12/02/2015 22:02

Neutralizing a Valve

The best way to exploit all the amplification available at a power amplifier in a MOPA is to have the power amplifier working as 'straight through', so not working as frequency multiplier. In the Literature it can easily be found that being 1 the output power as class 'C' amplifier, the output power would be 0.5 if the stage doubles the frequency, and 0.3 if the stage triples.

When working as straight through, usually the preceding stage is tuned, so that the power amplifier sees a resonant circuit. The impedance matching network towards the antenna is also resonating. Having two resonating circuits at the input and at the output, puts the tube into oscillation, i.e. it delivers power into the antenna without drive !

The situation is this: a portion of the output voltage is brought back to the input, with proper phase shift, so that, after amplification, it adds to the output voltage itself. If this deleterious effect is to be cancelled, then either the feedback should be cancelled (impossible, because it is caused by the internal capacitances of the tube), or it should be compensated.

In the following schematic, the feedback is provided by Cg1p (control grid to plate capacitance) and Cg1k (control grid to cathode capacitance). They are internal. You can't eliminate them. Their ratio in inversely proportional to the fraction of output voltage at the tube anode, that is brought back to the control grid.


Let's redraw the schematic by considering the parasitic capacitances, the tuned L-C network at the entrance and two additional capacitances Cn1, Cn2  connected in the circuit as illustrated. To understand the reshaping, follow the connection at the nodes marked 1,2,3,4:

This circuit is generally called a 'bridge': the internal branch, represented by the L-C parallel, bridges the two sides at nodes 2 and 4.

If the ratio between Cn1 and Cn2 equals that of Cg1p and Cg1k, than the two nodes 2 and 4 would have the same electric potential with respect to node 3, the ground. Therefore, 2 and 4 could be frankly shortcircuited and the overall network wouldn't realize that: they are equi-potential ... so no current would flow when bridged by a piece of wire. Therefore ... the L-C effect on the power amplifier, of inducing self-oscillation, is NEUTRALIZED.

From this point on, many are the comments. Some MOPAs described as having a straight through power amplifier are not provided by a Neutralization network: this is risky. It is practically impossible to have these circuit stable. It might happen if the driving preceding stage, despite its tuned output, has a low driving impedance with a very low resistive component, which would discourage further tendencies towards oscillations.

The Cn1/Cn2 possible ratios that equal the Cg1p/Cg1k are infinite, of course. In practice, the Cn1 capacitance is kept SMALL so that the neutralization setting results better independent against the antenna impedance network settings. Usually a Power Amplifier is neutralized at the highest working frequency because the output-to-input feedback tends to be higher, due to the lower reactive value of the stray capacitances in the tube. Typically, Cn1 is in the range of very few Pico Farads.

Cg1k depends not only on the internal, stray, capacitance of the tube, but also on external stray capacitances due to the components layout and the wiring. So, by considering only the declared internal capacitance values in the tube datasheet, it is possible to have only an estimate of the most probable Cn1/Cn2 ratio that allows the neutralization. Cn1 is usually variable, but in the 12BY7/6DQ6 MOPA, described in this website, it is Cn2 that is variable.

A tetrode is much more easily neutralizable than a triode. Its 'Screen Grid' greatly reduces the Cg1p capacitance, which is responsible for the bothering feedback. In the old WW II transmitter BC610, by Hallicrafters, the final tube, a 250TH, is a triode ('T') with high amplification ('H'): it is neutralized and works up to 18 MHz as straight through.

If the neutralization is understood first, and if is properly designed, then ... why sticking to adopt the 'Cathode Driven' configuration in Power Amplifier for amateur use ?

In my experience, the best way to neutralize a Power Amplifier tetrode provided with its suitable neutralizing Cn1/Cn2 network, is to disconnect the Screen Grid supply and inject a driving power in its Control Grid so that some power can be measured at a dummy load connected downstream to its Impedance Matching network. The Impedance network is set to read the maximum power delivered to the external load, the tuned input is also tuned for the same purpose, and the neutralizing network is set to read the minumum power delivered to the same external load: this condition means that the power amplifier tube does NOT GENERATE ITSELF high frequency power, by self oscillation !

It is a REAL satisfaction to operate a miserable, but extremely minimal and effective two tubes MOPA, from 1.8 up to 30 MHz ! And to see its ouput stage delivering the best it can because it works as straight through amplifier ever.

If you make your tubes cherry red by doing neutralization experiments ... they will forgive you. When this art is learned, it discloses very rewarding and satisfying use of wireless sets made the old way.

Try yourself.

Radioman, Dec. 2015

Posted by Radioman | Permanent Link