The Screen Grid and the Pentode

Find the offending component, which is not a component at all.

1. With a triode valve out of circuit and cold, measure the capacitance between anode and grid with an LCR meter.
2. Record it — it will be a few picofarads.
3. Now compute what that capacitance appears as at the input, multiplied by the stage gain plus one.

A few picofarads of anode-grid capacitance behaves like tens or hundreds of picofarads across the input, because the anode swings in the opposite direction to the grid and by a much larger amount. This multiplication is the Miller effect, and it means the input capacitance of an amplifier depends on its own gain.

Two consequences follow immediately, and both were serious. The huge effective input capacitance detunes and damps the tuned circuit feeding the stage. And because the feedback path exists at all, with the right phase shift the stage will oscillate on its own.

Fit a tetrode and repeat every measurement.

1. Build the same amplifier stage using a tetrode, with the screen grid connected to a steady positive supply and bypassed to earth by a good capacitor.
2. Measure anode-to-control-grid capacitance again.
3. Measure the stage gain.
4. Try to make it oscillate by increasing the load impedance.

Anode-grid capacitance falls by a factor of hundreds, gain rises by a factor of tens, and the stage is far harder to provoke into oscillating. The screen is doing two jobs at once: electrostatically it is an earthed shield between the two electrodes, and electrically it accelerates electrons past the control grid so the anode voltage barely influences the current at all.

The bypass capacitor is not optional. The screen must be at earth potential for signals even while sitting at a high DC voltage. Leave it unbypassed and the screen swings with the signal, the shielding evaporates, and you have an expensive triode.

The tetrode introduced a new fault, and its shape on the curve tells you exactly what it is.

1. Plot anode current against anode voltage at a fixed control-grid voltage, sweeping the anode from zero upward.
2. Pay close attention to the region where anode voltage is below the screen voltage.
3. Mark any region where current FALLS as voltage RISES.

There is a dip — a region of negative resistance — where increasing the anode voltage decreases the anode current. The cause is secondary emission: electrons arrive at the anode fast enough to knock further electrons out of its surface, and while the anode is at a lower potential than the screen, those knocked-out electrons are collected by the SCREEN instead of returning. Current that should have gone to the anode goes to the screen, so anode current falls.

A negative-resistance region is not merely untidy. It can sustain oscillation on its own — the dynatron oscillator was built deliberately on this effect — and it badly distorts an amplifier whose signal swings into that region.

Fix the secondary-emission problem with a third grid, and understand why it is at earth.

1. Build the same stage with a pentode, its suppressor grid connected to the cathode.
2. Repeat the anode current against anode voltage sweep.
3. Compare the curve with the tetrode’s.

The kink is gone; the curve rises and then flattens into a long horizontal plateau. The suppressor sits between screen and anode at cathode potential — negative relative to both — so it repels the slow secondary electrons knocked out of the anode and returns them to the anode, while the fast primary electrons from the cathode sail straight through.

The flat plateau is worth dwelling on: anode current is now almost independent of anode voltage. That is very nearly a constant-current source controlled by the grid, which is exactly what an amplifier wants — enormous voltage gain, and a stage whose behaviour barely depends on its load.

Put the pentode to work in the job the triode could not hold.

1. Build a tuned RF amplifier stage using the pentode, with a high-Q tuned circuit as its anode load.
2. Measure gain and check the tuned circuit’s Q with the stage running.
3. Increase the load impedance progressively and see how far you can push gain before instability appears.
4. Repeat the whole test with a triode in an equivalent circuit.

The pentode gives far more gain before oscillating, and it loads the tuned circuit far less, so selectivity survives. That combination is what made multi-stage IF strips practical — and therefore what made the superheterodyne buildable as a mass-market product rather than a laboratory curiosity.