Ken Fox
From:
Nashville GA USA
|
Posted 5 Aug 2022 3:03 am |
|
The following is part of an email from Jim Evans many years ago. He was my mentor and engineering hero for sure. I have many emails from Jim that I have kept in the Cloud. He was way ahead of his time with the Compactra amp
I hope you guys enjoy this excerpt from Jim. We were discussing my possibility of building a super Compactra with four output tubes. I finished the design years ago but realized the cost and weight of such an amp would limit sales
I would also still recommend the easy-balance of triode as the phase splitter. For best results it would be driven by a pentode in the same envelope. Here we are back to the tube 7199, which was very forgiving to variable-circuit-parameters, that result from component drifts or voltage drops. Thje 7199 will also take a great deal of negative feedback in the power loop, without bordering on instability. If you notice from the schematic I used a single 56pf mica capacitor across the feedback resistor, and never even came close to parasitic oscillation.
Other pentode/triode tubes have been used for this, but I have never tried any that come close to the 7199 from my evaluation standpoint. I noticed on the internet that the 7199 was again being made.
My friend Robert Crooks (who created Standel) tried some of the other Pentode/Triode tubes, which worked fairly well, but then dropped back to using the same dual triode as his preamp tubes,(12AX7 / 7025) for both the driver and the splitter. This will work, but the gain of the driver does not permit the degree of negative feedback that I would prefer, to still maitain the desired gain of the power amplifier. Still I would consider this probably the next best approach, if the 7199 is not available or affordable.
Now in the phase spiltter, many engineers went to great extremes to use precision resistors, to get the AC signal-balance for the power tube grids. Even then, two other factors are not automatically corrected, namely unmatched output tubes and slight impedance-mismatches in the output transformer's center tapped primary.
What I did here was to use the old McIntosh technology of an AC-Balance trimpot in the cathode side of the splitter. In this way, the signal can be set for optimum performance by simply using an intermodulation-distortion meter to set the AC balance pot. This not only takes care of all the variables with one simple tweak, but also permits a "100,000-mile-tuneup" every year or so, to restore linearity as the components age or drift.
The next item that would merit my concern is regulating the screen voltage of the output tubes. This is far more important that regulating the plate voltage by some "super-designed" high-voltage main-supply. If you hold the screen-grid voltage constant in the output tubes, you can get away with all kind of "sags" in the main DC supply to the plates. In fact RCA engineers learned that it would take more than ten times the regulation of the main supply (lots of weight in the power transformer core, heavier windings, and much larger filter capacitance) to match simply regulating the screen-grid voltage. The technique I used for this came from a friend and master-engineer (Al Hart) in Chicago, who designed all the "Grommes" and "Precision: amps in the early Hi-Fi revolution. This was simply an "emitter follower regulator" which for convenience and affordable performance was a standard (inexpensive at the time) glass 6L6.
Interestingly Robert Crooks likewise regulated the screen voltage for the (TV horizontal sweep) tubes he used for audio outputs. But since these required only a regulated +150 VDC he used the same gas tube regulator as was used in tube TV circuitry. He and I not only exchanged some ideas, but in 1960 when I worked in Dallas in Industrial Electronics, I bacame his area service-rep, and repaired Standel amps for his Dallas dealer.
One more approach I used in the Compactra-100 ,for output balance, was the "DC-Balance" I picked up this technology from the upper-end Hi-Fi amps, built by Fisher.
The same argument can be presented here as for the AC-Balance: Instead of expensive precision resistors, simply use an inexpensive trimpot to balance the plate current of the two output tubes.
One final operating parameter that was important in this power amp was the reduced bias (around -35 VDC) that made the tubes pull slightly more idling current, and eliminated even the slightest trace of Class-B (crossover) distortion. This runs the output tubes at adling current of about 135 mA, instead of the typical 70 or 80 mA recommended by most manufacturers. If the output tubes had been standard Beam-power-pentodes like the 6L6 this would have reduced their life expectancy by 10 to 15 percent. But I trusted the assurance of RCA engineers that the 7027 was still just "coasting" at this idling current.
The negative bias supply was obtained by dropping the power transformer's secondary HV-AC with the reactance of a capacitor rather than dealing with the I/R (wattage-heat) drop of a power resistor. It was then rectified by a simple diode and filtered before being sent to the bias pot. In later amps like the Hybrid-300 I had the power transformers custom wound and employed a special lower-voltage winding for th bias. But the Compactra-100 was built from off-the-shelf parts, and quite often from the surplus market that had over runs, where I could schedule purchases in increments 50 or sometimes even 10 at a time. Bear in mind this was the early days, and I think dinosaurs may have still roamed the earth, at least figuratively speaking. |
|
