Journal of Research of the National Bureau of Standards Vol. 48, No.

4, April 1952 * Research Paper 2312

Pressure-Humidity Apparatus 1
Arnold Wexler and Raymond D. Daniels, Jr.

An a p p a r a t u s for producing atmospheres of known relative h u m i d i t y , based on t h e

"two-pressure principle", is described. I t has a working space (test chamber) of 1 cubic
foot, in which t h e relative h u m i d i t y m a y be varied a n d controlled from 10 t o 98 percent, t h e
t e m p e r a t u r e from —40° t o + 4 0 ° C, t h e air flow up t o 150 liters per m i n u t e , a n d t h e test-
chamber pressure from }i t o 2 atmospheres. T h e h u m i d i t y in t h e test chamber m a y be set
a n d m a i n t a i n e d t o an accuracy of a t least }i of 1-percent relative h u m i d i t y .

1. Introduction (l — KPt)l(\—KPa) is shown in table 1 for the

applicable range of test-chamber pressures (% to 2
The pressure-humidity equipment is a laboratory atm) and relative humidities (10 to 100%). With at-
apparatus for producing atmospheres of known re- mospheric pressure in the test chamber the maximum
lative humidity by control of the pressure in the test error does not exceed }i of 1-percent relative
chamber of air saturated at a higher pressure* As humidity and hence, for most work, may be neglected.
this apparatus was developed primarily for hy- Even at a test-chamber pressure of 2 atm, the
grometer research and calibration, especially on maximum error is less than 0.5-percent relative humid-
electric hygrometer elements, which, during use, are ity. For very precise work, it is preferable to use
subjected to a wide temperature range, it was essential the empirical equation.
that the temperature of the working space should be
T A B L E 1. Error due to nonideal behavior of water-vapor-air
adequately controlled and maintained at any desired mixtures
value from —40° to + 4 0 ° C. Futhermore, it was
desirable that any relative humidity, from 10 to 98
Relative humidity, percent
percent, should be conveniently produced and that a
change could be made rapidly from any one value of Test-
chamber 100 80 60 40 20 10
relative humidity to any other value. Finally, it pressure
was also desirable to have a means of varying the rate Error in relative humidity, percent
of air flow through the working space, selected to
have a volume of about 1 ft3. oft* atm
H 0 0.02 0.05 0.0* 0.10 0.11
An apparatus has been developed that successfully 1 0 .05 .10 .15 .20 .23
meets the above requirements. I t operates on what 2 0 .10 .20 .30 .41 .47

may be called the "two-pressure principle." Basi-

cally, the method, shown in elemental schematic form The pressure method was developed by Weaver
in figure 1, involves saturating air, or any gas, with and Riley (see footnote 2) for the calibration of
water vapor at a high pressure and then expanding electrically conducting hygroscopic films used in the
the gas to a lower pressure. If the temperature is measurement of water vapor in gases. Their equip-
held constant during saturation and upon expansion, ment was designed for low rates of gas flow and was
and the perfect gas laws are assumed to be obeyed, used under ambient room-temperature conditions.
then the relative humidity, RH, at the lower pressure,
Pt, will be the ratio of the absolute values of the
lower pressure, Pt, to the higher pressure, P s , that is,
RH=100XPt/Ps..
Water vapor-air mixtures depart from ideal gas
behavior, so that the simple pressure ratio does not
strickly define the relative humidity. Weaver 2 3
has shown that an empirical equation of the form
Pt (l-KPt)
RH=100X
Ps (l-KP,)'
where the constant K has a value of 0.00017 when
the pressure is expressed in pounds per square inch,
more closely yields the true relative humidity. The
magnitude of the correction introduced by the term F I G U R E 1. Simplified schematic drawing of the principle of
1 operation of the pressure-humidity apparatus.
The development of this apparatus was sponsored by the Aerology Branch
of 2the Bureau of Aeronautics. A, high-pressure air source; B, pressure regulator; C, saturator; D, pressure
E. R. Weaver, and R. Riley, J. Research NBS 40,169 (1948) RP1865. gage; E, expansion valve; F, test chamber; G, pressure gage; H, exhaust valve;
* E. R. Weaver, Anal. Chem. 23,1076 (1951). I, atmosphere or vacuum source; J, constant-temperature bath.

269
Their saturator was a small cylinder, containing heat exchangers A17, A19, and A21 and centrigfugal
water and filled with fragments of pumice or stream- saturators A18, A20, and A22 precipitates this excess
washed gravel through which the gas could be water vapor so that just complete saturation is
bubbled under pressure. obtained in the final saturator, A22. The bath
saturators and heat exchangers are shown in figure 5.
2. Description of Apparatus The saturators are simple in design and function
equally well below as well as above the freezing
The apparatus is shown in block diagram in figure point of water. They are similar to a type pre-
2 and schematically in figure 3. I t consists, essen- viously used 4 with considerable success. Each
tially, of the following functional components, through saturator consists of a cylinder to which water is
which air flows continuously: A high-pressure air added to a convenient depth. Air is discharged
source, a low-temperature drying system, a filtering through a nozzle into the chamber above the water
system, a warm-up unit, a pressure regulator, a surface and tangential to the vertical walls and is
humidifying system, an expansion valve, a test exhausted through a central port in the top. T h e
chamber, and an exhaust control system. centrifugal force creates a whirlpool action that
Air is supplied from a 250-psig reciprocating thoroughly mixes the water vapor with air. Spray
compressor, Al, figure 3, capable of delivering 5 and liquid water are forced to the walls by centrif-
ft 3 /min of free air at room temperature. This air ugal force, so there is little tendency for liquid water
is filtered (A2) to remove pipe scale, dirt, dust, or
other solid material, and is then introduced into the
drying system (fig. 4).
Water from the air at the supply pressure is
removed by freezing in a train of four drying unite
immersed in a bath, D, containing a mixture of dry
ice and Stoddard solvent. T h e first two units, A4
and A5, are large-capacity centrifugal water sep-
arators. These are followed by a copper coil, A6,
and a baffle dryer, A7. Particles of snow or ice and
droplets of oil (from the compressor) are caught by
a filter, A8, which is maintained immersed in the
same low-temperature bath, D.
The air emerges from the dryer at about —78° C.
I t is then heated in the warm-up unit to a tempera-
ture somewhat greater than that being maintained
in the thermostatted liquid bath, B. The warm-up
unit consists of an electric heater, A9, that is con-
trolled by a thermoregulator, A10. Two pressure
regulators, A13, and A14, in series reduce the pres-
sure and maintain i t constant in the hymidifying
system.
Saturation is accomplished in four stages. The
air is first passed through the external gross satura-
tor, A15, which is kept at a higher temperature than
the bath, B, and then through the three bath satu-
rators, A18, A20, and A22. The gross saturator,
because of its higher temperature, introduces water
vapor in excess of that required for complete satura-
tion at bath temperature. The combination of

HIGH LOW
AIR WARM-UP
PRESSURE TEMPERATURE
FILTRATION UNIT F I G U R E 3. Schematic flow diagram of the pressure-humidity
AIR SOURCE DRYING SYSTEM
apparatus.
i '
9 <? Al, high-pressure source; A2, filter; A3, valve; A4, centrifugal water separator;
A5, centrifugal water separator; A6, copper cooling coil; A7, fin air dryer; A8,
low-temperature filter; A9, electric heater; A10, bimetal thermoregulator; A l l ,
TEST EXPANSION HUMIDIFYING PRESSURE air reversal valve; A12, shut-off valve; A13, pressure regulator; A14, pressure regu-
CHAMBER VALVE SYSTEM REGULATOR lator; A15, external gross saturator; A16, resistance thermometer and indicator;
A17, 19, 21, 25, copper coil heat exchangers; A18, 20, 22, centrifugal saturators;
1
A23, 27, pressure gages; A24, expansion valve; A26, test chamber; A28, linear
i
flowmeter; A29, exhaust control valve; B, insulated liquid (Stoddard solvent,
constant-temperature bath; D, insulated dry-ice bath; HI, bimetal thermoregu-
lator; H2, electric heater; H3, input voltage; T l , Stoddard solvent cooling coil;
T2, positive rotary displacement pump; T3, motor; T4, Stoddard solvent bypass
EXHAUST valve; T5, Stoddard solvent control valve; T6, centrifugal stirrer; T7, constant
ROOM OR
CONTROL
VACUUM SOURCE
electric heater; T8, intermittent electric heater; T9, 10, thermistors; VI, vacuum
VALVE source; V2, vacuum shut-off valve.

F I G U R E 2. Block diagram of Jfie pressure-humidity apparatus. 4 Arnold Wexler, J. Research NBS 4 5 , 357 (1950) RP2145.
 F I G U R E 4. Drying system and Stoddard solvent cooling coil.

F I G U R E 6. Test chambt

FIGURE 5. Bath saturators, heal exchangt and test chamber.

to emerge t h r o u g h t h e exit p o r t . A multilayer fine-

wire screen baffle is used at the exit as a furl her guard
against the escape of liquid w a t e r . As air does not
bubble t h r o u g h t h e w a t e r b u t only passes over its B S
exposed surface, the w a t e r may be frozen without
FIGURE 7. Temperature control circuit.
impairing the functioning of the s a t u r a t o r .
Upon emerging from the final s a t u r a t o r , A22, T, Thermistor; K, precision helical rheostat; B, 22y±-x battery; S, battery switch;
the pressure of t h e air is reduced by expansion A, microammeter; (), Vi a galvanometer relay; P, input from pulsing circuit; V,
- fusts J, power relay; E, input voltage for power relay coil; U, intermittent bath
valve A24. Because a temperature drop max occur heater; V, input voltage for intermittent neater; L, pilot lamp.
in the air at the expansion valve, a final heat ex-
changer, AJ85, is provided for bringing the air to
bath temperature before it enters the test chamber, the room air. T h e c h a m b e r is suspended from a.
M6. counterweight system so that it can be easily raised
The working space, A26, is a, cylindrical chamber above or immersed into the Liquid bath, />. A length
h a v i n g a nominal volume of 1 ft 3 . I t is shown with of flexible metal hose, between the test c h a m b e r ,
the cover removed in figure (>. T u b u l a r o u t l e t s A2(>\ and the final heat exchanger, A25, p e r m i t s the
extend from the c h a m b e r to allow electric leads to test, c h a m b e r to have the required motion.
be brought in and out of the working space. T h e T h e s a t u r a t i o n pressure, Pt, is measured in the
air discharges from the c h a m b e r into a, linear (low- final s a t u r a t o r , A'22, by gage A£8, and the test
ineler, A28, and then t h r o u g h a n exhaust, control pressure is measured in the test chamber by gage A27.
valve, A£9} into a v a c u u m source or simply into These m e a s u r e m e n t s are made with high-precision

271
laboratory test gages that have been calibrated they will maintain a new pressure in the saturators.
against the National Bureau of Standards pressure To maintain a constant air flow, a minor adjustment
standards. For atmospheric or reduced pressures, of the expansion valve is also made.
a high-quality calibrated aneroid barometer is used. When atmospheric pressure is desired in the test
From atmospheric pressure up to 2 atm, a mer- chamber, the air emerging from the chamber is
curial manometer is employed. The higher pressures allowed to exhaust directly into the room. Elevated
are determined with either a 0- to 50-psig or a 0- to pressures in the chamber are obtained by throttling
200-psig Bourdon tube gage. the flow from the chamber by means of the exhaust
The temperature of the liquid bath, B, can be control valve, A29, figure 3. Reduced pressures in
adjusted to and closely regulated at any temperature the chamber are achieved by attaching a vacuum
from —40° to + 4 0 ° C by a simple on-off thermo- source to the exhaust control valve and adjusting
statting system. Stoddard solvent, which is used as the valve to give the required reduced pressure.
the bath liquid so that low temperatures may be In operating below freezing, one precaution must
attained, is actively agitated by a centrifugal stirrer, be observed. The level of the water in each satura-
T6, and is circulated from the bath, B, by a positive tor must be kept below the inlet nozzle, otherwise,
rotary displacement pump, T2\ through a copper on freezing, the opening will be sealed by ice.
coil, Tl, immersed in a mixture of dry ice and The equipment may be operated continuously for
Stoddard solvent in bath D and back into bath B. 8 to 16 hr, after which the accumulated water in
By proper manual setting of a bypass valve, T4, the dryer should be removed. Failure to do so may
and a control valve, T5, the rate of Stoddard solvent result in clogging of the dryer by ice and the reduc-
flow is adjusted so that the bath B tends to cool tion, or even complete stoppage, of air flow.
slightly. The desired bath temperature is estab- The removal of water from the dryer is accom-
lished by resistances (coupled helical precision plished in two steps. First the dry ice bath is
rheostats), R, in a Wheatstone bridge circuit, figure 7. raised to room temperature by a thermostat, HI,
Thermistors, T, with temperature coefficients of 4 and heater, H2. Then suction is applied by vacuum
percent/deg C, are employed as the temperature- source VI and room air drawn through the reversal
sensitive elements. Any cooling disturbs the bridge valve, Ally and dryer until all the water has been
balance, which is sensed by a galvanometer relay, evaporated. Overnight operation usually suffices to
G. A current of %pa deflects the galvanometer remove most of the water.
pointer against a magnetic contact, actuating a
power relay, P , and, in turn, intermittent bath 4. Performance and Accuracy
heater, H. The operation of an electromagnetic
plunger returns the pointer to a sensing position. The psychrometric and dew-point methods of
If the bridge is unbalanced, the pointer will deflect humidity measurement were used independently to
and again throw on the heater; if the bridge is in evaluate the accuracy of the humidity produced by
balance, the pointer will remain in a null position. the equipment. A thermocouple wet-and-dry-bulb
An electronic pulsing circuit, P , periodically triggers hygrometer was employed over a wide range of
the plunger so that the pointer may sense the bridge relative humidities and at temperatures from 0° to
balance. 30° C. A dew-point instrument having a thermo-
couple embedded just below the surface of a small
Qi in. in diameter) mirror for temperature measure-
3. Operation of Apparatus ment, manually controlled heating and cooling of the
mirror, and visual observation through a telescope
The method of operation of this equipment is for dew and frost detection, was constructed and
simple. The instrument, material, or device under used to measure dew points from room temperature
investigation is inserted into the test chamber, the down to —27° C. A series of experiments was made
latter closed and immersed into the liquid bath. in which the relative humidity measured by the
Distilled water is added to each saturator to an above two methods was compared with the relative
appropriate depth. Solid carbon dioxide is then humidity given by the ratio of the test-chamber
added to the dry-ice bath, D. The temperature of pressure to the saturator pressure. The results are
the liquid bath is brought to and maintained at the shown in table 2. The average difference in per-
desired value. Air from the high-pressure source is centage of relative humidity between the psychro-
allowed to pass through the apparatus and the metrically determined values and the apparatus
pressures in the saturators and test chamber adjusted values given by the pressure ratio is ± 0 . 4 percent,
to give any preselected relative humidity. The and the average difference in percentage of relative
thermoregulator controlling the temperature of the humidity between the value determined by dew-
air passing through the warm-up unit is set to main- point measurement, and the apparatus value given
tain a temperature in the external saturator several by the pressure ratio is ±0.6 percent. Similarly,
degrees higher than in the liquid bath. When ther- the algebraic average differences are —0.2 and 0.0
mal equilibrium had been established in the com- respectively. I t may be assumed that as there is
ponents in the liquid bath, the pressure ratio indi- no marked tendency for the differences to be either
cates the correct relative humidity in the test cham- positive or negative, the air passing through the
ber. Changing the relative humidity primarily saturators emerges neither supersaturated or under-
involves adjusting the pressure regulators so that saturated.
272
 TABLE 2. Summary of calibration solvent). However, by substituting water for Stod-
dard solvent as the bath liquid, the upper end of
Difference in rela- the temperature range m a y be extended to about
tive humidity
Relative Relative humidity
between pressure- 90°*C.
measured by—
Nominal humidity
ambient produced
humidity appa-
ratus and—
The accuracy with which any desired relative
Date of humidity may be established is a function of the
bath by pres-
run temper- sure-hu-
ature midity Dew- Dew- Psychro- uniformity of temperature in the apparatus, par-
apparatus point metric
h $ S L Psychrometerj measure- measure-
ticularly in the final saturator and the test chamber.
eter ment ment The relative humidity in the test chamber will be
equal to the pressure ratio of the test chamber to
% % % final saturator only if these two units are at the
96.1 96.2 -0.1
90.1 91.1 -1.0 same temperature. The distribution of temperature
81.6 82.3 -0.7 within the saturators, test chamber, and surrounding
3-28-51 23.5 74.4 75.2 -.8
59.6
50.1
59.7 -.1
-.4
liquid bath was explored by means of thermocouples,
50.5
38.0 38.4 -.4 located at the inlet, outlet, and in the water of each
25.1 25.4
saturator, near the front and rear of the test chamber,
92.9 92.4 +.5 and at four separate points within the liquid of the
79.8 79.8 .0
3-28-51 9.7 61.0 61.0 .0 bath. As an indication of the variations in temper-
49.5 49.3 +.2 ature that may exist in the apparatus, data are
37.2 37.0 +.2
24.6 24.2 +.4 presented in table 3 of the average temperatures
92.6 92.9 -.3 at various locations for three 2-hour rung at different
78.3 78.6 -.3 ambient temperatures. I t may be seen that the dif-
3-28-51 59.6 59.7 -.1
49.2
37.0
49.4 -.2 erentials are of minor magnitude, especially between
37.2 -.2
24.9 24.8 +.1 the final bath saturator and the test chamber.
97.9 96.7 +1.2
88.8 88.7 +0.1
77.6 75.9 +1.7 TABLE 3. Temperature distribution
4-6-51 24.7 77.4 77.9 -0.5
54.8 55.4 -.6
36.9 37.3 -.4 Location Temperature
23.0 23.9 -.9
96.9 97.0 96.9 -.1 0.0 Initial b a t h saturator: ° C ° C ° C
90.6 92.1 91.4 -1.5 -.8 9.55 23.33 30.24
4-16-51 6.0 77.9 78.6 78.5 -0.7 Air i n l e t .
-.6 Airoutlet __ 9.57 23.30 30.25
57.7 58.5 58.2 -.8 -.5 Water 9.66 23.29 30.30
42.7 42.9 42.6 -.2 +.1
Intermediate bath saturator:
96.2 96.3 94.7 -.1 +1.5 9.63 23.35 30.45
Air i n l e t . __
53.9 54.7 53.8 -.8 +0.1 Airoutlet . 9.65 23.36 30.44
4-18-51 6.0 36.6 36.6 36.7 .0 -.1 W a t e r _. _._ 9.63 23.29 30.39
28.3 28.4 28.8 -.1 -.5
23.2 23.1 23.5 -.3
+.1 Final bath saturator:
9.64 30.48
Air i n l e t . . 23.39
96.8 97.1 — .3 Airoutlet __ 9.65 23.40 30.44
4-19-51 86.0 86.7 -.7 Water. 9.66 23.36 30.43
74.7 74.7 .0
Test chamber:
76.4 75.9 +.5 Front 9.64 23.35 30.40
-9.6 68.3 68.2 +.1 Rear 9.61 23.35 30.37
54.1 53.5 +•6
Bath:
96.3 95.6 +.7 Side of test c h a m b e r 9.64 23.38 30.40
88.0 89.1 -1.1 Expansion valve. .. .. 9.64 23.39 30.40
4-25-51 -19.8 78.7 78.2 +0.5 R e a r of test c h a m b e r 9.61 23.39 30.39
66.4 65.0 +1.4 Bottom __ 9.58 23.38 30.39
56.7 55.6 +1.1

Arithmetic
avg ±0.6 ±0.4 The constancy of bath temperature is of impor-
Algebraic
avg ... .0 -.2 tance, for quite often materials or hygrometers under
investigation are temperature dependent. The con-
The relative-humidity range obtainable is limited trol system has effectively regulated the bath at
by the range of ratios of test-chamber pressure to temperatures from —40° to + 4 0 ° C. For periods
saturator pressure. The maximum saturator pres- of time of 2 to 5 hours, average fluctuations of
sure that can be employed with this apparatus is 0.02 to 0.05 deg. have been observed.
about 150 psi. When the test chamber is maintained
at its maximum pressure (about 2 atm), the mini- 5. Discussion
mum relative humidity is about 20 percent. At
atmospheric pressure, a relative humidity as low This equipment has been used successfully for
as 10 percent is readily produced, and at a reduced calibration testing and research. The working space
pressure of % atm, the minimum relative humidity of 1 ft3 is ample for most instruments, materials,
decreases to 5 percent. and devices that have to be completely immersed
The temperature range of the equipment extends in an atmosphere of known relative humidity. There
from —40° to + 4 0 ° C. The upper end is limited is no theoretical limitation on the size of the test
by the flash point of the bath liquid (Stoddard chamber that may be employed with this type of
273
equipment; a larger sized chamber would simply easily be executed within 30 sec. At low rates of
require a larger surrounding liquid bath. Neither flow, the limiting factor ceases to be the time re-
is there any limitation on the geometry of the test quired for performing the above mechanical opera-
chamber. A cylinder was chosen in this case for tions and becomes, instead, the time involved in
ease of construction, but any other space configura- purging air of one relative humidity, with air of
tion can be substituted. Even in the present design, another relative humidity. The component with
the cylindrical test chamber can be uncoupled from the maximum air volume is the test chamber. II
the setup, and, within the limitation of the available has a space of about 1 ft3, so that the purging time
bath space, any other size, shape, or design of depends upon the rate of air flow through this 1-ft3
chamber can be attached. volume.
Occasionally air of known or preestablished dew 6. Summary
point is required at a place or instrument remote
from the test chamber. The desired dew point An apparatus of versatility and convenience for
can be readily produced by the apparatus, and all producing atmospheres of known relative humidity
or part of the air from the test chamber can be has been developed and constructed at the Bureau.
piped wherever needed. The only limiting factor I t operates on the "two-pressure principle/' whereby
involves the temperature of the ambient air, which air is saturated at a high pressure and expanded to a
must not drop below the dew point of the air flowing lower pressure, the relative humidity at the lower
through the transmission tubing, because condensa- pressure being the ratio of the lower to higher pres-
tion may occur in the lines. sure, provided the operation is performed at con-
The range of relative humidities obtainable with stant temperature.
this type of equipment may be extended to much Important parameters can be varied and controlled
lower values by using a higher pressure source. A over wide ranges: relative humidity from 10 to 98
250-psig compressor, operating between 150 to 200 percent; temperature from —40° to + 4 0 ° C; flow
psig, is used in the present design and provides up to 150 liters/min; test-chamber pressure from %
relative humidities that are sufficiently low for most to 2 atm. The relative humidity can easily be
purposes. Much higher pressures would necessitate changed from one value to another within 30 sec.
components capable of withstanding those high Independent checks on the accuracy of the rela-
pressures. Similarly, for flows in excess of 5 ft 3 /min, tive-humidity production with the psychrometric
a compressor having a larger volume capacity would and dew-point methods have yielded average agree-
be required. ments of ± 0 . 4 to ± 0 . 6 percent. As the latter
The rapidity with which the relative humidity methods of measurement are probably no more
may be changed depends primarily on the time in- accurate than about ± 0 . 5 percent, it is reasonable
volved in adjusting the pressure regulator, which to assume that the apparatus produces relative
controls the saturation pressure. Minor adjust- humidities that are known to at least ± 0.5-percent
ments of the expansion valve and the exhaust con- relative humidity.
trol valve may be required after the major pressure
adjustment has been made. These operations can WASHINGTON, December 12, 1951

274