Anemometer Experiments - Part I
Anemometer Experiments
Most people would instantly recognise an anemometer, even if they did not know what it was called.
Exhibit 1Note 1
The idea that 3 half cups, symmetrically arranged would catch the wind, and rotate at a rate proportional to the wind speed almost seems like the place a child would start. These devices have been around since the 15th century in various forms. A quick check on Wikipedia reveals there are many types of anemometer, each exploiting different physical attributes of air movement and pressure.
This article documents an experimental variation on the Hot Wire anemometer, and notes some of the attributes and issues to be resolved in making viable instrument.
Essentially a Hot Wire Anemometer utilises the positive temperature coefficient of wire (resistance of wire increases with increase of temperature) and the cooling effect of air to resolve the relative velocity of the wind. As the wind blows so the wire cools, and the resistance is lowered.
Note 1: Image courtesy of www.dreamstime.com
Experiment
As with most everything in life, every approach has it's trade-offs. The hemispherical cup anemometer is a mechanical device, which has moving parts and is exposed to the elements. It also requires some form of analog translation to be able to measure the velocity of the air, either mechanically or electrically or hydraulically.. One the plus side, it can be self powered
In this experiment we heat an electronic sensor (MCP9700) Note 2 to a temperature above ambient with a constant voltage source, then as the air moves across the sensor we measure the temperature change.
Hypothesis
The hypothesis is that with more air movement there will be a greater temperature differential
Exhibit 2
We used 3 resistors to be the heating element to increase the temperature of the sensor from the 'ambient' to an 'elevated' temperature, typically +30degreesC. With a constant voltage supply the elevated temperature will stabilise at a level where the amount of energy entering the circuit is the same as the amount of energy leaving.
When a moving airflow occurs the cooling effect will drop the temperature, as heat is given up to the air. The more air, the more heat is given up and hence the temperature drops.
In our somewhat amateurish formula in Exhibit 2, what we attempting to show the air has a relatively fixed capacity to absorb heat (shown here as a constant)
The obvious question here is what happens if the ambient temperature changes. , we deal with this question in part II of the anemometer experiment, suffice to say subtracting one from the other leaves us only with the differential i.e. the change due to the moving air
Note 2: Data sheet provided at www.noyoo.net
Exhibit 3
Exhibit 3 shows part of the circuit, firstly the heater comprising three resistors, in our case they are arranged equally spaced around the temperature sensor (MCP9700). The numbers and values of the resistors are not really important so so long as the ratings are not exceeded and there is enough heat to raise the temperature of the sensor.
In our case we used 15ohms, connected to 4v supply to generate 1060mW which raised the temperature of the sensor by +30degrees C
The second part of the circuit shows only the sensor, which in our case was connected to an analog port of Arduino Nano with a small amount of software which reads the voltage at Vout. The MCP9700 is a linear temperature sensor to 0 to +70 degrees C, +/- 1.0 degrees. The data sheet provides traces, so that you can correct in software to +/- 0.5 degrees. The output of the MCP9700 is a voltage somewhere between Vref and Gnd, and is linear at 10mV per degree. So calculating the temperature becomes a simple maths question. It is very important that the reference voltage Vref is as stable as it can be made to be.
You could use many and varied tools to determine the temperature, in our case we used an Arduino Nano, a simple voltmeter and calculator would suffice
Ambient Dashboard
Introducing the Ambient Dashboard, this is a prototype we will be using to record the data we collect from our experimental anemometer
Exhibit 4
The top display is the barometric pressure, and the second (red), is the relative humidity. The displays of interest are the two bottom displays, these are connected to our heater/sensor combination
The ambient dashboard is mounted on a sheet of ply and clamped to the wall.. in future articles we will do a full exposé..
Sensor Board
Given this is experimentation, not even prototyping no effort has been taken to make thinks pretty. This truly is a mock-up.
Exhibit 5
For our purposes there are two sensor boards (Exhibit 5), they are made on perf-board (Veroboard in some places). You can see the MCP9700 in the middle of the board (T0-92 case, looks like a transistor) surrounded by the three resistors. The messy grey material is heatsink compound, we found it made a profound difference is equalizing the temperature
The sensor boards are each mounted in a plastic tube (Exhibit 6), one of which will be force fed air, from a small fan which has control around the speed (hence volume of air), and the other blocked at one end, so that it is exposed to the ambient, but not the airflow
Exhibit 6
Full Test Setup
The full test setup includes the addition of a PC fan and a cardboard tube to direct airflow. Also to blank off one end of one tube with masking tape to reduce airflow.
Exhibit 7
Procedure
In coming to understand the effects of airflow , we varied the speed of a PC fan. Fortunately PC fans often come with a tachometer wire, and hooking this to a oscilloscope gives a proxy for volume of air, based on the frequency of revolution
Voltage applied to fan
|
Frequency on the tachometer
|
Approximations of airflow
|
0 volts
|
0Hz
|
0%
|
4.0 volts
|
35 Hz
|
25%
|
7.5 volts
|
77 Hz
|
50%
|
10.0 volts
|
113Hz
|
75%
|
12 volts
|
131Hz
|
100%
|
So we took the fan speed as a proxy for airflow, and measured the temperature overtime to watch for some level of settling.
In general it took about two minutes for each change to settle on a new temperature +/- 1 degreeC
Sample Readings – Initial
With no fan (speed of 0%) the initial readings look like : Each sensor within +/- 2degreesC
Sample Readings – 25% Fan Speed
At a fan speed of 25%, the temperature on the sample sensor has dropped approximately 17 degreesC
Sample Readings – 50% Fan Speed
At a fan speed of 50%, the temperature on the sample sensor has dropped approximately 25 degreesC
Sample Readings – 100% Fan Speed
At a fan speed of 100%, the temperature on the sample sensor has dropped approximately 27 degreesC
Exhibit 11
Note that from 50% to 100% the differential is quite slight, and that ambient sensor has remain at +/- 1 degreesC throughout.
The pictures here are samples, readings each thirty seconds for 10 minutes on each fan speed setting and produced the following results
Plotting these results delivers a fascinating picture :
Exhibit 12
The largest temperature drop occurs between NO fan and fan at 25%, and you can see the temperature stabilises after about 120 seconds.
There is clear delineation between each fan setting, even given the crudity of the test setup.
Conclusion
In the words of Adam and Jamie (Mythbusters), you would have to say this is PLAUSIBLE, whilst the relationship is not exactly linear it should be possible to program a profile to have a reasonable confidence limit, taking into account settling time and ambient changes
On the plus side
|
On the minus side
|
|
|
Acknowledgments
With special thanks to our visiting Physicist for the creativity and sensible head to keep us right whilst undertaking this somewhat esoteric project ...many thanks Orlando !
This content is published under Attribution-Noncommercial-Share Alike 3.0 Unported license.
Comments
Post a Comment