38 THE BELL SYSTEM TECHNICAL JOURNAL, JANUARY 1956 



testing purposes, it was rather complex and there were some problems 

 in its construction and use. It was difficult to obtain suitable microwave 

 crystals to match the waveguide at low levels in the expander. Tliis 

 would make it even more difficult to build this type of pulse generator 

 for higher frequency ranges. Stability also proved to be a problem. The 

 frequency multiplier had to be very well constructed to avoid phase 

 shift due to drifting. The gate pulser also required care in design and 

 construction in order to get a stable and flat output pulse. It was some- 

 what troublesome to keep the gain adjusted for proper operation, and 

 the gate pulse time adjustment required some attention. The pulse 

 frequency could not be changed. For these reasons, and in order to get 

 a smaller, lighter and less complicated pulse generator, work was carried 

 out to produce pulses of about the same length by a simpler method. 



If the gated output amplifier of Fig. 1 were to have a CW instead of a 

 pulsed input, a pulse of microwave energy would nevertheless appear at 

 the output because of the presence of the gating pulse. This gating pulse 

 is applied to the beam forming electrode of the tube to obtain the gating 

 action. If the beam forming electrode could be pulsed from cutoff to its 

 normal operating potential for a very short time, very short pulses of 

 output energy could be obtained from a continuous input signal. How- 

 ever, it is difficult to obtain millimicrosecond video gating pulses of suf- 

 ficient amplitude for this purpose at a 100-kc repetition rate. 



A traveling-wave tube amplifies normally only when the helix is 

 within a small voltage range around its rated dc operating value. For 

 voltages either above or below this range, the tube is cut off. When the 

 helix voltage is raised through this range into the cutoff region beyond 

 it, and then brought back again, two pulses are obtained, one during a 

 small part of the rise time and the other during a small part of the return 

 time. If the rise and fall times are steep, very short pulses can be 

 obtained. Fig. 2 shows the pulse envelopes photographed from the 

 indicator scope screen when this is done. For the top trace, the helix was 

 biased 300 volts negatively from its normal operating potential, then 

 pulsed to its correct operating range for about 80 millimicroseconds, 

 during which time normal amplification of the CW input signal was ob- 

 tained. The effect of further increasing the helix video pulse amplitude 

 in the positive direction is shown by the succeeding lower traces. The 

 envelope dips in the middle, then two separated pulses remain — one 

 during a part of the rise time and one during a part of the fall time of 

 helix voltage. The pulses shown on the bottom trace have shortened to 

 about six millimicroseconds in length. The helix pulse had a positive 

 amplitude of about 500 volts for this trace. 



1 



