The Linearity of the Pentode Amplifier

In this article, we're going to look at the pentode to see how it compares with the bipolar transistor and the triode tube amplifier.

Using the circuit shown at right, the pentode current is measured as the grid voltage is varied from -15 volts to -5 volts. The graph below shows how the pentode responds:

Compare this with the common emitter bipolar amplifier curve. Look familiar? The pentode and the bipolar transistor both respond to an exponential transfer curve. For this reason, I expect it's distortion characteristics will share a lot with the bipolar amplifier.

Once again, it is instructive to see the effects of negative feedback through degeneration in the cathode circuit. Here we have used a typical cathode resistor, inspired by the VOX AC30 amplifier.

In this test, we varied the grid from -15 volts up to -3 volts (the response changes with R3 installed in the cathode circuit). Notice this time how much the exponential curve has been straightened out between a grid voltage of about -10 volts to -5 volts.

To test the EF86 pentode in the preamp configuration, the circuit shown at right was used, with and without a bypass capacitor. Then the outputs Out0 and Out1 were plotted in the spectral charts shown below:

The input signal used was 220Hz and it's amplitude was 1.2 volts. This input signal is nowhere near grid overdrive or grid cutoff. Yet much distortion was the result.

You can see both the even and odd harmonics showing in both.

Once again the bypassed cathode resistor preamp suffered the most distortion, since no degeneration was possible with the cathode at AC ground.

Grid Bias

One thing that must be noted immediately, is that this preamp appears to be purposely biased to produce distortion. The grid bias sits at about -2 volts! Earlier, when we did our linearity test with the cathode resistor in place (unbypassed), the tube went out of linear operation above a grid bias of about -5 volts. So clearly, this amplifier is not operating at the cleanest operating point.

So it is no wonder that both of these test preamp configurations produced a lot of distortion.

5kohm Cathode Resistors

Increasing the cathode resistor puts the grid bias at around -5 volts, which is still leaving the linear region of operation. However, as shown in yellow below, the preamp without the bypass capacitor (Out0), did clean up quite a bit with this bias change.

Note that no real improvement was had for the bypassed preamp Out1. All we did for that preamp was to move its range of operation, causing the strengths of the different harmonics to change slightly. In fact, it's 2nd harmonic almost disappeared.

The preamplifier as designed in the VOX AC30 is clearly not intended to be a clean preamp. It appears that the designers have deliberately put the operating point above the pentode tubes's linear range of operation. The other likely reason is that they wanted the guitar signal to drive the grid positive, with enough amplitude.

However, despite all of this, we can make a few observations:

1. The characteristic transfer curve is very much like the bipolar transistor (exponential).
2. an unbypassed cathode resistor allows the device to operate more linear (at reduced gain).
3. with bypassed cathode resistor, there is added gain and distortion
4. 2nd order harmonic content is generally much lower than it was for the triode tube.