Aunt Corey's Homemade Buffered Passive Preamplifier Page 2
In a conventional system based on an active preamplifier, the signal source (phono stage, CD player, tape deck) is treated as a voltage source only; the high input impedance of the preamp enables the source component to drive its interconnect easily. In effect, the active preamp buffers the source component from the job of driving the interconnect feeding the power amplifier, tackling that job itself with its beefier output stage.
But a passive preamp has no such circuit; while the source component "sees" a moderately high input impedance (in most cases, 10k ohms), all the onus of driving the cables downstream of the control unit is placed back on the shoulders of the source, and most gear just isn't up to the task. Because when you use a passive preamp, the signal source has to work not only as a voltage source, but as a current source as well, and most phono stages and CD players (especially op-amp-based circuits) aren't designed to source much current. As a result, many listeners have complained about reduced bass impact and a lack of dynamic bloom to their systems' sounds with passive preamps.
There are other drawbacks to a purely-passive preamp; one is the high-frequency rolloff it can introduce when asked to drive long and/or highly capacitive cables to the power amplifier. The output impedance of a preamp and the shunt capacitance of a cable form a first-order RC filter that rolls off the high frequencies at 6dB/octave. This isn't a problem with most active preamps, as their output impedance is sufficiently low to raise the RC rolloff point (footnote 3) to well above the audio range, but some of the most highly regarded cables have fairly high capacitance, and using long lengths with a passive preamp can muffle the high end.
Passive preamps using a typical volume pot value of 10k have a worst-case output impedance of approximately 1.6k (depending on the position of the control), and this can also increase the cable's susceptibility to RF and noise pickup. Not only couldn't I use unshielded but otherwise excellent cables like Kimber KCAG with my passive preamp, but even fully shielded cables had to be positioned well away from AC cords, power transformers, etc., if I didn't want to hear a loud HUMMM from my speakers. In all, there are just as many reasons not to use a passive preamp as to use one, and it was this whole slew of limitations that started me thinking of A Better Way.
The passive preamp I had built was similar in most respects to those offered by the Mod Squad, EVS, and others: the signal source was brought to a Penny & Giles stereo 10k, conductive-plastic, audio-taper volume pot, and the pot fed the output jacks. Why does everyone use a 10k pot? Because it happens to be the best tradeoff between input and output impedance. Ideally, a preamp should have an infinitely high input impedance and an infinitely low output impedance, but as a volume pot's output and input impedances are directly related, you can't raise one without raising the other. Use a 100k pot so you don't load down your source components, and the output impedance becomes too high; lower the pot to a 5k unit to drive longer cables, and the input impedance becomes too low for most gear to comfortably drive.
So how could I simultaneously raise the input impedance and lower the output impedance? By changing my mantra from "Less Is More" to "Sometimes, A Little Less Less Is More Better," and using an active buffer.
The mighty BUF-03
A buffer is a unique device; among its many characteristics are an extremely high input impedance, an extremely low output impedance, hyper-fast signal-handling speed, and the ability to source a goodly amount of current. These devices are often called on to drive ultra-high-frequency video signals over great lengths of coax in 75 ohm circuits, and needless to say, a good buffer wets its pants with laughter at the thought of carrying mere audio signals; to call these devices "overkill" for audio would be an understatement!
But I happen to LOVE overkill, and so I began rifling through every data book I could get my hands on, looking for a suitable buffer. After listening to many promising units, I finally settled on the Precision Monolithics BUF-03. Offering a voltage gain of unity, this little FET-input, metal-can buffer boasts some pretty scary specs: 250V/µs slew rate; 63MHz bandwidth; 70mA peak current drive capability; an input impedance of 500 trillohms (5 x 10 (footnote 11) ohms!); and an output impedance of only 2 ohms! Not long after I decided upon the BUF-03, I learned that Kinergetics uses it for the output stage of their KCD-40 CD player. Mike Moffat also uses it for the output buffer in his Theta DS Pre and Pro converters; according to Mike, the BUF-03 is internally biased to class-A. So if you've marveled as I have at the sound of one of the Thetas, you've already heard audio through the BUF-03. It's fast, clean, and about as neutral as any single active device I've yet encountered (footnote 4). It's also expensive. [My 1986 edition of the TAB Books Linear IC Handbook quotes a small-quantities price-range of $11 to $35 each for the BUF-03, depending on specification---Ed.]
I'm a widdle ALPS pot, short and stout
By placing the BUF on the output of the pot, I achieved both of my original design goals: it lowered the output impedance to a couple of ohms, and since the BUF's own input impedance is so insanely high, it freed me up to go to a higher-value pot, thus raising the input impedance of the preamp. Of all the different pots I tried, the best-sounding was the high-grade black ALPS 100k pot Joe Grado uses in his HPA-1DC headphone amplifier. Unfortunately, while Joe was very gracious to sell me one, he's not in a position to sell any more; it seems ALPS only wants to ship these super-pots in very large quantities, so now that Joe's got his stash, the only places they're going are in his headphone amps (and Melos's, who get their black beauties from---you guessed it---Joe).
I feel bad telling you about this great pot that you can't get, so the second I find someone willing to sell them, I'll let you know. Other pots that worked well included the 10k and 20k versions of the Penny & Giles and, surprisingly, a $2 ALPS pot recommended to me by no less than John Curl! Radio Shack carries it under part #271-1732, and it's actually not bad at all; sure, the interchannel tracking isn't so great at lower volume positions, and it is a little grainy-sounding when compared with the P&G and the high-grade ALPS, but this little pot tied to a couple of BUF-03s kills a lot of highly regarded preamps out there for just a handful of change. And it's just the pot to use while you wait for me to figure out a way to get you the black beauty ALPS.
Footnote 3: Using the formula f = 1/(2xPIxRxC)---where PI is 3.1416, you can calculate the frequency at which the response will be down 3dB. An output impedance of 2.7k ohms feeding a cable with a shunt capacitance of 400pF will give a -3dB point of 147.44kHz, for example, which is of no consequence. But use three times the length of the same cable and the -3dB point will drop to 49kHz, meaning that some HF rolloff will appear in the audio band.
Footnote 4: Though the BUF-03 is a little noisier than the best integrated-circuit op-amps, this will not be apparent in systems that don't use ultrasensitive power amplifiers driving horns.---JA