FROM

Issue 98   Jan-Feb, 1992
 

THE SELF-MODULATED SCREEN GRID
Bacon, WA3WDR

Most amateur AM transmitters in the 50 to 200 watt class use plate-modulated tetrodes, and feed the screen grid through a series dropping resistor from the modulated plate supply.  Examination of the modulation linearity of this design shows a kink at about 85% negative.

Dean, WA1KNX, found a way to eliminate that kink by feeding the screen from a resistive voltage divider providing less modulation to the screen than to the plate. I tried that, and it worked for me, also.  (See "Screen Modulation Optimization for the Heathkit DX-100" by Dino, WA1KNX.  You can also see the original May, 1989 AM Press/Exchange issue containing Dino's article, "Class C Optimization for Ultra Low Distortion"  Note: the online version has the corrected schematic; there was an error in the original print version schematic.)

Then, I read about the effect of plate voltage on screen current.  With a given voltage on the screen, when the plate voltage is high, the screen does not catch as many electrons from the cathode, because the strong attraction of the high-voltage plate snatches most of the electrons away.  When the plate voltage is low, the screen catches more electrons from the cathode, because the low voltage plate is not attractive enough to snatch as many electrons away.  The screen will draw LESS current when the plate voltage is high, and MORE current when the plate voltage is low.

It looked as though a resistive screen supply ought to self-modulate in phase with the plate, without connection to the modulated plate supply.  I tried this in my Viking II - I simply connected the screen dropping resistor to the unmodulated side of the modulation transformer secondary, and it worked!  The kink was gone, and modulation linearity was excellent.  This simple change can be done in any of the screen-dropping-resistor rigs such as the Apache, DX-100, Ranger, Valiant, Viking I and II, etc.

Another advantage of this arrangement is that the screen current comes directly from the plate supply, and not through the modulation transformer.  This is an improvement, because it reduces the amount of modulator power required, and it reduces the amount of DC flowing through the modulation transformer secondary.  No new parts are required; just move the screen dropping resistor from modulated B+ to unmodulated B+ (the other side of the modulation transformer secondary winding).  It's not "something for nothing" - it's the elimination of a wasteful design mistake!

By the way, this self-modulating phenomenon is also how the "screen choke" modulators work:  a change in the current flowing through an inductor causes a change in the voltage appearing across that inductor.  An interesting factor appears here, though:  the higher the frequency, the more voltage will appear across an inductor for a given change in current through the inductor.  How is it, then, that the screen is not severely undermodulated at low audio frequencies, and severely overmodulated at high audio frequencies?

The answer is that the screen current varies in such a way as to "force" the voltage towards where the screen "wants" it to be!  If for some reason the screen voltage would start to go too HIGH, the screen would draw MORE current, and the voltage would go DOWN.  If the screen voltage would start to go too LOW, the screen would draw LESS current, and the voltage would go UP.  This negative feedback action compensates for the effect of reactive impedance over frequency and produces very accurate screen modulation with either the resistive or the inductive technique.

BACK TO HOME PAGE

THE AM PRESS/EXCHANGE