And if the amplifier does not use a global feedback loop, the sound will suffer. Unless, of course, we match the sagging negative rail voltage with a negative drive voltage. The circuit below shows the top device seeing the negative power supply rail noise reflected to its input, which will make it to the output. Is the answer to replace the cathode resistor with a constant current source?   

    The solution is to shift the power supply filtering capacitor connection from ground to the output stage's positive rail. A small change in layout, a large reduction in noise. Now both the top and bottom device's see the their required noise referenced drive voltages. The bottom device sees the same noise signal as its cathode or source or emitter sees. Thus without a difference there cannot be any amplification. The top device sees the sum of both the positive and the negative rail noises, which cancel to null. Once again, nothing to amplify.

    The circuit below shows the addition of the constant current source and the inversion of the noise at the output. What happened? The constant current source only shifted the source of the noise to the bottom device.

    The cathode resistor's value must equal half the plate resistors value or the noise cancellation will not occur. Thus the cathode resistor should not be replaced by a constant current source. The same holds true for replacing the plate resistors with current sources. So the answer to the question, "Are current source a good idea in tube equipment?" is sometimes yes, sometimes no, and sometimes it makes no difference.
   Look carefully at this last circuit and you will see that the circuit compensates for power supply rail sagging and expansion. Should the negative rail voltage creep upwards, the current flowing through the cathode resistor will decrease, which in turn will increase the plate voltages, as the voltage drop across the plate resistor will collapse.

< PREVIOUS

NEXT >