Getting what you want
When we set out to prove one thing or another we arrange tests to prove our theory.
For example, if you’re trying to prove there are no differences between cables or amplifiers there are any number of ways to prove that. One would be the difference or null test where an identical signal is passed through two samples: say an expensive audio interconnect vs. a cheap one. If there were actually a difference it would show up on the scope as such.
Since we know that changing input cables—a high-end version vs. a dimestore copy—on a power amplifier in a highly resolving system is easy to hear, the null test should show the difference. Yet, it may not. Do we then conclude there are no differences?
If our goal is to understand why we hear a difference then it’s incumbent on us to dig deeper. Our hypothesis didn’t give us the results we were looking for. Our ears detect a difference our meters and methods fail to uncover. The proper conclusion is not to stop there but to march forward until it can be satisfactorily explained.
Garth Powell of Audioquest proposed a method that just might have some answers. Since the change we hear comes out of the loudspeakers and affects the entire audio chain, it’s only logical we measure the entire chain to seek differences. This would involve using a microphone to capture the output of the system and then comparing the recorded files to find the differences. It’s essentially the same test I have done any number of times with the microphone in my iPhone which more than adequately picks up differences.
I haven’t the time nor the interest in performing these tests with any scientific rigor, but perhaps someone else wants to grab the flag and climb the mountain. It would have to be performed on a system where we actually do hear a difference.
Proving what we already know might be valuable to someone.
Just not me.
We started a little mini-series on how amplifiers work. I know that some of you gloss over these details because you’ve heard them before or you don’t care. Others have large appetites for learning and it is to those hungry readers I continue.
If we look at the block diagram of a power amplifier in this post, we note two main sections: input and output. The job of the input stage is to take a small signal from the preamp and make it 30 times bigger. Let’s talk about that stage today.
The input stage of a power amplifier is a big preamp. It is also the one stage that makes nearly all the sonic difference in an analog amplifier. This is the most critical stage to get right. Just like a preamp’s architecture has everything to do with its sound, the amplifier’s input stage is where all the magic happens.
If we look at an amplifier like the BHK, we’ve used a vacuum tube to provide the gain. In most solid state amplifiers—those that are not hybrids like the BHK—this task of amplifying the small input signal is handled by any number of clever schemes. I have engineered simple op-amp style architecture with a single diff pair feeding a gain stage, to more complex versions known as full complimentary where there are multiple diff pairs and gain stages. The means to build a high voltage preamplifier are as many as there are amplifiers. Every engineer has their take on what sounds best in this all critical stage.
One technique we pioneered many years ago, though I am sure we weren’t the first, was the use of a separate power supply for this input voltage gain stage. It’s what we’ve done in almost every amplifier we’ve ever built and the improvements are clear to hear. In this scenario, there are two power transformers (or at least two separate windings on the main transformer) inside the amp: a small and a large one. The small transformer and its associated power supply feed the input stage, while the behemoth transformer is kept separated for the output stage. Here’s where we can get tricky. We can easily regulate this input stage, we can make sure it is never impacted by demands for power on the output stage. Thus our internal preamp is pure and undisturbed by our subject of tomorrow’s post, the less important but certainly not unimportant, current amplifier.