64 BELL SYSTEM TECHNICAL JOURNAL 



observation of the true shape of the spike has not yet been made, the dura- 

 tion is estimated to be of the order of 10~' seconds, a time interval that is 

 probably short compared with the thermal time constant of the contact on a 

 converter crystal. However, it is possible to measure the energy' content of 

 the spike, and such measurements indicate that this energy is fairly inde- 

 pendent of the length of the pulse and of the transmitted power level, 

 although it is definitely dependent upon the steepness of the wave front of 

 the transmitted pulse. The spike clearly represents energy transfer through 

 the TR box during the period required to establish the discharge conditions 

 which exist during the flat. The energy contained in the spike varies be- 

 tween a few hundredths of an erg to perhaps one erg per pulse, depending 

 upon a variety of factors. By way of comparison, the conventional crystal 

 rectifiers are proof tested in manufacture with a single spike of 0.3 erg to 

 5.0 ergs, depending upon the crystal type. It is generally believed that the 

 spike is more damaging than the flat in most radar systems. 



The energy in the spike is found to depend upon the repetition rate of the 

 transmitting pulses, presumably because of residual ionization in the gas 

 discharge gap. At low repetition rates (that is less than roughly 1,000 

 pulses per second), the spike energy may be materially decreased by a d-c 

 glow discharge near the radio frequency gap. This discharge provides a 

 continuous supply of ions and free electrons and so aids in establishing the 

 desired condition in the r-f discharge path. A discharge is supplied in all 

 the standard TR tubes. An auxiliary electrode called the "igniter" or 

 "keep-alive" is used as the cathode, with the back or inside portion of one 

 of the high frequency electrodes acting as the anode. A small amount of 

 radioactive material is placed in the tube to insure that the igniter discharge 

 starts on the application of the igniter voltage. Fig. 18 is a plot of the way 

 in which the spike energy varies with the repetition rate both with and with- 

 out an igniter discharge. Igniter oscillations sometimes occur as a result of 

 the negative resistance characteristics of the igniter discharge. This causes 

 a cyclic variation in the number of free electrons and ions with a resulting 

 fluctuation in the spike energy. Inadequate protection may result from 

 such oscillations. It is customary to mount a current limiting resistance 

 very close to the igniter cap to minimize the effects of these undesirable 

 oscillations. When such oscillations still occur they are usually evidence 

 of an insufliciently high igniter voltage or of tube failure. The margin of 

 safety in the igniter operation may be increased by increasing the discharge 

 current but at the expense of greatly reduced tube life. 



When a radar system is first turned on, the first pulse occurs without the 

 benefit of residual ions in the discharge, and for the first few pulses the spike 

 energy may easily reach dangerously high values. While the magnitude 

 of this "turn on" eflFect is greatly reduced by the presence of the igniter 



