An excellent discussion! Thanks to all involved.
While I have not had any direct experience with Erhard-Audio's auto bias module, from the description given of its operation, it is clearly working to address a completely separate (but important) issue, versus that which EFB™ addresses. The discussion that follows is based on the use of either of these two biasing approaches in a Class AB1 output stage, which the Fisher SA-100 employs.
With that said, when operating under quiescent conditions (only), there is a vague similarity between the action of the AB module and EFB, in that both circuits will seek to either maintain a constant current flow through the output tubes (AB Module), or minimize current flow differences (EFB), due to varying power supply B+ conditions. Under varying quiescent conditions:
1. The AB module (by definition) will sense a rise or drop in current flow through the tubes it controls, and automatically adjust the bias voltage to each tube to maintain the current flow through the tubes at precisely the pre-established current flow level it seeks to maintain. Since the module seeks to maintain a constant current flow through the tubes, it will do so whether varying power supply conditions or aging tube conditions cause any current deviation to otherwise exist. And, because the tubes it controls are all set to maintain the same level of quiescent current flow, it automatically provides optimum DC Balance as well.
2. On the other hand -- and specifically under quiescent conditions only -- EFB acts much more like a simple cathode bias resistor. As B+ voltage is elevated, the bias voltage created across a cathode bias resistor increases, which then minimizes the rise in current flow through the tubes under elevated B+ conditions. Conversely, when B+ falls the bias voltage across a cathode bias resistor also falls, which then minimizes the drop in current flow through the tubes under reduced B+ conditions. This action also describes very closely the way EFB action operates under quiescent conditions.
As emphasized however, the above analysis is only for simple static (quiescent) conditions. Under dynamic conditions however, the two approaches couldn't be more different, and spotlight the real differences between the two approaches. Consider:
1. As power output increases in a Class AB output stage, average current draw rather quickly rises. In this scenario then, the AB module will again seek to minimize this increase in current flow by raising the bias voltage applied to the tubes so as to return the increased current flow back to the pre-established value. Now no doubt, there are time constants built into the design, so that under what would be considered as normal residential use conditions with a musical (i.e., transient) signal -- even with the use of low efficiency speakers -- the module would not be allowed to react to every little musical peak that comes along. After all, if it did -- and did so perfectly -- there would be no increase in power output from the amplifier beyond that which the class A portion of operation could produce, since no increase in current flow would ever be allowed above the preset level. So the built in time constants are necessarily required so that the module will keep all the tubes at the pre-established current flow level under quiescent conditions, yet also allow for reasonable transient current increases to pass without adjustment by the module so that the amplifier can produce useful work.
The down side of this approach occurs when power output is increased sufficiently such that average output stage current draw becomes notably elevated in a continuous fashion. With a continuous elevation of current, an auto-bias approach would then typically seek to return the increased average current draw of the output stage back to it pre-established quiescent current level, limiting power output and increasing distortion. To some extent, the exact same thing happens with resistive cathode bias, as any average increase in output stage current flow increases the bias voltage across the cathode bias resistor, which again reduces power output and increases distortion. In stereo amplifiers that use a common cathode bias resistor for all the tubes, this approach works passingly well enough under transient dynamic conditions. Under steady state dynamic conditions however, the problems of this approach are spotlighted, where power output then falls significantly and distortion rises notably when both channels are driven, versus operating each channel one at a time. Of course a resistor is a passive device, whereas an auto-bias approach represents active management, which could enhance these concerns even further. From a purely theoretical standpoint then, an auto-bias approach would be ideal for a true Class A push-pull output stage, where no increase in average output stage current draw occurs with increasing power output.
2. In the same Class AB output stage however, EFB reacts very differently under dynamic conditions. Here, the EFB regulator now seeks to hold the bias voltage for the tubes absolutely stable in the face increasing output stage current draw -- whether it be from the application of a steady state or transient signal being applied to the tubes. As long as the B+ conditions remain unchanged, the bias voltage to the tubes remains unchanged, meaning that the tubes by definition are then operating with fixed bias. This of course allows for maximum power output to be produced, with minimum distortion. But it is the "Enhanced" portion of Enhanced Fixed Bias where the circuit really shines. This unique portion of the design allows the control grid voltage to remain in lockstep with the screen and plate voltages, which then maintains the optimum operating point regardless of how the B+ voltage might vary under high level signal conditions. Very specifically then, with EFB, bias voltage to the output tubes actually drops as the main B+ supply droops under high level power output conditions. That surely won't happen with resistive cathode bias, but this is exactly as it should be: As plate and (most specifically) screen voltages drop under high power conditions, then so should the bias voltage drop as well. In this way, EFB is a simple and uniquely effective back door approach to providing the same benefits as a fully regulated power supply would provide for a fixed bias Class AB amplifier. With a regulated power supply, no voltages change at all. With EFB, the main power supply voltage is allowed to droop as it typically will, but the relative voltage relationships at the tube elements all remain the same. In fact, it will be found that if the voltage level of an UNregulated power supply under conditions of maximum power output droops down to the same level to which a fully regulated supply operates at, then the performance of an amplifier using EFB and powered from the unregulated supply, will be virtually identical to that when powering it from the fully regulated supply without EFB installed.
Both approaches then act quite differently under dynamic conditions, and particularly so under elevated power output conditions. For the everyday set it and forget it use of a vacuum tube amplifier, where optimum quiescent current and DC Balance is always maintained as the tubes age or line voltage fluctuates, then an auto-bias module is hard to beat and a good way to go. Or, for those like myself who what to know that all the horses are available with maximum performance whenever the need arises, EFB has been tested and proven to provide that literally the world over for over a decade now.
I hope this helps!
Dave
