Balanced vs RCA preamps

atmasphere

It is a fact in high end audio though that most producers of balanced line products pay no attention to AES48 (the balanced line standard) ... That is part of the reason there is a balanced vs RCA debate.

Not really, the benefits of balanced designs are pretty well known. Designing to the AES spec is just one approach to balanced design.

...the Topping DACs; The 40 Ohm value is actually the two 20 Ohm output impedances put in series when really they should be in parallel, which would be about 10 Ohms. WRT to the ARC stuff they are doing the same thing.

The Topping and Audio Research schemes are different than your designs. That doesn’t make your design right or "proper" any more than it makes the ARC design "improper," as you’ve claimed. They each take a different approach to balanced amplification, either method offers improved CMRR. (Similarly, some speaker designers use sealed boxes, some use ported boxes and some use no boxes at all. Each approach can be valid and its success will depend on implementation.)

The standard for balanced line studio line level input impedance is 600 Ohms ... Our MP-1 preamp can drive 600 Ohms directly ... Most tube preamps will fall flat on their face trying to drive loads like that. In fact most solid state RCA preamps will too.

Most home users have no need to drive 600 ohm loads.

You have to be careful about things you read like the specs above! I own the two of the Topping DACs; The 40 Ohm value is actually the two 20 Ohm output impedances put in series when really they should be in parallel, which would be about 10 Ohms.

@atmasphere  One should be very careful when interpreting the impedance of a true balanced XLR cable. The two conductors (pin 2 and pin 3) carry both positive and negative signals from the source to the load simultaneously, behaving like a parallel connection. However, because they are of opposite polarity, the signals travel in differential (v.s. common) mode. The resulting voltage is the so-called differential voltage, which is twice the amount of the individual phase voltage.  This explains why the voltage from an XLR port is twice that of its RCA counterpart.

Similarly, the resulting impedance, known as differential impedance, is also twice the impedance per phase.  The manufactural reports either impedance per phase or the differential impedance.  Hope this helps.

The resulting voltage is the so-called differential voltage, which is twice the amount of the individual phase voltage. This explains why the voltage from an XLR port is twice that of its RCA counterpart.

@lanx0003 Actually this isn’t true if the source is AES48 compliant.

An output transformer is a good way to look at this if you understand how they work. In the case of a an output transformer driving a balanced line, there is a simple secondary with no taps. One end of the secondary output is tied to pin 2, the other end to pin 3. Pin 1 is typically tied to chassis.

Now if you want to run this output single-ended, pin 3 is tied to pin 1. You’ll note that the actual output Voltage is unchanged as is the output impedance. IOW either way the Voltage and impedance of the source is the same, which is to say there’s no 6dB increase when running balanced as opposed to single-ended. The transformer doesn’t care if one side is grounded or not. Your conclusion in the quote above is false WRT AES48 compliant sources, such as our MP-1 or MP-3.

Now if you have a non-compliant source (such as a Topping DAC or ARC preamp) then your comment is true.

But that isn’t how balanced lines are supposed to work. Here is a link to a Neumann microphone. Note the output impedance and the load its expected to drive (1KOhm). This mic is set up to drive balanced lines properly. A device that behaves as you described above does not.

Put another way, the idea that the output impedance of a balanced source is twice as high is true only if the balanced source isn’t designed properly to drive balanced lines!

To understand this issue better take a look at this page on the Rane website- its a quick read. If you scroll down to the ’Absolute Best Right Way To Do It’ you’ll see I’m not making this up. You might also read portions of this book by Bill Whitlock who designed most of the Jensen transformers. It gets more relevant to this discussion on page 13; if you look at the diagrams which show balanced connections, you’ll see no connection to ground except when it goes single-ended (such as driving an opamp in the case of an input transformer). Page 15 shows a microphone input. Page 16 shows a line output. I think you’ll see what I’m talking about.

atmasphere

... the idea that the output impedance of a balanced source is twice as high is true only if the balanced source isn’t designed properly to drive balanced lines!  ..

That is your opinion, and you're certainly entitled to it. You express it here frequently, as is your right.

But it is just as opinion, and one which many of the world's foremost manufacturers of audio equipment reject. That appears to trouble you greatly, perhaps because you seek to promote your trademarked and patented Balanced Differential Design® components. And that's fine, too. But when you claim that those whose circuit topologies don't embrace your favored methods somehow aren't "proper," you really make yourself look kinda silly, imo.

I have a lot of respect for you Ralph, as do many others on this forum. But for just about everything in life, there simply is no One Absolutely Right Way, and that includes the design of balanced circuits.

I'm telling you this Ralph because I think you do yourself a disservice by denigrating the designs of others.

But it is just as opinion, and one which many of the world's foremost manufacturers of audio equipment reject.

@cleeds I'm not sure that's true; AES48 is an industry standard but you'd have to check with them to see. As far as I can tell, a good number of them don't know what AES48 is. I've seen manufacturers state their gear supports AES48 but then you found out it really doesn't. That makes me wonder if they know what it is.

You mistake my reason for being active on threads where balanced line is discussed. The simple fact is I've been doing this longer than anyone else and I've seen an awful lot of misinformation online. Apparently if I don't say something, the chances are high that no-one else will step in and set the record straight. Would you prefer that I just don't participate?

Just to be clear WRT to the 'one absolutely right way' remark,  I didn't make this stuff up. We use a direct coupled technique in our products, transformers have been in use for the same purpose for the last 70-some years and That Electronics makes a chip at the link that is an example of the other way to drive balanced lines properly. There may be more ways but I am not aware of them and would love to be enlightened- after years of doing this, these three methods are the only ones I know of.

You are correct that my way isn't the only way. Before we introduced balanced lines to high end audio the only balanced products you came into contact with was studio gear. To be clear, the innards of the product doesn't even have to be balanced. My Ampex 351 tape machine was single-ended inside. That machine was the recording industry workhorse from about 1957 to well into the 1960s. 

I am not denigrating anyone else's work; I think you misunderstand. Instead what I am saying is that if AES48 is not supported, then you loose a lot of the benefit that balanced line operation offers: immunity to ground loops and reduced or non-existent cable coloration/interaction issues (this being the 'cable makes a difference' phenomena).

No-one would argue stating that you don't have to support the USB standard for cable because such a cable might not even work. Balanced lines are different in that they will still play despite the standards being ignored, much like speakers in this regard; you can throw some drivers in a box but unless you did your homework its unlikely to sound right. If you want a speaker to work right you have to pay attention to Thiel Small parameters, for a USB cable you support the USB specs, for balanced line its AES48.