|
|
OTL Output Stages Having scared many readers with the list of transformer faults, let us now examine the alternative, no output transformer. Tube OTL amplifiers have intrigued tube fanciers for decades. The eliminating the output transformer is like releasing the tube from a prison. If you a balking at this metaphor try the following experiment: place two tube output transformers in opposition to each other, i.e. connect the primaries together and attach one secondary to the output of an amplifier and the other secondary to a loudspeaker. If the transformer are perfect the sound from the loudspeaker should not differ from what it sound like when attached directly to the amplifier.
|
|
|
Note the use of the a plate resistor rather than a current source in the first circuit; it was used purposely. It is common knowledge that a push-pull amplifier will reject the amplification of common mode noise, and while this statement needs some fine tuning, it is basically correct. What is not common knowledge is that noise leaving a driver stage or phase splitter may not offer a balanced common mode noise for the push-pull amplifier to reject. For example, if one half of the balanced signal is thick with power supply noise, but other half is wonderfully noise free because of some power supply modification, then the noise will be amplified along with the music. Removing the modification will have the paradoxical effect of lowering the output noise by reintroducing the noise. Noise rejection requires balance in a push-pull amplifier. (Here is a question: How much noise can be rejected when one tube ceases to conduction in a Class AB, B amplifier?) The long tail phase splitter at first glance would seem the easy winner in balanced noise race and the split load phase splitter the clear loser, as it offers a very asymmetric PSRR on its outputs. The reality is that as long as we factor this asymmetry in our calculations, it is not really a concern. As all of this has been covered here before, I will jump to the punch line: if the split load phase splitter's grid receives half of the power supply noise, then the amount of noise on each output phase will be equal in both phase and magnitude. This magic results from the noise leaving the cathode in phase and equal with grid and the noise leaving the plate in anti-phase to the power supply's noise: 1 - ½ = ½.
|
|
|