Audio Research D-250 II Servo power amplifier:
Page 2
All large-value capacitors in the signal path have polypropylene dielectric, and are bypassed with small-value polystyrene caps. There are only five such banks in the signal path: two per side plus an additional one for the phase inverter.

Phase inversion is accomplished in an unusual manner. Instead of drawing opposite-polarity signals from the plate and cathode of a single triode stage, Audio Research simply uses an extra triode stage (padded down to unity gain) to invert polarity for the "pull" side of the circuit. This eliminates much of the imbalance between the + and the - signals, which tends to occur as the result of differing source impedances between the plate and cathode of the classic phase inverter, yielding more effective cancellation of distortion.

The power supply is even more complex than the audio circuitry, seemingly embodying everything that could possibly serve to enhance and refine the amplifier's performance. There are, for example 11 separate secondary windings on the power transformer! Seven of these supply heater voltage to different banks of tubes, and all heater lines are DC-biased to minimize hum from heater/cathode leakage, and to increase reliability by reducing the voltage difference between the heater and cathode.

Four more 6550s regulate the output screen grid voltages and the front end HT, two more small dual triodes control the regulator circuits, and more than 10 assorted transistors and op-amps control and stabilize the regulator control tubes and other sources of critical operating voltages. I consider myself rather well versed in tube design topology, but there were several parts of the D-250's schematic whose functions I simply could not figure out. All I can say is that it all seems to work quite well.

Although the D-250 I tested was marked only "D-250" on its front panel, the instructions identified it as a D-250-II; an accompanying sheet further qualified it as a "Servo" model. (If you find this confusing, bear in mind that Audio Research and FMI co-founded the audiophile tradition of the Mod of the Month (footnote 2).) The "II" is self explanatory; the current D-250 is an improved version of the original D-250; the Servo is an improved version of the D-250 II. While I can only applaud this kind of evolutionary refinement of a no-holds-barred product, I wonder how the Mark II Servo designation makes owners of the Mark I feel. ARC does offer a I-to-II upgrade option (for $300), and an additional $750 will buy you the latest Servo version, but I really think ARC's customers would be a much happier breed of audiophile if there was less (or no) flaunting of every move to make obsolete last month's $6500 component outlay (footnote 3).

The "Servo" enhancement is basically just a means for stabilizing the DC balance between the two halves of the input stage. That may not sound like much, but it's a very important factor in minimizing changes in the amplifier's performance characteristics as the tubes gradually change during their operating life.

The D-250 is one of the few power amplifiers available which has input level-set controls—a feature that will rarely be needed in most installations, but which is of inestimable value when it is needed. For example, Infinity's RS-1B crossover provides no control of HF gain, and has barely enough gain in the LF section to balance out the upper-range speakers. The D-250's input controls allow one to get the additional 1 to 2dB of HF attenuation needed to match the LF precisely to the HF drivers. (The input controls also allow one to feed a high-level source directly to the power amplifiers, without any intervening preamplifier. But then, if you mate the D-250 with ARC's SP-11 preamp—and why not?—you won't want to eliminate the preamplifier anyway.)

Setup & System
Very early on during my tests of the D-250, I found that the amp—like the SP-11 preamp—must have an adequate warmup time before it can function at its best. Unlike some solid-state amps (the Threshold SA-1, for example), which reach peak performance in half an hour if no more than a day or so has elapsed since the last listening session, the D-250 seems to require at least two hours of warmup for each session, and continues to show gradual improvement for about four hours each time. (Fortunately, most of this warmup time can be done in the amplifier's Standby mode, which applies all operating voltages to the system but biases down the output tubes to about half their normal current draw.)

---

Footnote 2: By the way, in case you wondered what ever became of FMI, Bob was back at CES last month, showing a new line of Fulton loudspeakers. FMI is no longer the brandname; don't ask me why.

Footnote 3: In defense of Audio Research, it should be pointed out that the person who buys an ARC product when it first comes out and subsequently updates it to the latest version ends up paying the same as, or less than, the person who waits for the latest version before buying. Although I think that changes in model number and version do have the psychological effect JGH refers to, ARC has adopted a pricing policy that softens the blow, and eliminates the possibility of the early buyer feeling gypped. In our experience, all companies modify products on an ongoing basis, but many don't let you know what the changes are or how to obtain them; ARC at least lets you know how to get the best out of your older product.—Larry Archibald