1210 
cordingly corrections must be applied for the tempera- 
ture effect. 
As in the case of the temperature element, the ab- 
solute values of the humidity-resistance relation may 
vary over wide limits between elements, provided the 
shape of the humidity-resistance curves remains rea- 
sonably unaltered. This latter requirement permits use 
of a simple evaluator similar to that employed for 
temperature reduction. 
The humidity strip is subject to several serious er- 
rors, some of which can be evaluated and corrections 
then applied. Others are not easily evaluated, and cor- 
rections are therefore difficult to apply. The more 
important sources of error are polarization, “washing 
out effect,” variation in the shape of the humidity- 
resistance curve, effect of direct exposure to moisture, 
lag, sensitivity to humidity changes at very low 
temperatures, and insensitivity to humidity changes 
at very high humidity. 
Polarization is a phenomenon exhibited by the strip 
whenever direct current passes through it. The effect 
is an increase in the resistance of the strip and a con- 
sequent lower humidity indication. In the early design 
of humidity elements, polarization was a rather serious 
source of error. However, with the use of wide strips 
(2 in. wide) and consequently higher resistance, polari- 
zation has been all but eliminated. For example, a 
change of less than 1 per cent relative humidity takes 
place during a 30-min continuous exposure of the ele- 
ment to constant humidity and temperature. 
Humidity elements exposed to near-saturation (96 
per cent relative humidity) for 30 min and then to air 
at 70 per cent relative humidity retained their cali- 
bration within 1 per cent relative humidity. However, 
the effect of direct exposure to moisture is far more 
serious. At this writing there is no unanimity of agree- 
ment among investigators as to the extent of damage 
caused by water droplets on the strip. This point cer- 
tainly requires more detailed treatment. 
A very important characteristic of any humidity 
element is its lag constant. According to Wexler [83] 
this constant for the lithium chloride strip (4 im. by 
34 in. by 46 in.) depends upon the magnitude and 
direction of the relative-humidity change and upon 
the relative humidity from which the change was made, 
in addition to the marked dependence upon tempera- 
ture. Although the response characteristics of the strip 
do not follow an exponential law, for convenience the 
lag constant will be given for a 63 per cent change in 
humidity, for the range of 50 to 70 per cent relative 
humidity. This procedure will give low lag constants. 
For ascensional rates of 1000 ft min~, the lag con- 
stant of the present element is approximately 4.0 sec 
for a temperature of 25C, increasing rapidly with a 
decrease in temperature, reaching values of 15 sec at 
OC and values in excess of 120 see at —30C. Although 
these high lag constants of the electrolytic elements 
leave much to be desired, they are superior to the hair 
hygrometer in this respect. 
Under proper conditions the electrolytic strip can 
be quite reliable. If the element is not subjected to 
METEOROLOGICAL INSTRUMENTS 
moisture condensations, it will give relative-humidity 
measurements with a probable error of 2.5 per cent 
for temperatures to —10C and a humidity range of 15 
to 96 per cent. 
Radtosonde Transmitter. A diagrammatic sketch of 
the radiosonde transmitter is shown in Fig. 5. This 
RE 
SY 
=n 
DESCRIPTION 
cl CAPACITOR, FIXED, PAPER .1 MFD + 10% 
c2 OO! 
cs -02-.03 
c4 ADJ.,CERAMIC 7-45 MMFD 
Lt COIL RADIO R.F, 
L2 R.F. CHOKE 
Pi CONNECTOR, MALE 
RI RESISTOR, FIXED, COMPOSITION 1500 OHMS 
R2 | AMPERITE D6M2 REG 
RS | | 22002 10% 
R4 WIRE 100,000 12 + 1% 
vi TUBE R.F. CAVITY OSC JAN 5794 
v2 v JAN 1U4 
Fic. 5.—Circuit diagram of the present 1680-me transmitter. 
V-1 is the new cavity oscillator. 
transmitter consists of a radio-frequency oscillator oper- 
ating at 1680 me sec and an electronic modulator for 
imposing intelligence onto the radio frequency carrier. 
As shown in the diagram, the radio-frequency oscil- 
lator is of the cavity type and is capable of yielding 
14 watt of radio-frequency power. 
“The electronic modulator is basically a blocking os- 
cillator whose repetition rate of 10 to 190 eps is 
controlled by the variable resistance of the meteoro- 
logical sensmg elements. From basic considerations, it 
can be shown that the oscillation time of the blocking 
oscillator should be as short as possible. To meet this 
requirement, the time constants have been adjusted so 
that the oscillation time is about 65 microseconds. 
While the blocking oscillator is on, the radio frequency 
carrier is quenched, and thus the radio signal reaching 
the ground has 65-microsecond segments deleted at an 
audio rate of 10 to 190 eps. To the direction finder, this 
represents an essentially continuous wave. 
The basic reason why the Diamond-Hinman radio- 
sonde lends itself so successfully to mass production is 
the fortunate relationship that exists between the audio 
frequency and resistance parameters of the blocking 
oscillator. It can be shown from an analysis of the 
circuit that the audio frequency f is given by the fol- 
lowing general relation: 
1 
dra ‘ 
