YSI 58 - Discussion of Measurement Errors

SECTION 6 DISCUSSION OF MEASUREMENT ERRORS
There are three basic types of dissolved oxygen errors. Type 1 errors are related to limitations of
instrument design and tolerances of instrument components. These are primarily the meter linearity
and the resistor tolerances. Type 2 errors are due to basic probe accuracy tolerances, mainly
background signal, probe linearity, and variations in membrane temperature coefficient. Type 3 errors
are related to the operator's ability to determine the conditions at the time of calibration. If calibration
is performed against more accurately known conditions, type 3 errors are appropriately reduced.
Type 1 Errors
A. Meter linearity error: ±0.5% of full scales reading, or ±0.04 mg/l at 25°C whichever is greater.
B. Component and circuitry error: ±0.04 mg/l
Type 2 Errors
A. DO errors caused by temperature compensation for measurements at ±10°C from calibration
temperature: ±1% of 25° C (±0.08 mg/l)
B. DO errors caused by temperature measurement errors: A maximum ±0.2°C temperature error is
equal to ±0.5% (0.04mg/L at 25°C).
Type 3 Errors
A. Altitude: The maximum DO error caused by calibrating to altitude in increments of 100 feet:
±0.18% (< 0.015 mg/l at 25°C)
B. Humidity: Errors occur if calibration is performed at less than 100% humidity. The worst
possible case would be calibration at 0% humidity. The error varies with the calibration
temperature as follows:
Temperature Calibration Error at 0% humidity
0
o
C 0.09 mg/l
10
o
C 0.14 mg/l
20
o
C 0.21 mg/l
30
o
C 0.33 mg/l
40
o
C 0.50 mg/l
Approximating the Error
It is unlikely that the actual error in any measurement will be the maximum possible error. A better
error approximation is obtained using a root mean squared (r.m.s.) calculation:
r.m.s. error = ±[1a
2
+ 1b
2
+ 2a
2
+ 2b
2
+ 3a
2
+ 3b
2
]
½
mg/l
NOTE: This calculation is for a near extreme set of conditions.
If the probe is calibrated in water-saturated air, then type 3B errors (humidity), the largest error of all
types, is virtually eliminated and the maximum possible error is in the order of 0.1 mg/L for the case
of calibrating around 25°C.
YSI Incorporated YSI 58 DO Meter 20