About fixed bias
Considered from the outside, this subject is still somewhat mysterious. Pittman managed to write one page while staying vague, some talk about an amount of current according to the tube, others speak about the grid/earth voltage. Everybody knows that a minimum of idle current is needed but not too much.

To try to understand all that, let's study the performance of an output stage for different values of idle current. The two parameters we decided to observe are the output power and the distortion. A Fender Princeton output stage is used for the measurements (6V6 push pull, modified fixed bias). The current is measured at the middle point of the output power transformer. Tubes are new and matched (medium grade).

The method consists of tracing the distortion curve acording to the output power, and for different values of the idle current. The result of the measurements shows us the famous "S" curves. We will notice that the scales are not linear. Wen you see the curves, some obvious facts appear:

  • A weak idle current is bad for the tone because the crossover distortion is big. Moreover, the output power is limited.

  • A medium idle current reduces crossover distortion and allows the nominal output power to be reached.

  • A high idle current reduces crossover distortion slightly more. The output power does not increase anymore. This high current leads to overheating of the tubes and finally, their premature wear.

  • The right value is slightly higher than the one allowing the amp producing its theoritical maximum power. Notice that the maximum power also depends on the tubes' grade. A tube with a high grade will produce more power than a tube with a low grade.

    For this amp, a total current of 42mA (21mA per tube) seems to be suitable. For this value, the amp is "hot" enough. Increasing the current will slightly reduce the distortion but the maximum power will stay the same.

    In practice, to view the crossover with an oscilloscope before clipping is useful. It allows us to visualise globally the amp performance and to detect a possible dissymetry. As it concerns the influence of the bias on the amp tone, as Ken says, the matter is far from being over...

    Back to the lab corner