The above circuit is an interesting one, in that it does not use a differential amplifier at its input. Instead, this balanced line stage amplifier employs two grounded-cathode amplifiers, with symmetrical loading by the same triodes, as used in the first half of the Aikido amplifier. (The circuit uses pure single-end operation with not a speck of SRPP, all appearances to the contrary.)

And since there is no differential amplifier, with its long tail—its large valued common cathode resistor—there is no need for a negative power supply rail. I know that many tube fanciers will breathe a long sigh of relief upon reading this, as many tube-loving solder slingers just hate negative power supplies. But before we move on to the rest of the circuit, let’s examine why a differential amplifier is used as the input circuit in most balanced amplifiers.

 

Differential Amplifiers

The differential amplifier is a simple circuit that holds two triodes (or pentodes, transistors, FETs, MOSFETs, heptodes…) and two plate resistors, but only one common cathode resistor. When a balanced signal is applied to the two grids, one triode conducts more, while the other conducts less, thereby creating balanced output signals at the triodes’ plates. The formula for this amplifier’s gain is a simple one:

             

(This is also the formula for a grounded-cathode amplifier with a bypassed cathode resistor or with a grounded cathode and fixed bias.) But what happens when the input signal is not balanced, when both grids see the same signal in the same phase? Then the differential amplifier offers very little gain indeed, as the formula below—for gain with a common-mode input signal—reveals.

             

(If you are troubled by the 2Rk in the denominator, imagine that the differential amplifier is made up of two grounded-cathode amplifiers being joined at the cathodes by connecting a wire from cathode to cathode, so that two cathode resistors have been placed in parallel, thereby halving their values.)

From inspecting the above formula, we readily see that the cathode resistor can make a huge difference in gain, as it is greatly magnified in value by the triode’s amplification factor. In fact, if an infinitely large-valued common cathode resistor is used, an infinitely small amount of gain results. Is this a good thing? Indeed it is. We want to amplify only the differences between the two input signals and reject and ignore all that is common between them, which—pretty much by definition—means noise. Thus, the bigger in value the common resistor is the better the common-mode-rejection ratio (CMRR). And thus, we see the popularity of replacing the common cathode resistor with a constant-current source.

             

or in dBs:

             

or as one easy formula:

            

 

Broskie Cathode Followers
If a high CMRR is so desirable, why not use a differential amplifier? Or, put a little differently, how are we going to get an equally high CMRR without a differential amplifier front end? The solution is found in the output stage, which consists of two Broskie cathode followers. This follower accepts a balanced pair of input signals and delivers a single phase output, which explains why two of them are required. In addition, the Broskie cathode follower offers an excellent CMRR. In other words, if the same input signal is applied to the top and bottom inputs of a Broskie cathode follower, very little of this common-mode signal makes it to the output of the follower.

 

Back to the Aikido Balanced Circuit
The input stage is able to forgo the usaul differential amplifier, because the output stage is effectively a differential follower, passing only what is different, rejecting what is common. In other words, the input stage amplifies both balanced and common-mode signals equally, but the output stage rejects common-mode signals. And since the input signal must be balanced and the output from this line stage amplifier is balanced, the power amplifier will must likely be a push-pull type that offers a balanced input, which means that we can expect a futher layer of CMRR applied to the signal.

Great. Wonderful. But how can I buy some of those shiny new PCBs or kits? Hang in there: I am still testing the the boards and I am also doing my taxes. In other words, until the taxes are done, nothing else can get done. This includes testing, writting user guides, adding products to the GlassWare Yahoo Store. But soon...or at least so I pray.