output of the positive going half of the wave will not be exactly the same as the bottom going half compared to the input. The output stage of a class AB amplifier where both halves use the same type of transistor are called “quasi” complimentary symmetry amplifiers. For true complimentary symmetry amplifiers, both halves of the cycle have to be treated identically which means NPN and precisely matched PNP transistors throughout including the output stage. Does it matter? Will it make a difference? Is it better? How the hell do I know? It only matters to people who stay up at night worrying if they bought the right brand of connectors.

    1. I guess the electrons being freed by a lot of energy fed to the heating plate in a vacuum tube have a totally different behavior than seen in transistors (semiconductors) showing p-type (hole) conduction. Thus the reaction to sudden (!) (transient) changes in voltage should be different too.

  5. Paul, what is R 11 for? That can’t be right unless it’s for a headphone jack ??????? No, that’s not it. If it is, where is the other speaker terminal connection, ground? Something doesn’t look right here at all. Is this a test to see who is paying attention?

    1. If it was a test, you failed, reading comprehension 101, you need to brush up on your reading skills. From Paul’s Post ” (for those technically inclined, ignore R11, it is not supposed to be there).” And it was brought up by someone else, earlier.

      Paul I haven’t much experience with schematics, and that was about 40 years ago in an industrial electricity class at a Tech college. I think flowcharts would be easier to follow, for everyone without experience with schematics. Maybe a follow up with flowcharts would be worth a post.

Leave a Reply

Stop by for a tour:
Mon-Fri, 8:30am-5pm MST

4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

Stop by for a tour:
4865 Sterling Dr.
Boulder, CO 80301

Join the hi-fi family

linkedin facebook pinterest youtube rss twitter instagram facebook-blank rss-blank linkedin-blank pinterest youtube twitter instagram