Over a dinner of wonderful Mexican food (for my tastes, California has the best Mexican food—far better than what passes for that here in Colorado) and with the help of a margarita or two, Bob laid out on a napkin his input stage to his new amplifier design's voltage gain stage (see yesterday's post).
This design was very much like our own: an input differential pair fed by an active current source, feeding a gain stage that then connected to an output buffer (this buffer is the input of the current amplifying section I wrote about yesterday).
Basically, a discrete op amp. Here's a basic picture of one from us long ago.
Q1 and Q2 are the input differential pair, Q3 is the gain stage, and Qs 4 and 5 are the output buffer (the current source isn't shown).
What was different with Bob's design was his was fully complimentary meaning he had twice the parts: one of these circuits for the positive going waveforms and another, upside down, for the negative going signal. This is called full complimentary, a more linear and far lower distortion technique than ours.
Here's an example I found on the internet. It probably looks complicated but if you take a moment to compare, you'll see the top and bottom input diff pairs and gain stages.
We of course knew all about such a circuit but had rejected it for one simple reason. A traditional full comp circuit has extraordinarily high open loop gain.*
This is because the gain transistor (Q3 in the first schematic) has an opposite sex mate connected to its collector. When two transistors are connected in this way they produce maximum uncontrolled gain. In our circuit, Q3's collector is tied down by R3, the 10K resistor which affixes it gain.
*Too much open loop gain is the nemesis of good sound. You have to add gobs of feedback to get the amplifier's output where you want it and this winds up squashing the sound, producing the same sort of harsh, lifeless audio so prevalent in the receivers and integrateds of the day. They had great low distortion specs but sucky sound.
Why did Bob's amp sound so damned good when we knew it shouldn't?
Bob grinned from ear to ear. He had discovered that a simple resistor to ground, tied at the junction of the two gain stage transistor collectors limited the open loop gain to where it sounded amazing.
"Doh!" It was so obvious. And yet, no one had ever done it before.
Genius is often the simplest solution.
