On Friday, February 19, 1999 Mrs. Jan Heuberger's 3,4,5,6, and 7th period Physical Science classes assembled 48 magnetometers. Students were grouped into teams of 3-4 students who were able to put together the magnetometers, one per group, during a 30-minute standard period. Dr. Odenwald provided a 10 minute introduction at the start of each period, and then the student teams worked diligently during the rest of the period. Each magnetometer was labeled with a strip of masking tape at the bottom, with the names of the students in the assembly team.
On Monday, February 22, the students were instructed that they would be taking the magnetometers home to make their own measurements to search for magnetic disturbances. The plan was for one student from each team to take the soda bottle magnetometer home, set it up, and for the next three evenings, take readings every 30 minutes or so, from 7 PM to 9PM. Additional readings at later times would also be encouraged. The students were all told that these readings were needed by Dr. Odenwald who was only able to make daytime measurements between 9AM and 5PM, and that night time measurements were needed to monitor auroral activity during North American night time conditions.
The following information is a step-by-step guide for setting up the magnetometer at home, and making and recording the measurements.
1) During each of the participating school periods, ask for a volunteer from each of the groups to bring the magnetometer home. 2) Have the student pick up the magnetometer after school to minimize damaging the system. 3) Once the magnetometer arrives at home, the student will need to find a room were the instrument will remain undisturbed for the next three days. The student will have to inspect the instrument for damage during transport from school, and make the necessary repairs so that the sensor card hangs freely inside the bottle and does not scrape the inside of the bottle as it moves. 4) Obtain a high-intensity lamp, or a desk lamp with a CLEAR bulb. Do not use a bulb with a frosted lamp because you will not be able to see a glint off of the mirror with such a bulb. The glint/spot you are looking for is actually the image of the filament of the lamp. 5) With the magnetometer positioned 1 meter ( 39 inches) from a wall on a table, position the lamp so that the center of the bulb shines at a 45-degree angle to the mirror. Search for a glint or spot of light from the mirror on the wall. Make sure the table is stable and not rickety because any vibration of the table will make reading the spot location very difficult. You may also have to relocate the magnetometer several times until you find a convenient location in your house where the spot falls on a wall 1-meter from the magnetometer. 6) Once again, make certain that the sensor card is free to rotate horizontally inside the bottle after you have finished this set-up process. 7) On an 8 1/2 x 11 piece of white paper, draw a horizontal line along the center of the long direction of the paper so that you have a line that divides the paper into two parts 4 1/4 x 11 in size. 8) With a centimeter ruler, draw tic marks every 1 centimeter on this line starting from the left-hand end of the line. Label the first mark on the left end '0', and then below the line, label the odd-numbered marks with their centimeter numbers. '0, 1, 3, 5, ...' If you label every tic mark, the scale will be too cluttered to easily read from a 1-meter distance. 9) With the lamp turned on and properly positioned, find the spot on the wall, and position the paper with the centimeter scale, horizontally on the wall. Before securing to the wall, make sure that as the spot moves from side to side on the wall, that it travels along the centimeter scale in a parallel fashion. It is convenient to have the spot moving in a parallel line offset about 1 inch above the centimeter scale. 10) At the start of your 3-day observing sequence, you will need to shift the centimeter scale so that the spot falls at '15 centimeters' which should be about at the middle of the page. It is very important that you perform this step so that your data can be properly compared to other data being taken. If at any time your magnetometer is disturbed, you will have to reposition the scale so that the spot lands on this number before resuming with your measurements.
Now that we have set-up the experiment, here some tips for recording data:
1) For three days of recording, you will be able to fit Day 1 and Day 2 on the front side of a sheet of ruled paper, and Day 3 on the back side. For each day, leave a blank for the date, followed by 4 columns which you will label from left to right 'Time' 'Position' 'Amplitude' 'Comments'. 2) In the 'Time' column, write down the following times in a vertical list: 5:00 PM 5:30 6:00 6:30 7:00 7:30 8:00 8:30 9:00 9:30 10:00 10:30 11:00 3) The first reading you will make on the first day will always be '15.0' because that is where you set-up the scale on the spot in Step 10 in the instructions above. For the subsequent measurements, you will record the actual spot location on your scale. Do NOT reposition the spot every day. You just need to do this one time at the start of your 2, 3, 5 ...day measurement series. 4) When making a measurement, turn on and off the lamp from the wall plug only. This will avoid accidental vibration or lamp motion if you were to try using the switch on the lamp. You want to avoid disturbing the lamp, magnetometer and centimeter scale during the three-day session. 5) If you know, for a fact, that the set up was disturbed, recenter the centimeter scale on the current spot position at the '15 centimeter' point. Make a note that you did this on the data table at the appropriate time, you can then resume taking normal data at the next assigned time in the data table. Warning, do not assume that just because a big change in the readings occurred, that the instrument was disturbed. You could have detected a magnetic storm!! Only recenter the scale if you physically saw the instrument disturbed, or someone told you that they accidentally touched it. 6) It is important that you make your measurements within 5 minutes of the times listed in the data table. If you are unable to do this for any entry, leave it blank and do not attempt to 'fudge' or estimate what the value could have been. Chances are very good that another student in the network will have mad the 'missing' measurement. 7) The spot on the wall will probably be irregular in shape. Make yourself familiar with what the spot looks like as it moves, and find a portion of the spot that has a good, sharp edge, or some other easily recognized feature. You can also estimate by eye where the center of the spot is if the spot has a simple...round..shape. Try to make all of your measurements in a consistent way each time, and to estimate the spot location to the nearest 0.5 centimeter. Record this number in Column 2 in your data sheet. 8) You may notice several 'behaviors' of the spot. It will either just sit at one location, or it may oscillate from side to side. At a 1-meter distance from the magnetometer, if the spot swings back and forth horizontally by an amount LESS than 0.5 centimeters, consider the spot 'Stationary' and write 'S' in Column 3 after your measurement. If it is obvious that the spot is oscillating back and forth, write 'O' in Column 3 and in Column 4 write down the range of the swing in centimeters along the scale. Example, if it moves from 13.0 centimeters to 17.0 centimeters, write the average position of '15.0' centimeters in Column 2, and then write '13.0 - 17.0' in Column 4. 9) The last thing you would want to note in your data log is local weather conditions IF there is a lightning storm going on. Note the time that the lightning began and ended as a 'Note' on the data page, but don't write this in the data table itself. You also want to mention if the street outside your house is busy with traffic or not. An estimate of how often a car passes would be good to note. 10) When your assigned time is finished, bring the data table and magnetometer back to school.
At the end of the three-day session, students will turn-in their data tables to the teacher. Because students in several class periods may be making measurements on the same days and times, you will have to combine the data into a single set of readings. The way to do this is illustrated by the following example for 'Day 1' in which 5 students reported their measurements:
All students reported no lightning storms. One student, Eve, noted that she lived on a street where cars passed every 2 minutes. We will treat her data separately.
Ben Sarah Eve John Emily
5:00 PM
5:30 15.0
6:00 13.0 15.0
6:30 12.0 15.0 11.5 15.0
7:00 11.0 11.5 11.5 11.0
7:30 10.5 11.0 11.0 15.0 11.0
8:00 8.0 7.5 9.0 8.0 7.5
8:30 6.5 6.0 7.0 6.5 6.5
9:00 9.0 8.0 9.5 9.0
9:30 10.5 10.5 11.0 10.5 10.0
10:00 10.0 11.0 10.5 10.5
10:30 11.0 10.5 11.0 11.0
11:00 11.5 11.5 11.0 11.5
Each student began with '15.0' as the first measurement on the first day of
the series. Sarah and Eve were not able to make some of the measurements, but
because multiple students were 'on the air' at the same time, there are
readings available for all of the times between 6:00 and 11:00 PM.
On a separate piece of paper, which will be faxed/mailed to Dr. Odenwald each Friday by 5 PM, create four columns and label them as follows:
Time Number Average S. Deviation
'Number' will indicate the number of students that reported data for that hour, 'Average' is the simple average of the valid measurements not including the '15.0' calibration. 'S. Deviation' is the Standard Deviation of the data which will be defined below. As a note on this paper, also include information that summarizes relevant neteorological information such as 'between 8:00 and 11:00 PM there was lightning reported in the area'. We now compute a straight average of the readings for the corresponding times
Here is a worked example of the above final data table based on the sample data above:
Example:
Ben Sarah Eve John Emily
8:30 6.5 6.0 7.0 6.5 6.5
9:00 9.0 8.0 9.5 9.0
At 8:30, the measurements we will average are 6.5, 6.0, 7.0, 6.5, 6.5
but because of the high-traffic conditions for Eve's measurements, we will
first average the other data points to get:
( 6.5 + 6.0 + 6.5 + 6.5 )/4 = 6.37
In the Final Report table on the line for 8:30PM, we enter 'N=4' in Column 2
because the data from four students was used in computing the average. We
enter '6.4' in Column 3 where we have rounded UP to the nearest tenth.
Next, we calculate the Standard Deviation for this data as follows:
The basic formula is
S.D. = ( [ ( x1 - x0)^2 + (x2 - x0)^2 + ..... ] /(N-1))^1/2
Here are the steps for the measurements at 8:30PM
1) Subtract the average value of '6.4' from each of the data points:
6.5 - 6.4 = 0.1
6.0 - 6.4 = -0.4
6.5 - 6.4 = 0.1
6.5 - 6.4 = 0.1
2) Square these numbers and add them together:
(0.1 x 0.1) + (-0.4 x 0.4) + (0.1 x 0.1) + (0.1 x 0.1) = 0.19
3) Now divide this 'sum of the differences squared' by the number of
measurements minus 1 ( N-1 = 4-1 = 3)
0.19/3 = 0.063
4) Take the square root of this number and enter it in column 4:
0.25
Time Number Average S. Deviation
8:30 4 6.4 0.25
Now, what do we do with Eve's suspect data? It is suspect because cars can be
sensed by the magnetometer, and if a car happened to pass by her house when a
measurement was being made, it could affect the readings. If we compare Eve's
8:30 measurement with the average of 6.4, we can see that it is '7.0'.
Compared to the other measurements it seems to be slightly higher. Also, if we
checked her data log, we would see that while the other students recorded no
oscillations in Column 3 of their data tables, Eve noted oscillations were
going on. This is a sign of possible interference by passing cars. Because of
this, we elect NOT to include Eve's measurement in computing the average and
S.Deviation for this measurement.
For those of you with statistics background, Eve's measurement differs from the average by 7.0-6.4 = 0.6. The computed S.D was 0.25, so this means that Eve's measurement was 0.6/0.25 = 2.4 standard deviations from the mean. This kind of deviation can be expected to happen about 15% of the time in a random sample, so actually, her measurement is not statistically different from what the other students got. We could actually have elected to use her data anyway, but since she reported different recording conditions ( oscillations) than the other students, we are justified in not using her data. Note, a 1 S.D. difference happens 33% of the time, 2 S.D. happens about 15% of the time, and 3 S.D. happens about 5% of the time. Usually, only 3 S.D. or higher is considered significant in a random sample.
Although the above student experiment is based on evening 'homework' operation, daytime readings are also of interest and could be performed regularly during the year so that longer-duration trends could be examined more conveniently.
Once the data have been averaged, they can be Faxed or emailed ( preferably ) to
Dr. Sten Odenwald, Goddard Space Flight Center, Code 630, Greenbelt Maryland, 20771. The email address is odenwald@bolero.gsfc.nasa.gov. Please put 'MDATA' in the subject field of the letter! Fax 301) 286-1771