ADMINISTRATIVE USE ONLY
NATIONAL WATER QUALITY NETWORK
APPLICATIONS AND DEVELOPMENT REPORT
#6
Evaluation
of the Field Performance
of a Hays Dissolved Oxygen Analyzer
John D. Weeks
X
Water Quality Section
Basic Data Branch
Division of Water Supply and Pollution Control
U. S. Public Health Service
1014 Broadway
Cincinnati 2, Ohio
October, 1962

-------
TABLE OF CONTENTS
PAGE
S VI lull l3 ITy ...00.00000000.00. ...a.oo.o..	^
Introduction..............			2
Location of Test			3
Description of Analyzer..... .		4
G 6 X16 IT Si 1 ooooooooaooo ooo.o....oooo..	^
Theoretical...............			5
Physical..........................	6
Component s . 			7
Amplifiers			7
Control Valve...................	7
Filter				7
Gas Control Valves 			8
Analyzer Cellooo.oooooooooooa.oo	8
Level Control...................	8
Flood Control				10
Test Procedure......................	11
Results and Discussion of Results...	13
Oxygen Concentration			13
Percent Saturation..............	15
Thermometry				17
Gas Consumption. 			17
Required Purge Time.............	18
Bibliography........................	19

-------
1.
2.
3.
4.
5.
FIGURES
PAGE
Schematic diagram of Hays Dissolved Oxygen Analyzer..4(a)
13(c)
15(c)
Comparison between manifold dissolved oxygen con
centration and recorded concentration.........
Comparison between manifold percent saturation
and recorded percent saturation
aoooooooeeoetfooooov
Percent saturation differences as a function of
manifold percent saturation
bceoooooeoooo
e o o o « o
Required purge time following standardization.„.
15(d)
18(b)
TABLES
1. Chronological sequence of dissolved oxygen
concentration, submersible pump. .......... * .... .... .13(a)
2
3,
4,
5.
6.
7.
8,
Chronological sequence of dissolved oxygen con-
centration, jet pump0„.
ooooooooooovoooo
0*000000
Analysis of deviation, absolute values, Hays-
Manifold comparison,
'oooatrootiooo^oooooooe
Analysis of deviations, algebraic values, Hays-
ManifoId comparison,
O0?0OOOOOO«0OOO
Computation of standard deviation of algebraic
values, Hays-Manifold comparison,
'ooeeooooooeeooo
oooeoeoeooooo
Solubility of oxygen in fresh water
Percent saturation tabulation and computation...
Chronological sequence of observed temperatures,
bo th pumps o.ooo.oooeooooo.aoo..o0ooco9os..oofio
Computation of absolute and algebraic temperature
deviations-, Hays-Manifold comparison
e • o • o «
10. Required purge time,
oooocopoeooooooooeoooo
13(b)
.13(d)
.13(e)
.14(a)
.15(a)
.15(b)
.17(a)
.17(b)
.18(a)

-------
- 1 -
SUMMARY
The Hays Dissolved Oxygen Analyzer was operated at
the Little Miami River Field Test Station of the National
Water Quality Network from July 17, 1962, to September
27, 1962. This evaluation was undertaken in order to
determine if the instrument is suitable for utilization
in the National Water Ouality Network's field activities.
The instrument yielded oxygen concentration values
which in average deviated by 0.15 ppm from the oxygen
concentration of the influent water. Percent saturation
errors varied with the saturation value of the sample
and indicate the necessity for a recalibration of the
percent saturation circuit. The instrument was capable
of recording temperature with an average deviation of
0.19°(C).
Throughout the period of this evaluation, no lost
time may be charged against the Hays Analyzer. However,
some time was lost due to failure of auxiliary equipment.
In the opinion of the writer the Hays Analyzer is
suitable for field application. Its application is
limited, however, by initial and operating costs, weight,
size, and required maintenance by high level personnel.
The analyzer is not compatible with the integrated sys-
tems approach now being developed by the National Water
Quality Network in cooperation with several instrument
manufacturers.

-------
- 2 -
INTRODUCTION
The National Water Ouality Network is interested in
the testing and application of instrumentation to the
measurement and recording of water quality parameters.
A parameter of major importance is dissolved oxygen.
There are three methods of instrumentally measuring
dissolved oxygen available at this time. These involve
(a) polarographic, (b) conductometric, and (c) partial
pressure techniques.
The polarographic method was tested as a portion
of another study and will be reported on separately.
The conductometric method with its requirement of
deionized water is limited to either waters of extremely
low mineral content or short term studies. For these
reasons, this method is not considered applicable to
long-term continuous sampling of surface waters.
The objective of this study was to evaluate the
field performance of a dissolved oxygen analyzer operating
on the partial pressure principle and to investigate the
factors which affect its performance.
The analyzer evaluated was a production model manu-
factured by the Hays Corporation of Michigan City,
Indiana. The theoretical principles, and the method-
ology underlying the dissolved oxygen analyzer were
originally developed at the Robert A. Taft Sanitary
Engineering Center. 1/ 2/
1/ Levine, H. S. , and Walker, W. W„ American Chemical
Society, 1955.
2/ Levine, H. S., and Kleinschmidt, R. S. Public Health
Service, DHEW.

-------
- 3 -
LOCATION OF TEST
This study was conducted in the National Water Ouality
Network's Field Test Station on the Little Miami River at
Cincinnati, Ohio.
This station is equipped with both a jet and a sub-
mersible pump. The intake to ejector piping on the jet
pump was 1-1/4 inch diameter and 1-1/2 inch diameter from
ejector to pump. The return line from pump to ejector was
1-1/4 inch in diameter. The submersible pump was installed
with a one-inch diameter pipe line. Polyethylene pipe was
used in both systems. The total length of each of these
lines was 400 feet. The lift from river to pump manifold
is 44 feet.
A rudimentary laboratory was set up at the station
so that bench analyses of dissolved oxygen could be made
quickly as required in the field by the aside-modified
Winkler method. Reagents and standard solutions were
provided by the National Water Quality Network's Service
Laboratory.

-------
- 4 -
DESCRIPTION OF ANALYZER
GENERAL: The Hays Dissolved Oxygen Analyzer records the
temperature of the water and dissolved oxygen content in
either parts per million or percent saturation. The
temperature is detected by a thermistor in the analyzer's
incoming water line.
Conversion of oxygen content in ppm to percent satura-
tion is accomplished by a thermistor located in the cyclone
separator, Figure 1. This thermistor corrects the oxygen
content for the temperature of the water. The selection
of operating mode (ppm or percent saturation) is controlled
by a switch on the front of the recorder case.
The Hays Analyzer uses a chart calibrated from 0-15
units with 0.3 unit being the width of the finest division.
The range for parts per million concentration is 0 to 15.
Temperature range is 0-30°(C). This requires that the re-
corded temperature values be multiplied by two. The range
of percent saturation is 0-1507o9 Thus, the chart reading
must be multiplied by 10.
This chart scale calibration limits the readability
of parts per million of dissolved oxygen to 0.1, satura-
tion to 17o, and temperature to 0.2°(C). Readings to a
greater degree of precision are limited by the width of
the pen trace.
Red ink was used in the oxygen recording pen and
green in the temperature recording pen.

-------
FIGURE I SCHEMATIC DIAGRAM OF HAYS DISSOLVED OXYGEN ANALYZER
&
ANALYZER RECORDER
MIXING
GAS
CYLINDER
GAS PRESSURE
GAS
CONTROL
VALVE
_l_
THERMISTOR ""
S_r-J
>TO RECORDER
I
PRESSURE
FILTER REGULATOR
TO L
MANIFOLD
FLOW CUTOFF
VALVE
k/1
i	r
j—L
PRESSURE
REGULATOR
ASPIRATOR
i	r
DRAIN
03
CHAMBER PRESSUR5
CALIBRATION FLOOD
VALVE CONTROL
_r~i_r
~l_j~i	r
TO GAS CONTROL
11
T=ai!r

WATER
LEVEL
.CONTROL
ELECTRODE
CYCLONE
SEPARATOR
DRAIN
-TO FLOOD CONTROL

-------
- 5 -
THEORETICAL: The Hays Dissolved Oxygen Analyzer operates
on Henry's Law which states "the quantity of gas which
dissolves in a solvent at a given temperature is directly
proportional to the partial pressure of that gas,"
Henry's Law is limited to those cases where the gas does
not react chemically with the solvent. The oxygen-water
system obeys Henry's Law,
In the Hays Analyzer nitrogen gas is injected into
the water sample, mixed, and then the oxygen-nitrogen
gas mixture is stripped from the water solution. The
gas mixture is fed into the analyzer cell and then is re-
cycled back to the aspirator unit. The gas system com-
prises a closed loop which insures that equilibrium is
reached between the dissolved oxygen in the water and
the oxygen gas in the loop. Since this loop has a vol-
ume, there is a definite time required for the system
to respond to a change in the dissolved oxygen content
of the influent sample. The time required for the sys-
tem to reach 90% of a sudden change is termed the res-
ponse time and for the instrument used was about 3.5
seconds.
The schematic diagram of the Hays Analyzer has
been shown in Figure 1. The performance of each sub-unit
will be discussed under operational description.

-------
- 6 -
PHYSICAL: The analyzer consists of two major units.
These are the recorder and its associated amplifiers
and the magnetic analyzer unit. On the back of the
magnetic analyzer unit are mounted the gas control sys-
tem, water control valves, cyclone separator and inci-
dental piping and tubing. The recorder weighs about 60
pounds and the magnetic analyzer unit weighs about 100
pounds. These sub-assemblies are mounted in a house
which measures 2.25 feet wide, 4.89 feet high, and 3.10
feet deep. The overall volume is 33,7 cubic feet. In-
side the house are the electrical distribution lines,
sample pipes, and pressure gages. Space was provided
for installing the nitrogen gas cylinder inside the
house. The housing used in this study was deeper than
required since space was available for the installation
of an automatic data logging device. This device was
not used in this study.
The overall weight of this unit complete with a
large gas cylinder is estimated at 400 pounds.
A floor area of 4.0 feet by 5.5 feet should be
provided for this instrument to allow the doors to be
fully opened and for water and drain connections to be
made. Future designs of housing should place some of
the control valves and gages at the front of the instru-
ment to permit ready control of the system variables.

-------
- 7 -
COMPONENTS : The Hay's Dissolved Oxygen Analyzer is a
complex of many components and systems. For example,
there are electronic amplifiers, motors, fluid pres-
sure regulators and gas control valves. There are
hydraulic, gas and electrical systems within the an-
alyzer. The following discussion will be limited to
those units which have a bearing upon the operation
and necessary theory.
Amplifiers: Two identical amplifiers were used
in the Hays Analyzer, One was in the temperature chan-
nel and the other was in the oxygen channel. These
amplifiers had a gain control which was found to be
quite sensitive. At one setting the gain was insuf-
ficient and with just a little more gain an over-
sensitive condition prevailed. When such a system is
in an over-sensitive condition, it is subject to "hunt-
ing" which is a violent searching for the correct re-
balance point. Once the optimum gain was obtained,
however, there was little or no drift in the setting.
Control Valve: This valve controls the flow
rate through the analyzer. It can be fully opened to
permit an uninhibited discharge of water. When it is
closed, water flows through a 1/8-inch diameter hole
in the valve gate. In operation, this valve is fully
closed and it is opened only for standardizing the
analyzer.
Filter: Before the water enters the analyzer, it
passes through a Cuno edge-type filter which has brass
plates,spaced to 0.020 inches. This will take out only
large particles and leave the smaller particles to pass
through. No flow stoppages occurred during this study
as a result of particulate.matter clogging the water
lines, valves, or filter.

-------
- 8 -
Gas Control Valves: This unit, which is actuated
by the level control electrode consists of two solenoid
valves in series. No operational difficulty was trace-
able to its operation.
Analyzer Cell: This component operates on the
paramagnetic properties of oxygen. It consists of two
chambers in which are mounted two matched thermistors.
One.chamber is under a very intense magnetic field.
The chambers and thermistors are heated. In operation,
a mixture of nitrogen and oxygen enters the analyzer,
and oxygen is attracted to the chamber which lies in the
magnetic field. The oxygen becomes heated, loses its
paramagnetism, and is swept out of the cell by incoming
cooler paramagnetic oxygen. Since nitrogen is non-
magnetic, equal flow rates of this gas exist in each
chamber. Thus the thermistor in the magnetic field is
preferentially cooled by an amount proportional to the
oxygen content of the sample. Its resistance changes
are measured by a self-balancing Wheatstone bridge cir-
cuit and applied through an amplifier to a chart.
Level Control: The water level in the cyclone
separator is maintained fairly constant by use of con-
trolled gas flow. When the water level in the separa-
tor rises, contact is established between the separator
case and the level control electrode. This signal con-
trols the solenoid valves which permit nitrogen gas to
enter the system. The incoming gas forces down the
level in the separator until contact is broken at which
time the gas flow ceases.
The solenoid valve should be actuated from two to
six times per minute in normal operation. The frequency
of actuation is a function of the differential of the
incoming gas pressure over the chamber pressure. The

-------
- 9 -
normal operating differential is 1/2 to 1 psig. In this
evaluation study, the chamber pressure was maintained
at 3.0 psig and the nitrogen gas pressure was 3.5 psig.
As the pressure differential is increased, the frequency
of actuation is decreased. This causes the recorded
valves to "cycle" or wander, back and forth over the
correct value. The amplitude of this wandering has
approached ^ 0.15 ppm at 12 ppm. Reducing the gas pres^
sure to allow more frequent actuation reduces the cycling.
There is no one optimum setting which will produce cycle
free recording over a wide range of dissolved oxygen
concentrations. This is particularly noticeable when
the saturation exceeds one-hundred percent.
Initially, a single stage regulator was used to
drop the gas pressure from about 2000 psig in the tank
to 3.5 psig in the analyzer. This was, a very critical
setting and changed as the tank temperature changed.
Later, a two stage regulator was installed, and found
capable of maintaining a constant output pressure with
changes in temperature.
The dissolved oxygen analyzer as delivered had a
three-inch long level control electrode. Knight (3/)
has reported frequent flooding with this length of
electrode. Throughout the period of this study a five-
inch electrode was used and no routine flooding was
experienced.
3/ Knight, W, E., personal communication to M. B.
Ettinger, 1962.

-------
- 10 -
Flood Control: Under certain conditions, it is
possible for water to be forced out of the top of the
cyclone separator and to enter the oxygen-nitrogen gas
loop where it could damage the thermistors in the analy-
zer cell. In order to prevent such an occurrence, a
flood control device has been installed on the analyzer.
This device consists of a conductivity cell installed
in the gas loop. When water reaches this cell, a relay
system closes a solenoid valve which cuts off the flow
of water into the instrument.
The conditions under which flooding can occur are
(a) an obstruction in the analyzer water drain; (b) an
empty gas cylinder; (c) an obstruction in the gas input
line (d) a dirty or corroded water level electrode.
Two flooding events occurred during this study.
One was caused by a leaking pressure regulator on the
nitrogen tank which exhausted the gas supply and the
other event was traced to a blocked nitrogen gas line.
The line was blocked by dirt introduced by the change-
over from one nitrogen tank to another. It is felt that
these lost-time periods are chargeable to the auxiliary
equipment and not to any design or construction detail
of the Hays Analyzer.

-------
- 11 -
TEST PROCEDURE
Before this study began, all electronic components
were tested to insure peak performance. The instrument
received a complete inspection for any worn items which
may have required replacement. The lubrication procedure
was followed in order to place the mechanical systems in
the best operating condition.
A warm-up period of 24 hours was allowed before
water was introduced into the instrument. A warm-up of
four hours is required to stabilize the temperature in
the analyzer cell compartment. This extended warm-up
was used as a short-term life test of the electronic
equipment. After this warm-up period, the temperature
channel was adjusted to record the temperature of the
manifold which was determined by a conventional partial
immersion laboratory thermometer inserted into the
manifold by-pass.
Following the adjustment of the temperature chan-
nel, the water pressure relief valve on the manifold was
adjusted until a pressure of 50 psig was maintained at
the analyzer input. Next, the gas pressure regulator
was adjusted until 3.5 psi was indicated on the gas feed
line gage. The water line pressure regulators were then
adjusted until the chamber pressure gage read three
pounds per square inch.
The instrument was then standardized and zeroed by
procedures specified in the maintenance manual. Water
samples were collected at the manifold and analyzed
using the Winkler method as described in the National
Water Quality Network's Operating Manual. The analytical
results were compared with the recorded values. The
manifold concentration is based upon a single determina-

-------
- 12 -
tion rather than the average of repetitive determinations.
At each sampling, temperature, concentration and percent
saturation were read from the Hays Analyzer. Temperature
and concentration were determined at the manifold and
theoretical percent saturation computed.
The instrument was operated for ten days on the
submersible pump and for ten days on the jet pump. These
periods do not include weekends, holidays and other
periods when the instrument was in operation but no samp-
ling was in progress. The instrument was in an ''on!'
condition from July 17 to September 27, 1962, and no
failure of its operating components occurred.

-------
- 13 -
RESULTS AND DISCUSSION OF RESULTS
The Hays Analyzer was tested for its ability to
record dissolved oxygen concentration, oxygen percent
saturation, and temperature, accurately. Pipe and
pumps may have an effect upon the dissolved oxygen.
For this reason, the Hays Analyzer results will be
compared against the results obtained from samples
collected at the manifold rather than collected in
the river.
Oxveen Concentration: The Hays Analyzer was op-
erated principally as an oxygen concentration recorder.
This mode was selected because prior knowledge indi-
cated that the Little Miami River was quite likely to
exceed 1507o saturation but not likely to exceed a 15
ppm concentration,, Secondly, the analytical method
determined oxygen concentration and not saturation.
During this study, eighty-one comparisons were made
between the manifold concentration and the recorded re-
sult,. These data are tabulated in Tables 1 and 2 and
are presented graphically in Figure 2. Overall, the
Hays Analyzer had a mean absolute deviation of 0.15
ppm, a mean absolute deviation of 0.13 ppm on water
supplied by the submersible pump and a mean absolute
deviation of 0.16 ppm on water delivered by the jet
pump. (Table 3)„ The oxygen concentration delivered
to the Hays ranged from 1/4 ppm to 11.8 ppm with the
jet pump and from 5„2 ppm to .13.1 ppm with the submer-
sible pump. The deviations listed in Tables 1 and 2
were further refined by computing the algebraic mean
deviation. (Table 4 ). If- the deviations were randomly
distributed, a value of zero for the algebraic deviation

-------
TABLE #1
Hays-Manifold Comparison Data
Chronological sequence of dissolved oxygen concentrations
Submersible Pump
Date
1962
Manifold
ppm
Hays
ppm
Deviation
ppm
Date
1962
Manifold
ppm
Hays
ppm
Deviation
ppm
7/26
9 . 2
9.3
-0.1
8/16
10.6
10.5
+0.1
7/27
7.4
7.4
0.0
8/16
11.5
11.3
+0.2
7/27
8 „ 0
8.0
0.0
8/17
8.0
7.9
+0.1
7/27
8.9
8.8
+0.1
8/17
9.1
9.0
+0.1
7/27
7/27
10,5
10.5
0.1
8/17
10.1
10.1
0.0
11.0
11.0
0.0
8/17
11.4
11.3
+0.1
7/28
12.0
11.8
+0.2
8/17
11.7
11.7
0.0
7/28
12.7
12.4
+0.3
8/17
12.0
12.0
0.0
7/28
13.1
12.5
+0.6
8/17
12.3
12.2
+0.1
8/10
10.3
10.5
-0.2



8/13
6.2
6.4
-0.2




8/13
6.5
7.1
-0.6




8/13
7.2
7.7
-0.5




8/.13
9.5
9.4
+0.1




8/14
6.4
6.5
-0.1




8/14
7.0
6.9
+0.1




8/14
8.0
7.9
+0.1




8/14
8.4
8.5
-0.1




8/14
8.9
8.8
+0.1




8/15
7-1
7.0
+0.1




8/15
8.-4
8.5
-0.1




8/15
9.1
9.5
-0.4




8/15
10.0
10.0
0.0




8/15
10.4
10.4
0.0




8/16
7.4
7.4
0.0




8/16
8.5
8.4
+0.1




8/16
9.2
9.2
0.0





-------
TABLE #2
Hays-Manifold Comparison Data
Chronological sequence of dissolved oxygen concentration
Jet Pump
Date
1962
Manifold
ppm
Hays
ppm
Deviation
ppm
Date
1962
Manifold
ppm
Hays
ppm
Deviation
ppm
8/27
9.5
9.4
+0.1
9/17
6.6
6.6
0.0
8/27
10.9
10.8
+0.1
7/17
7.8
8.0
-0.2
8/28
6.7
6.6
+0.1
9/18
5.5
5.4
+0.1
8/28
7.4
7.4
0.0
9/18
6.1
6.0
+0.1
8/28
9.0
8.9
+0.1
9/18
7.4
7.5
-0.1
8/28
11.2
11.0
+0.2
9/18
8.2
8.2
0.0
8/28
11 o 7
110 5
+0.2
9/18
9.1
9.1
0.0
8/29
9.3
9.1
+0.2
9/18
10.1
10.1
+0-1
8/29
10.0
9.9
+0.1
9/19
6.9
6.8
+0.1
8/29
11.4
11.1
+0.3
9/19
6.7
6.8
-0.1
Si/ 29
11.6
11.3
+0.3
9/20
6.6
6.6
0.0
a/so
7.2
7.0
+0.2
9/20
6.3
6.1
+0.2
8/30
7.8
7.5
+0.3
9/20
8.1
7.9
+0.2
8/30
8.4
8.1
+0.3
9/20
10.3
10/6
-0.3
8/30
8.9
8.8
+0.1
9/20
9.7
9.6
+0.1
9/30
11.1
10.7
+0.4
9/21
7.0
6.9
+0.1
8/31
6.3
6.0
+0.3
9/31
8.0
8.3
-0.3
8/31
6.2
6.1
+0 f 1
9/21
8.4
8.5
-0.1
8/31
6.7
6.7
0.0



8/31
1.4
1.1
+0.3




8/31
5.1
4.8
+0.3




9/13
1.3
1.0
+0.3




9/13
5.4
5.2
+0.2




9/13
7.4
7.4
0.0




9/13
10.7
11.1
-0.4




9>/13
11.8
11.8
0.0




9/17
5.10
5.1
0.0





-------
13(c)

-------
TABLE #3
13(d)
Hays-Manifold Comparison
Analysis of deviations, absolute values
Deviation (x) !
Jet
Su
bs.
Total
f
fx
f
fx
f
fx
0.0 ppm
9
0.0
11
0.0
20
0.0
0.1
16
1.6
16
1.6
32
3.2
0.2
8
1.6
4
0.8
12
2.4
0.3
10
3.0
1
0.3
11
3.3
0.4
2
0.8
1
0.4
3
1.2
0.5
0
0.0
1
0.5
1
0.5
0.6
0
0
2
1.2
2
1.2
Totals
45
7.0
36
4.8
81
11.8
1
Mean deviation,
ppm
0.
156
0.
I"	
i
i CO
1 ^
0.
146

-------
TABLE #4
Hays-Manifold Comparison
Analysis of deviations, algebraic values
Deviations
ppm (x)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
Jet
Submersible
Total
Negative
Positive
Negative
Positive
Negative
Positive
f
fx
f fx
f
fx
f fx
f
fx
f fx
0
3
1
2
1
0.0
0.3
0.2
0.6
0.4
9 0.0
13 1.3
7	1.4
8	2.4
1 0.4
0
4
2.
0
1
1
1
0
0.4
0.4
0
0.4
0.5
0.6
11	0
12	1.2
2 0.4
1 0.3
0 0.0
0	0.0
1	0.6
0
7
3
2
2
1
1
0.0
0.7
0.6
0.6
0.8
0.5
0.6
20 0.0
25 2.5 !
9 1.8
9 2.7
1 0.4
0	0.0
1	0.6
Totals
7
1.5
38 5.5
9
2.3
27 2.5
16 j 3.8
65 8.0
Residual fx
Total
Deviation, ppm
+4.0
45
+0.089
+0.2
36
+0.006
+4.2 |
81 !
+0.052

-------
- 14 -
would indicate equal weighting above and below the regres-
sion linec The algebraic mean deviations are +0.052 ppm,
+0.006 and +0.089 ppm for the total population, submer-
sible and jet pump waters, respectively. From the sign
convention used, a positive value indicates that the mani-
fold concentration is higher than the recorded concentra-
tion. The small value of the algebraic mean shows that
the Hays Analyzer is nearly free of a bias in its record-
ing. If a bias exists then the Hays is under-recording
the correct concentration by a small amount, by an aver-
age algebraic deviation of 0.052 ppm for a large number
of samples.
The deviations between the Hays Analyzer results
and the results determined at the manifold ranged from
+0.6 to -0.6 ppm with a mean absolute deviation of 0.15
ppm. The standard deviation of the deviations is + 0.19
ppm, Table 5, This standard deviation is strongly af^
fected by the extreme values. There is evidence presen-
ted in the original data sufficient to reject the fol-
lowing deviations: +0.6, -0.6, -0.5, and -0.4 ppm.
If these values are dropped from consideration, the mean
absolute deviation, total population, computes to 0.13
ppm. Retaining the questionable data places a more
severe test upon the Hays Analyzer. For this reason
these data will be retained in this study.
Applying statistical theory to the values computed
above, the Hays Analyzer should produce data with an
average deviation of 0.15 ppm from the concentration of
water supplied. In the future 68.37* of the observations
should be within a range of deviat±rrrrs o-f -0.14 to +0.24
ppm; 95.4% would be within a range of -0.33 to +0.43 ppm,
and 99.1% of the data would be within the range -0.52 to
+0.62 ppm.

-------
14(a)
TABLE #5
Total Hays-Manifold Comparisons
Computation of standard deviation
Hay s-Hani f oId
Deviation
PPm
(d)
Frequency
(f)
fd
Deviation
From Mean
(x)
x2
fx2 1
j
¦
-0.6
1
-0.6
-0.652
0.425,104
0.425,104 !
-0.5
1
-0,5
-0.552
0.304,704
0.304,704
-0.4
2
-0.8
-0.452
0.204,304
0.408,608
-0.3
2
-0.6
-0.352
0.123,904
0.247,808
-0.2
3
-0.6
-0.252
0.063,504
0.190,512
-0.1
7
-0.7
-0.152
0.023,104
0.161,728
0.0
20
0.0
-0.052
0.002,704
0.054,080
+0.1
25
+2.5
+0.048
0.002,304
0.057,600
+0.2
9
+1.8
+0.148
0.021,904
0.197,136
+0.3
9
+2.7
+0.248
0.061,504
0.553,536
+0.4
1
+0.4
+0.348
0.121,104
0.121,104
+0.5
0
0.0
+0.448
0.200,704
0.000,000
+0.6
1
+0.6
+0.548
0.300,304
0.300,304
Totals
81
+4.2


3.022,224
d mean = 4.2 = 0.052 ppm
81
x = d - d mean
£T= / Zf(x)2 = /3.022.224 = ^/o377778
^ n-1	80
= 1 0.19 ppm

-------
- 15 -
Percent Saturation: The Hays Dissolved Oxygen An-
alyzer is capable of presenting the dissolved oxygen
content in terms of percent saturation. The conversion
from parts per million to percent saturation is made by
a thermistor located in the body of the cyclone separa-
tor. This thermistor corrects the oxygen concentration
for the solubility of oxygen at the sample temperature.
In this study, 65 comparisons of manifold and rec-
ords data were obtained. From these data, a percent
saturation values was computed using the solubility
figures in Steel. 4J This data is repeated for con-
venience in Table 6. The observed and computed values
are listed in Table 7.
The computed values of saturation ranged from 16
to 149 percent; while the recorded values ranged from 9
to 149 percent. There were 10 events of the recorder
reading off-scale and these are not included in the
above 65 comparisons. From the plotting in Figure 3,
there appears to be a drift in the data, particularly at
the higher values of percent saturation. The differences
in percent saturation, Column 7, Table 7, were then
plotted against manifold theoretical percent saturation,
in Figure 4. This figure shows that the differences are
lowest in the vicinity of 70-percent saturation and high-
est near 125-percent saturation. Since the Hays Analyzer
has proven capable of yielding accurate concentration
and temperature values, it must be concluded that these
departures are inherent to the percent saturation ther-
mistor or its circuitry. Other workers have reported
4/ Steel, E. W. Water Supply and Sewerage.

-------
15(a)
TABLE #6
Solubility of Oxygen
In Fresh Water

Dissolved

Dissolved
Temperature
Oxygen
Temperature
Oxygen
°C
ppm
°C
ppm
0
14.62
16
9.95
1
14.23
17
9.74
2
13.84
18
9.54
3
13.48
19
9.35
4
13.13
20
9.17
5
12.80
21
8.99
6
12.48
22
8.83
7
12.17
23
8.68
8
11.87
24
8.53
9
11.59
25
8.38
10
11.33
26
8.22
11
11.08
27
8.07
12
10.83
28
7.92 !
13
10.60
29
7.77
14
10.37
30
7.63
15
10.15

1
Source ¦:
Steel, Ernest W„, "Water Supply and
Sewerage", Second Edition, 1947.
McGraw Hill, New York, 1947, P0 483.

-------
TABLE #7
Percent Saturation Tabulation and Computation

Manifold
Manifold
Theoretical
Theoretical
Hays
Difference
Date
Dissolved
Temperature
Mani-f o Id
Manifold
Percent
Mani-f old PS.
1962
Oxygen
°C
Oxygen Content
Percent
Saturation
- Hays PS

ppm

at Saturation
ppm
Saturation

7/26
9.3
24.6
8.44
110
119
-9
7/27
7.4
21.0
8.99
82
87
-5
7/27
8.0
22.6
8.74
92
97
-5
7/27
10.5
25.0
8.38
125
144
-19
8/13
6.5
25.2
8.35
78
89
-11
8/13
7.2
25.4
8.32
86
97
-11
8/13
9.5
26.8
8.10
117
125
-8
8/14
6.4 .
23.0
8.68
74
77
-3
8/14
7.0
22.9
8.70
80
84
-4
8/14
8.0
23.1
8.66
92
95
-3
8/14
8.4
23.0
8.68
97
103
-6
8/14
8.9
23.0
8.68
102
105
-3
8/15
7.1
21.2
8.96
79
81
-2
8/15
8.4
24.Q
8.53
98
105
-7
8/15
9.1
25.1
8.36
109
122
-13
8/15
10.0
25.3
8.33
120
132
-12
8/15
10.4
25.8
8.25
126
145
-19
8/16
7.4
22.8
8.71
85
89
-4
8/16
8.4
25.3
8.33
101
108
-7
8/16
9.2
26.0
8.22
112
120
-8
8/16
10.6
26.2
8.19
129
145
-16
8/16
11.5
26.3
8.18
141
off scale

8/17:
8.0
24.2
8.50
94
98
-4
8/17
8/17
9.1
25.2
8.35
109
116
-7
10.1
26.2
8.19
123
136
-13
8/.17
11.4
26.4
8.16
140
off scale

8/17
11,7
26.2
8.19
143
off scale

8/17
12.0
26.2
8,19
147
off scale

8/17
12.2
26.2
8.18
149
off scale

8/27
9.5
25.4
8.32
114
123
-9

-------
TABLE #7 (Contd)
Percent Saturation Tabulation and Computation

Manifold
Manifold
Theoretical
Theoretica1
Hays
Difference
Date
Dissolved
Temperature
Manifold
Manifold
Percent
Manifold PS
1962
Oxygen
°C
Oxygen Content
Percent
Saturation
- Hays PS

ppm

at Saturation
ppm
Saturation

8/27
10o9
26.0
8.22
133
148
-15
8/28
6.7
24.6
8.44
79
82
-3
8/28
7.4
25.0
8.38
88
92
-4
8/28
9.0
25.2
8.35
108
114
-6
8/28
11.2
26.2
8.19
137
off scale

8/28
11.7
26.5
8.14
144
off scale

8/29
9.3
26.2
8.19
115
121
-6
8/29
10,0
26.8
8.10
123
134
-11
8/29
11,4
27 .0
8.07
141
off scale

8/ 29
11,6
27.0
8.07
144
off scale

8/30
7 o 2
25.4
8.-32
85
89
-4
8/30
7.8
26.0
8.22
95
97
-2
8/30
8,4
26.1
8.20
102
104
-2
8/30
8.9
27,0
8.07
110
117
-7
8/30
11.1
27.4
8.01
139
149
-10
8/61
6.3
25.8
8.25
76
75
+ 1
8/31
fr.2
26.0
8.22
75
76
-1
8/31
6.7
25,8
8.25
81
88
-7
8/31
1.4
26.0
8.22
17
14
+3
8/31
5.1
26.2
8.19
62
60
+2
9/13
1.3
25.2
8.35
16
9
+7
9/13
5 „ 4
24.4
8.47
64
61
+3
9/13
7.4
24.4
8.47
87
90
-3
9/13
10 0 7
24.8
-8.41
128
147
-19
9/13
11.8
25.2
8.35
141
off scale

9/17
5,1
23.8
8.56
60
62
-2
9/17
6 . 6
24 01
1 8.52
78
80
-2
•

-------
TABLE #7 (Contd)
Percent Saturation Tabulation and Computation
Date
1962
Manifold
Dissolved
Oxygen
ppm
Manifold
Temperature
°C
Theoretical
Manifold
Oxygen Content
at Saturation
ppm
Theoretical
Manifold
Percent
Saturation
Hays
Percent
Saturation
Difference
Manifold PS
- Hays PS
9/17
7.8
24.6
8.44
92
100
-8
9/18
5.5
22.7
8.72
63
64
-1
9/18
6.1
23.0
8.68
70
71
-1
9/18
7.4
23.0
8.68
85
91
-6
9/18
8 o 2
23.4
8.62
95
101
-6
9/18
9.1
23.8
8.56
106
114
-8
9/18
10.1
23.6
8.59
118
126
-8
9/19
6.9
22.0
8.83
78
80
-2
9/19
6.7
22.0
8.83
76
79
-3
9/20
6.6
20.5
9.08
73
75
-2
9/20
6.3
20.8
9.03
70
70
0
9/20
7.9
21.0
8.99
88
92
-4
9/20
10.3
21.0
8.99
115
126
-11
9/20
9.7
21.0
8.99
108
114
-6
9/21
6.9
20.0
9.17
75
79
-4
9/21
8.0
20.0
9.17
87
90
-3
9/21
8.4
19.8
9.20
91
96
-5
Ln
/	N
cr
n
o
3
rt
a

-------

-------
15(d)

-------
- 16 -
that the instrument is capable of giving values of per-
cent saturation accurate to + 2% over its range. 5/
To investigate the normal range of percent saturation
errors taken from a recorder, consider the following case:
Temperature values accurate to + 0.2°C.
Dissolved oxygen concentrations accurate to +0.15°C.
Saturated water at 25°C is being sampled.
The computations are shown in the following:

D.O. Error
-0.15 ppm
No D. 0.
Error
D.O. Error
+0.15 ppm
Error in temperature=
+0.2°C
8.23
8.36
25.2
98.4
8.38
8.36
25.2
100.2
8.53
8.36
25.2
102.0
No temperature
error
8.23
8.38
25.0
98.2
8.38
8.38
25.0
100.0
8.53
8.38
25.0
101.8
Error in temperature=
-0.2°C
8.23
8.40
24.8
98.0
8.38
8.40
24.8
99.8
8.53
8.40
24.8
101.6
i
LEGEND:
A
1 B
c
D
A = Concentration of D.O, in ppm,
B = Temperature in °C.
C = Saturation D.O. at temperature B.
D = Percent saturation of B & C.
From this square it can be shown that the worst error
incidental to normal operation is •+ 2„0%. An error in tem-
perature at 25°C of ± 0.2°C gives rise to errors in satura-
tion of + 0,2% for constant concentration„ If the tiempera-
ture is constant and the dissolved oxygen varies by + 0.15
ppm, an error of + 1.87o saturation-arises. The square shows
that the worst conditions arise when the errors are both plus
or both minus for temperature and concentration, respectively.
The erratic performance of this instrument should be further
investigated to see if a faulty adjustment or component ex-
ists somewhere in the circuitry.
5/ Levine, H. S., and Kleinschmidt, Principles and Problems
in Development of the SEC Dissolved Oxygen Analyzer.

-------
- 17 -
Thermometry: One parameter recorded by the Hays Dis-
solved Oxygen Analyzer is temperature. This is a necessary
measurement if dissolved oxygen concentrations are to be
converted to percent saturation.
A total of 68 observations of manifold and recorder
temperatures were made over a range of 19„8 to 28„0°(C)
at the manifold. These data are tabulated in Table 80
The recorder deviated +0.6 to -0o5°(C) from the manifold
reading, with a mean absolute deviation of 0.185°(C) and
an algebraic deviation of +0.065°(C), Table 9. It should
be stressed that no insulation existed on the copper pipe
which fed water from the manifold to the instrument.
The chart calibration and multiplication factor of two
would give rise to errors of +0.2°(C) for a +0.1°(C)
chart reading error.
Since the computed mean absolute deviation falls
within the anticipated deviation, it is concluded that
this system is performing satisfactorily.
Gas Consumption; In normal operation this instru-t
ment consumed enough nitrogen gas to cause the tank pres-
sure to fall by an average of 17 psi per day. This is an
inexact value since the tank pressure is affected by temp-
erature variations, and the test station temperature varied
over wide limits. If a drop of 17 psi per day is allowed,
a 220 cubic foot tank with an initial pressure of 2300 psi
should last 135 days or about 4-1/2 months. This agrees
with the manufacturer's statement that such a tank should
last several months. It must be recognized that the gas
consumption rate will increase with lower value of dis-
solved oxygen and for a different station may vary from
these figures obtained on the Little Miami River.

-------
TABLE #8
17(a)
Chronological. Sequence
of Observed Temperatures

Manifold
Recorded

Date
Temperature
Temperature
Deviation
1962
°€
®c
°C
7/20
25.0
24.6
+0.4
7/23
25.0
25.4
+0.6
7/24
23.0
22.6
+0.4
7/24
23.4
23.4
0.0
7/24
24.4
24.6
-0.2
7/25
26.5
26.8
-0.3
7/26
23.3
23.4
-0.1
7/26
24.8
24.8
0.0
7/27
21.0
20.8
+0.2
7/27
22.3
22.4
-0.1
7/28
25.8
25.8
0.0
8/10
27.9
28.0
-0.1
8/13
25.0
24.8
+0.2
8/14
23.0
22.8
+0.2
8/14
22.9
23.0
-0.1
8/14
23.1
23.0
+0.1
8/14
23.1
23.1
0.0
8/15
21.8
21.2
+0.6
8/15
24.0
23.8
+0.2
8/15
25.1
25.0
+0.1
8/15
25.3
25.8
-0.5
8/15
25.8
25.8
0.0
8/16
23.0
22.8
+0.2
8/16
25.3
25.2
+0.1
8/16
26.0
26.0
0.0
8/16
26.2
26.4
-0.2
8/16
26.3
26.4
-0.1
8/17
24. 2
24.0
+0.2
8/17
25.2
25.4
-0.2
8/17
26.2
26.4
-0.2
8/17
26.4
26.2
+0.2
8/17
26.2
26.4
-0.2
8/17
26.2
26.4
-0.2
8/17
26.3
26.4
-0.1
8/27
25.4
25.6
-0.2
8/27
26.0
26.2
-0.2
8/28
24.9
24.6
+0.3
8/28
25.0
24.6
+0.4
8/28
25.2
	1) M
25 „ 2
0.0

-------
17(a) Gontd.
TABLE #8 (Contd.)
Chronological Sequence
of Observed Temperatures
1
Manifold
Recorded

Date
Temperature
Temperature
Deviation
1962
°C
°C
°C
8/28
26.2
26.2
0.0
8/28
26.5
26 .4
+0.1
8/29
26.2
26.2
0.0
8/29
26.8
26.6
+0.2
8/29
27.0
26.8
+0.2
8/29
27.0
26.9
+0.1
8/30
25.4
25.4
0.0
8/30
25.8
26.0
-0.2
8/30
26.0
26.1
-0.1
8/30
27.0
27.0
0.0
8/31
26.0
25.8
+0.2
9/17
23.8
23.6
+0.2
9/17
24.1
23.6
+0.5
9/17
24.6
25.0
-0.4
9/18
22.7
22.4
+0.3
9/18
23.0
22.8
+0.2
9/18
23.0
22.8
+0.2
9/18
23.4
23.4
0.0
9/18
23.8
23.6
+0.2
9/18
23.6
23.6
0.0
9/19
22.0
22.2
-0.2
9/20
20.5
20.2
+0.3
9/20
20.8
20.6
+0.2
9/20
21.0
20.4
+0.6
9/20
21.0
21.0
0.0
9/20
21.0
21.0
0.0
9/21
20.0
20.2
-0.2
9/21
20.0
19.6
+0.4
9/21
1
19.8
19.6
+0.2

-------
17(b)
TABLE #9
Computation of Absolute and
Algebraic Temperature Deviations
Deviation
°C
Frequency
(f)
fd
d


-0.5
1
-0.5
-0.4
1
-0.4
-0.3
1
-0.3
-0.2
11
-2.2
-0.1
7
-0.7
0.0
15
0.0
+0.1
5
+0.5
+0.2
16
+3.2
+0.3
3
+0.9
+0.4
4
+1.6
+0.5
1
+0.5
+0.6
3
1.8

68

Absolute total = 12.6
Absolute mean deviation - 12.6.^68 = 0.185°C
Algebraic total = +4.4
Algebraic mean deviation = 404^68 =+0„065°C

-------
- 18 -
Required Purge Time: The standardization procedure
requires that atmospheric air be introduced into the an-
alyzer cells through the calibration valve. The instru-
ment is then adjusted to an oxygen saturation value
computed from temperature, barometric pressure and humid-
ity readings.
Opening the gas loop in this manner effectively con-
taminates the system. It was of interest to establish
the minimum time, following standardization, that reliable
dissolved oxygen concentrations could be read. This time
has been termed "purge time". It is a measure of the time
necessary for the oxygien in the gas loop to reach equili-
brium with the oxygen dissolved in the influent water.
On July 27, 1962, the instrument was standardized
and placed in a running condition. Readings were taken
at intervals following standardization and were plotted.
The tabulated data is in Table 10, and the plot is in
Figure 5. From the plot in Figure 5, it can be seen that
a time interval of one hour following standardization
must be allowed for purging.
The standardization procedure requires from 10 to
30 minutes depending upon the amount of adjustment re-
quired. A routine service visit would then require from
70 to 90 minutes to standardize the analyzer and to in-
sure that it was indicating the correct dissolved oxygen
concentration after purging.

-------
18(a)
TABLE #10
Required Puree Time Data
Time
7/27/62
Hays Dissolved
Oxygen
ppm
Remarks
1:45 pm
„
Manifold dissolved oxygen 9.2 ppm
3:20 pm
8.4
Standardization completed
3:25 pm
8.8
3:30 pm
8.7
Manifold dissolved oxygen 9.1 ppm
3:35 pm
8.6
3:50 pm
8.9

4:00 pm
8.8

4:05 pm
9.1

4:15 pm
9.3

4:27 pm
9.3

4:45 pm
9.3
Manifold dissolved oxygen 9.2 ppm

-------
18(b)
ii>o

-------
- 19 -
BIBLIOGRAPHY
Dixon. William S.--"Pollution Control By Continuous
Dissolved Oxygen Analysis and Associated Instruments",
P. 153. Proceedings of the 1960 Seminar- on Water Quality
Measurement and Instrumentation, U.S. Department of Health,
Education and Welfare 1961.
Levine. H. S „ . and Walker. W. W.--"The Continuous
Measurement and Recording of Dissolved Oxygen in Water,
Part 1, Theoretical Basis", presented at 127 American
Chemical Society Meeting, April 9, 1955.
Levine. H. S.. and Kleinschmidt--"Principles and
Problems in Development of the S.E...C. Dissolved Oxygen
Analyzer", U.S. Department of Health, Education and Wel-
fare .
Steel. Ernest W„--"Water Supply and Sewerage", Third
Edition, McGraw Hill, New York, N.Y., 1947.
Knight. W. E.--"Operation of a Hays D.0„ Analyzer",
June 27, 1962. Personal communication to Mr. M. B.
Ettinger.
Havs Corporation Maintenance Manual.--The Hays Cor-
poration, Michigan City, Indiana.

-------