Foreword by Constantine Soo
Steve Keiser is setting out to design a preamplifier of his ideal, and has agreed to share his thoughts on the priorities of the design with Dagogo’s readers.
Dagogo is privileged to be chosen as the exclusive platform for presenting such rare articles by the creative mind behind the original B&K Components ST-140.
At present, the Luminance Audio pre-amp circuit design is conceptual. There is no hardware constructed yet. A working prototype is at least two months away. I have, however, generated two designs on paper with all component values calculated.
One design is solid state and the other is vacuum tube. I decided that both would be built and compared with the sonically superior unit being the one actually going into production. I am still at the point where certain fundamental decisions are being made. The choices and individual parameters include the following list of options: tube or solid-state, remote control or no remote, cost-no-object or budge oriented, built-in phono stage optional or included in all pre-amps, balanced and single-ended input/output option.
At this stage of the article, I will address each one of these issues as they are decided.
Starting with the solid state design and considering the high level section. I set out with the high level section having 20db of gain. This is pretty much an industry standard. This means the input signal is amplified by 20db when the volume control is set to full volume. In consideration of this gain requirement, my design criteria is to make the circuit as sonically transparent as possible, with the degree of transparency subjectively verified by listening and comparing the effects that the line section have on the input signal by comparing with a passive pre-amp, and no pre-amp in circuit.
I started with the following basic circuit configuration: front-end featuring differential amplifier using J-309 FET transistors. The J-309 was selected for its high gain (200 milliohms at 2ma), wide bandwidth and low cost as well as abundant power supply. The differential is biased using bipolar 2N5210 with each FET having its own independent current source. The front end is loaded with a current mirror which translates the signal to the appropriate configuration for amplification by the next stage, which is a common emitter-amplifier consisting of a bipolar 2N5210 transistor, biased with an active current source which is also a 2N5210.
This stage accomplishes the function of providing the 20db voltage for the entire line section. During the design process, I considered two possibilities. Possibility number one is to use no “follower” and have the previously discussed stage drive the amplifier directly. Possibility number two is to have an additional follower stage drive the amplifier. If I went with first option, the line section would be a two stage design which would have the advantage of one fewer active device in the signal path, thereby intrinsically possessing a lower artificial signature sonic coloration on the input signal.
Using a follower stage has the advantage of significantly lowering the output impedance of the circuit, and thus increasing the current drive to the amplifier and long interconnect cables, as well as lowering the circuits sensitivity to different cable characteristics. Here are the numbers for those who are interested:
• Output impedance of circuit without follower stage is 2kΩ, output impedance with follower stage is 75Ω
• Current delivery capability of circuit without source follower is 1ma. Current delivery of circuit with follower is 50 ma
As you can see from the numbers for each of the possible configurations, the circuit using the follower output stage has far more flexibility for the end user, including the ability to drive loads down to 600Ω without sonic loss. The circuit without a follower stage can drive loads down to only 1.5kΩ, below which the circuit performance would be compromised. I decided to finalize the design of the basic topology using the follower stage because only the purist who’s only interested in maximum sonic transparency of the circuit would tolerate the numerous inconveniences that a no-follower topology would provide. Since I do have to consider the parameter of market share and product appeal to the greatest number of people, I decided to go ahead with the design with follower stage included. Of course, I suppose two versions of the pre-amp could be offered, one using the follower stage and one that does not. Or it may be possible to design the pre-amp with a user switchable option where the user could switch in and out the follower circuit as desired.
I therefore appeal to DAGOGO readers directly to give me feedback as to which configuration they think Luminance should incorporate into the final design.
- (Page 1 of 1)
