The examination of the Brook Amplifier Mod. 12A

 

Fig. 1
 


The Brook  Mod. 12A is a small power all-triodes push-pull amplifier. It is constitued by a common-cathode first stage, a “cathodyne” inverter, a push-pull driver with an inductive load and Class AB1 operation output stage, Fig. 1. This amplifier constitues a milestone for the audio electronics and


 

Fig. 2


you can see it as a counter-current choice during an historical period where the resort to pentodes appeared as an irrevocable imperative. During the simulation of this amplifier, I was stunned by the results. In fact, although a

Fig. 3
 

marked overall feedback is present, the harmonic content at the maximum output power is rather limited in number and amplitude and approaches a “spectral imprint” very close to that of a ZOF amplifier. This results are summarized in Fig. 2 when the output power is 10Wrms on 8W load. The meaningful value for the THD is concentrated in output to the first stage, a result coherent with the modus operandi of the Negative FeedBack. The THD in output is “only” 0.978%.

 

Fig. 4
 


You can accept as conceptual challenge the achievement of  similar results by renouncing to the overall feedback and applying the HCT. The removal of FeedBack Loop forces you to effect a series of operations, Fig. 3:

1.    Increase the capacitive value of coupling caps;

2.    Eliminate the compensation network;

3.    Lower the open loop gain in order to reduce the sensibility.

Particularly, the sensibility reduction by lowering the load impedance in an intermediary stage of power amplifier could allow us also an optimal application of the HCT. For instance you can easily reduce the amplifier gain  by decreasing the value of grid resistances in the output stage. Since this  resistance contribute notably to the dynamic load of the driver, a correlated effect  will be an increase of  higher order products from this stage and consequently a possible THD reduction  in the output node since driver and output stage present complemetary phase characteristics. Pspice simulations show an optimal behaviour in terms of THD when grid resistors (marked as RG in Fig. 4) are close to 9k. In Fig. 46 a low value for RG is deducible in the very high  THD value on V(d) and V(e) nodes respectively. Here you can see the action of the HCT : two “distorted” signals from the driver, V(d) and V(e), are input to output stage grids, in order to produce an altogether “low” THD in the output node.  Clearly “classic” drawbacks are involved  in the elimination of feedback loop as:

1.    Damping factor reduction;

2.    Bandwidth reduction;

3.    Worsening of noise immunity;

 

 

What did you think of this article?
Click here to send us your comments, feedback and suggestions