Badger Meter Series 4000 Sensor - 1/2, 3/4", 1" - Manual (Page 12)

CALIBRATION
If you are replacing an existing Series 4000 sensor and have already calibrated your flow monitor, no calibration changes are
necessary. For installation of a new flow monitor, please refer to the calibration instructions in the flow monitor manual.
The Series 4000 flow sensor, like all impeller or turbine flow meters, operates by converting kinetic energy (in the flow stream)
into rotation (of an impeller). Indeed, almost all flow sensors work on the principle of converting flow energy to output
signals. The only arguable exceptions are ultrasonic and electromagnetic sensors. The interaction of the flow stream and
the impeller depend, to a currently unquantified extent, on fluid properties (density, viscosity, and pressure), and on physical
properties of the impeller. The Badger Meter impeller design features the following:
1. A low mass polar moment of inertia
2. No magnetic drag
3. Very low eddy current drag
4. Low bearing friction
The impeller housing forms the periphery of a rotating fluid stream, the only source of drag tending to retard the impeller.
The efficiency of this design is the key to the repeatability of sensor output at very low flow rates, and is the reason that the
pressure drop across the installed sensor is so low.
The Series 4000 flow sensors have been calibrated on the Badger Meter flow bench, shown schematically on TD4000CAL. The
calibrating fluid is untreated tap water, at ambient temperature, and at various pressures.
All calibration tests are based on average flow rate and average frequency measured during the delivery of 200 pounds
of water over the time period required to deliver that weight of water at a given flow rate. Our flow bench and calibration
practice conform to ASME/ANSI MFC-9M-1988, measurement of liquid flow in closed conduits by weighing method.
Sensors are calibrated based on the average flow rate. Flow velocity in feet per second is calculated there from based on the
theoretical pipe diameters. Data accuracy, as noted on TD4000CAL, is estimated to be the following for each individual test of
sensor/flow rate combination:
+/-0.15 percent on average flow rate (q)
+/-0.10 percent on average frequency (f )
Tests on at least 12 points across the rated flow range of the specific sensor, including at least one slightly above maximum
rated flow, and at least one slightly below minimum, were performed on at least six different samples of each of the basic
design.
Output data points (72 minimum) for each basic model were analyzed using standard regression techniques, to find the best
linear least squares fit describing flow rate (q) as a function of frequency (f ) in the form:
q=kf+b(EQ1)
The values of k and b are developed to produce the minimum summed squared differences between the actual flow rate at
an observed frequency and the flow rate predicted by EQ1 at that observed frequency. The results of this analysis indicate
that the calibration constants given below are accurate within +/- 1 percent of full scale given the following constraints:
1. Constant pressure
2. Constant temperature
3. Water as the fluid medium
This 1 percent covers the effects of non linearity in the flow rate/frequency relationship and dimensional variations from
sensor to sensor.
Testing to date indicates that over a relatively broad range of operating pressure and temperature, factory calibration should
be adequate to yield results accurate within +/-1 percent of full scale in fluids not markedly different in density and viscosity
from water. Repeatability with most compatible fluids at constant pressure and temperature should be within 0.7 percent or
less, although the actual readings may be significantly in error if the viscosity and density of the sensed fluid depart drastically
from water.
PVC sensor calibration is based on flow through standard schedule 80 PVC pipe. The effects on calibration of joining the
standard schedule 80 pipe terminations in a system of other than PVC schedule 80 have not been investigated. We caution
you to anticipate some deviation in calibration. This effect can be minimized if an additional 10 to 20 diameter lengths of
Series 4000
Page 12 April 2012