tracks with each track having an inductance of 4.5 mh. An impedance converter and one isolation amplifier droveeach track. The erase head was a J. B. Rea magnetic erase head and, like the record head, congisted of eight tracks, each track having an inductance of 4 mh. The erase head was connected as a plate load for the erase amplifiers, with each track dissipating 2.5 watts of 350 ko signal. The playback head is J. B. Rea high-inductance, narrow-gap-width playback head. The playback head had an inductance of 125 mh, and a gap width of 0.0001 inch. The recording medium was magnetic tape 1'4 inches wide, 0.005 inch thick, produced by Reeves Sound Craft. This tape was selected for its resistance to wear and imperviousness to high temperatures. Playback: The signal-switching commutator consisted of eight segments and a wiper segment. The purpose of the commutator was to play back each track of the record se- quentially. The commutator wiper was connected mechanically to the recorder capstan drive through a 2-to-1 gear train. The gear reduction allowed each track to be played through twice before stepping to the next track. The stepping sequence was counter- clockwise, playing back Channels § through 1, in that order. The playback amplifier was a four-stage pentode amplifier with an open loop gain of 40,000 and a closed loop gain of 4,000. The primary response criterion was chosen to be a high signal-to-noise ratio at the high frequencies. The overall signal-to-noise ratio, from record to playback, was 18 to 20 decibela, with a recorded signal of 50 ko and having square wave form. : In operation, a frequency reduction of 16 to 1 resulted from the speed change from record to playback: (A recorded frequency spectrum of 500 cycles to 50 ke would re- suit in a playback frequency spectrum of 30 to 3,000 cycles.) The playback-frequency spectrum was well within the frequency capabilities of the ground-station frequency modulation discriminator and recording oscillograph. The overall frequency response is shown in Figure 2.9 and the frequency response of the record head, tape, and playback head is shown in Figure 2.10. These response curves indicate the effects of the high and low frequency pre-emphasis in the impedance converters. 2.4 CALIBRATION This section describes the calibration of that portion of the instrumentation between the detectors and the magnetic-tape-recording amplifiers. 2.4.1 Detector 1 (Cal). Wigure 2.11 indicates the procedure. The pulse generator delivered a pulse of amplitude V and duration t to the gamma~integrating network in the canister. Since duration t was small compared to the network time constant, the charge delivered was g = Vt/R, coulombs. The resulting high-frequency converter output was recorded on the tape at high speed, together with the output of the precigion 32-ke oscillator. The tape was played back at low speed, and the reduced converter and oscillator frequencies were measured with a counter. The tape-speed-reduction factor was deter~ mined from the oscillator data and properly applied to the observed converter frequency to deduce the actual converter frequency associated with the input charge g. The relaHon is shown in Figure 2.12. 2.4.2 Detector 2 (Lil). The current input-voltage output characteristic of the Log-R was determined by calculation, employing the known componert values and bias voltages. 27