OK -- Finally had a chance to draw something out, but a number of comments would be useful for those following along:
1. This project was my first effort at providing individual adjustment capability for all four output tubes while using one EFB™ regulator, and while it works just fine, time has shown that the adjustment process it requires is more tedious than I would like. That's partially because I intentionally included more adjustments for experimental purposes, and partially because of the configuration itself. Converting the adjustment procedure to a more conventional process could be accomplished a couple of different ways, but the most practical way of doing it is to simply use separate EFB regulators for each channel.
2. One approach would be to eliminate the EFB Range Control and fix it at a preset value, and then build more range into the existing bias controls that supply point C to each channel. This approach would allow to continue using just one EFB regulator to support both channels. The downside of this approach however is that to accommodate a usable range for typical tubes means adding more (potentially unneeded) voltage drop across the EFB regulator which it can certain handle (particularly with the new "bulletproof" measures applied), but goes against my approach of over-rating and under taxing SS devices. This is the reason that the EFB Range Control was included in the first place.
3. Ultimately then, the move to use dual EFB regulators has nothing to do with over taxing a single regulator for the amplifier. Even before the bulletproofing measures were added, continuous operation of both channels at sustained maximum power output caused the regulator to pass less than 15% of its rated current handling capability, at less than half of its rated voltage handling capability, and dissipate less than 3 watts in a worst case scenario. For a metal TO-220 device rated for 150˚C operation, it was uh, a walk in the park for the conditions it was operating under. With the bullet proofing measures in place, total worst case current passing through the regulator is now less than 8% of rated capability, and dissipation is reduced to just over 1.5 watts under a worst case scenario. As well, with the diode protection pack added that I developed and discussed in my bullet-proofing thread, the turn-on stress condition #2 identified in that thread is eliminated as well (stress condition #1 does not exist in this amplifier). So the choice to move to dual EFB regulators has nothing to do with relieving any normal operating stress on a single regulator design, and everything to do with being the most practical way to achieve a more conventional adjustment approach end. With dual regulators, the range of the DC Balance controls will allow for any reasonably good tubes to be easily balanced out, while allowing the EFB Regulator itself to set the bias voltage for the tubes eliminates any otherwise needless voltage drop across the regulator element.
4. In the sketch provided then, not only is the use of dual EFB Regulators shown, but included is the bullet-proofing elements now standard for all cathode-injected EFB designs, and some other improvements/conveniences in addition as well. Those include:
A. The Screen Grid Voltage Adjustment control has been eliminated, and fixed values shown for its replacement.
B. The power MOSFET used in the EFB Screen Grid Regulator now includes a 1K gate stopper measure -- not because it needs one, but because it is simply good design practice.
C. The power MOSFET is now partially bypassed which reduces its worst case dissipation to under 4 watts under full sustained power output in both channels. Under normal operating conditions, dissipation of the device is just over 1 watt.
D. As a result, I can now specify an STF10NK50Z device (or equivalent) for the MOSFET element. This is a 500 volt, 9A, 30 watt plastic device that (again) is significantly over-rated for the service it performs. In addition to no longer needing any insulated mounting hardware (because it's plastic), this device no longer requires the use of the Zener protection diode shown in the original schematic, either.
E. Note the value of the dropping resistor providing B+ to the AF Amplifier stage has been increased, and voltages shown as delivered to the phase inverter and AF Amplifier stages altered accordingly. The new voltages properly (i.e., better) center the operating parameters for optimum performance from the inverter stage.
Do note that NONE of these changes alter or improve the measured performance of the design as originally published. They are simply an update given for all the various reasons cited, as slowly applied over the last decade since the project was initiated.
That should about do it. I truly appreciate all the interest in the design. I am always loath to heap any praise on any of my own work, but this design remains on of my most favored amplifiers to listen to. The fact that it has remained so since the project started is the take away point to understand.
I hope this helps --
Dave
