PHYS194 Manual PDF

Physics-II Lab Manual
Physics for Engineering
Fall 2013
A. Abdesselam
Cover’s Photo: Aurora Phenomenon (http://en.wikipedia.org/wiki/Aurora_(astronomy))
© Abdelouahab Abdesselam (abdesselam12@gmail.com)

Table of Contents
Introduction ............................................................................................................. 1
Measurement and Error Analysis............................................................................. 3
Measurement and Errors Analysis ........................................................................ 19
The Gaussian distributions .................................................................................... 21
Ohm’s law .............................................................................................................. 26

The Fuse ................................................................................................................ 30
Dielectric Constant ................................................................................................ 33
Electric Conduction Through Semiconductor ..................................................... 36
RC dc Circuit .......................................................................................................... 39
Earth Magnetic Field ............................................................................................. 43
Oscilloscope .......................................................................................................... 47
Transformer .......................................................................................................... 54
Low and High Pass Passive Filters .......................................................................... 59
RLC Resonant Circuit ............................................................................................. 66
References ............................................................................................................. 70
Appendix A: How does a Vernier caliper work? ..................................................... 71
Appendix B: Least-square Fit ................................................................................. 74

Introduction
Phys-194 is a laboratory experimental course designed to accompany the theoretical

lectures given in the Phy-194 course. It therefore presents the student with unique
opportunities to have a firm grasp of the abstract concepts and laws he/she is exposed
to during the lectures. In this course indeed the student verifies experimentally Ohm’s
law, Biaot-Savard law and other laws and experiments with the oscilloscope,
transformers and different types of electric circuits. These subjects are nowadays
considered very basic but nevertheless they are essential to every serious sciences
student. They should constitute an important part of the basement for further higher
and more involved studies.
The laboratory practices are so important for another reason. It is the occasion where
the students can learn precious experimental techniques. They develop measurement
and data analysis skills. It is not rare in physics that the measurement of a quantity
(distance, speed, acceleration,…etc.) put us in front of insurmountable challenges. Only
very highly qualified people genius at times can imagine, design and build apparatus
that make such measurements realizable. Data analysis and extraction of useful physical
quantities could also be arts in their own right. By doing and contemplating lab
experiments the students acquire some of these skills which are also required in dealing
advanced studies and researches.
Science and Physics in particular would not have made the enormous progresses
apparent everywhere around us without uncountable number of meticulous
experiments giving sometimes unpredicted results. On another side, it often happens
that many mathematical theories present themselves as candidates to describe the
physical reality. But it is only by doing experiments that we know which theory is
acceptable and which are not. The experiment has thus the privilege of saying the last
word.
The student should be highly motivated and organized to be able to benefit from this
course. One way to demonstrate a sense of organization is to keep an log book (note
book) for the physics labs. This is the place where student record the data, make
sketches, express his/her thoughts etc. It should be hand be written and needs not to be
clean but should be clear though. The aim of log book is to contain enough information
such that the student could at any later time reconstruct the experiment. Another
1
benefit is that the log book contains data which are not easily obtainable elsewhere.
Although this course is a basic one, the equipments hereby used cannot be found in
anyone’s kitchen.
Finally the students should after each experiment return a well written report which is
graded by the instructor. These grades contributes to from 30% to 50% of the final
grade. Therefore it is of the students’ interest to take the reports quite seriously.
2
Measurement and Error Analysis
This is a quick introduction to Measurement and Error analysis. The material presented here is
crucial for conducting the experiments in the labs and correctly reporting their results. The
student should have a first superficial reading of the content and then come back whenever a
more accurate understanding is required. There are many books1 dealing in detail with the
subject of “Data and Error Analysis” which the serious student should consult to deepen her/his
knowledge. The student is also encouraged to look for other resources on the internet.
1 How do we Report a Measurement

The Data taking or measurement process is an essential step in conducting an experiment and
should be given a great deal of attention and care. Even so, all measurements do have errors
attached to them because of the inherent incapability of any instrument to give the true value
exactly (of course the aim of any measurement is to evaluate the true value). A measurement is
therefore incomplete, to say the least, if it is not accompanied by an estimation of the
committed error. An acceptable measurement of a quantity should then consist of the
measured value, sometimes also called best value, and its error :
2 The Meaning of Error

What do we mean by ? We can first give a rough answer and say that we
mean we do have some confidence that the true value of the quantity is somewhere within
the interval . This is illustrated in Figure 1.
Figure 1: the true could be anywhere within the uncertainty interval
1
See References at the end of this manual.
3
Said otherwise, this means that if we repeat the experiment many times and calculate , we expect
that the majority of the results (of ) will lie within the mentioned interval. A more precise
definition will be given later in this document (see the paragraph entitled “The statistical interpretation
of the error”).
The error above is given in an absolute form, i.e. without comparison to the measured value .
It’s sometime more informing to quote the committed error as a percentage or relative error
(evaluated as ). This, for example, makes it easier to judge if the measurement is a
good or poor one. A relative error of few per cent (i.e. less than 10%) is acceptable in our labs.
On the contrary an error of more than 15% indicates a poor measurement. It’s also mandatory
to use the relative errors when comparing the precisions of the measurement of dimensionally
different quantities. In other words, to be able to say which of such quantities is measured with
a better precision, one has to look at their respective relative errors.
3 Why is it Important to Know the Error

There are at least two reasons why we should know the value of the committed error.
1. The first reason is that the error gives us an idea on the quality of a measurement. A
measurement with a relatively big error is almost useless. Only a precise
measurement, i.e. one with a relatively small error, could be used with benefit in
practice. As said previously, in general, a relative error less than 10% can be
considered as acceptable in our labs.
2. The second reason is that we need the errors whenever we want to say something on
the equality of two measurements. It’s because we know the errors that we can make
a qualitative statement about the two measurements in figure 2-a and say that they
are probably equal. Similarly, because we know the value of the errors, we can
confidently say that those in figure 2-b are most likely to be different. A more precise
criterion in comparing two measurements is discussed in the paragraph entitled
“Comparing two measurement” (see below).
Figure 2: using the errors in comparing two measurements
4
Now, if the errors are that important, how should we proceed to estimate them? Here we
distinguish two situations: Estimating the error on directly measured quantities and indirectly
measured quantities. We treat each in turn.
4 How do we Estimate the Error on a Directly Measured Quantity

We have just seen that it’s very important to know the value of the error committed in any
measurement. We therefore proceed now to discuss how we can evaluate such an error. There
are three different ways of estimating the error on directly measured quantities depending on
the instruments or methods we do use to take the measurements:
4.1 Analog Devices:

Examples of these are the ruler, balance, Vernier caliper…etc. It’s, basically, any device that has
graduations on it. In our labs we will often take the smallest graduation as error. (See Giancoli,
4th Edition, page 3). This is because we cannot read, using the device’s scale, any value smaller
than the smallest graduation (there are no ticks to help us to do that). However, if you think
that this is a too big error, then consider taking half the graduation as error (or even a smaller
fraction). On the other side if you deem that your error should be greater than one graduation,
then take few graduations as error.
Example: if you are measuring the length of an object with sharp ends then a graduation as
error is generally safe. If, in contrary, the ends of the object are not sharp (i.e. they are not
clearly located) then it’s safer to take a few graduation as error.
4.2 Digital Devices

Examples are the stop watch, multi-meter…etc. It’s any device that displays the measurement
as digits. Here one can take the smallest digit that can be displayed as error. For example, if the
instrument displays 1233, then one should report the measurement as 1233 . (this is
because the smallest number that this instrument is able to show is 0001 (=1)).
We can motivate this rule in the following way. Because the device in the example above
cannot display decimal numbers (it gave use the number 1233 without a decimal point), even if
1233 has a decimal part, it could not be shown on the device. So, the measured number 1233
could be in reality any one of the list 1233.0, 1233.1, 1233.2, 1233.3, 1233.4, 1233.5, 1233.6,
1233.7, 1233.8, 1233.9. The device when it returns the number 1233, it does in fact some sort
of rounding to the real number (which could be any one of the above list, but we do not know
which one, because the device, through its limited precision, has lost the decimal part). Now we
can estimate the committed error by saying that it is of the order of the difference between the
highest and lowest value of the considered number: . And
that is why we write . (We could probably do a better estimation by taking an error
5
half of this and write the result as . But in our labs we prefer to stick to the simpler
rule of quoting an error equal to one full unit of the smallest displayed digit).
Note: it should be stressed here that error estimated by the previous prescriptions is often the
smallest error one should quote. It’s the error from the measuring instrument itself. One should
consider error from other sources (e.g. Human errors, time of reaction, difficulties to locate
precisely the ends of an object…etc.). To be more specific an example is given: when measuring
the time interval of an action (a moving object crossing two point), one should note quote the
error only from the stopwatch, which is often very small, but should take account for the
human time of reaction (reflex) and ability to synchronize events.
4.3 The Standard Deviation method.

This method can, and maybe should, be applied whenever one does have at hand many
measurements (say ). (For example one has repeated the same measurement times). The
rational of this method is that the spread of the measurements themselves should contain the
information on the precision of the measuring instrument. We have, therefore, just to find a
way to extract the error from a set of measurement. In fact, there already exists a well known
method, Standard Deviation method, which proceeds in three steps as follow:
a. First one calculates the mean. This will be your best estimate of the quantity under measurement2:
(1)
b. Then one calculates the standard deviation. This tells you how spread your measurements are (It
represents an estimate of standard deviation of the parent distribution3 and thus the error of one
measurement).
2
This can be shown by using the method of Maximum Likelihood. We assume, as is often the case, that the
measurements are distributed following a Gaussian distribution with means and standard deviations (for
instance both unknown). With this assumption, the probability for making one observed measurement is
. This follows from the statistical interpretation of the measurements and their
errors (with the Gaussian distribution as parent distribution). The probability for making all the observed ’s is,
following the probability multiplication rule, the product of the individual probabilities
. The best estimate of the parameter is
the one that gives the maximum probability . To find this maximum we set the derivative with respect to to
zero: . This gives or . Solving this
equation we get . (Here we abuse the symbols and use the same symbol for the true and estimated
value of the mean).
3
A similar calculation as in note 1 above shows that the best estimator of is given by .
(Here too we abuse the symbols and use the same symbol for the true and estimated value of the standard
deviation).
6
(2)
Why do we use and not just in the definition of is explained in detail in one of the reference
books4.
c. Finally one calculates the error on the mean5.

(3)
The result can then be written as . Please note that Eq. 3 shows that the more
measurement you do (this is ) the smaller your error will be ( decreases as .
Note: If it happens that all of the measurements give the same value and thus is null, one
should be careful to note quote a zero error. In this case you should use method one of the
previous two methods). One should also use the previous methods whenever the standard
deviation method gives a smaller result (i.e. one should be on the safe side and report the
biggest obtainable error).
4.4 Number without error

If we are given a number without the error being specified, we can then assume that the error
is equal to a unit of the least significant figure of that number (significant figures are described
a bit further down in this introduction). Here are a few examples:
1. 45678 has an error 1 .

2. 53.3 has error 0.1
3. 0.7830 has error 0.0001
4. 3450 has error 10
We can justify this rule in the following way (similar to the digital device rule explained above).
Consider the number given in item 1 above, i.e. 45678. Whoever gave us this number, he is not
sure about the digit that comes after 8. It could be any number between .0 and .9. he is not
sure about it and that’s why he did not specify it. Therefore 45678 could be in fact 45678.0 or
45678.1 or 45678.2,…, or 45678.9. This defines then the uncertainty in 45678. We can say that
4
See Appendix E in “An Introduction to Error Analysis: The Study of Uncertainties in Physical Measurements”, J. R.
Taylor, 2nd ed. Briefly, this is needed to make the estimator give exactly the true value of the distribution
parameter when an infinite number of experiment is used (
).
5
Eq.3, which gives the error on the mean, can be proved in the following way. The error on comes
from the errors on the single (i.e. . We can then use the method of propagation of errors (see Eq. 4)
, where . We have . Then
7
the error is of the order of the difference between the highest and lowest value of the
considered number: . And that is why we write
.
5 Statistical interpretation of the error

This is the right place to continue the discussion we have started in the paragraph entitled “The
meaning of the error” (see above). A more precise definition of the error relies on the assumption6 that
the distribution of the results of repeated measurements of a quantity follows a Gaussian (or Normal)
curve. Within this assumption, we expect that 68.2% of the results of the measurements lie in the
interval , 95.4% within the bigger interval
, and 99.6% inside the even bigger interval , where and
7
are the mean and the standard deviation of the distribution, respectively . This statistical
interpretation of measurement is depicted in figure 3.
Figure 3:Gaussian distribution of repeated measurements
The error on the mean (or best value) can howevr be much less than and is given by
. (The error get smaller with increased number of measurements . See the previous
section the validity of which is in fact based on the assumption that the measurements are
distributed following a Gauss distribution). The values of , obtained in repeated
8
experiments , also follows a Gaussian distribution, but with smaller standard deviation (to wit
). This is shown graphically in figure 4.
6
This is indeed the case for many practical situations.
7
We abuse the language and neglect the difference between the (parent) distribution parameters and their
estimators.
8
By experiment we mean a set of measurements (like the ones discussed in figure 2). Each experiment gives us a
different value of . By repeating the experiment many times we get a distribution similar to what is shown in
figure 3.
8
Figure 4: statistical interpretation of the error
Please note the differences with respect to the previous figure in both vertical and horizontal
axes labels. Figure 3 deals with measurements within one and only one experiment (where
each experiment consists of many measurements). In each experiment we estimate .
Figure 4, on the other side, deals with many of such experiments. It shows the distribution of
the ensemble of obtained in the many experiments. The Gaussian distribution tells us
that we should expect that 68.2% of the results of the repeated experiments will lie in the
interval , 95.4% within the bigger interval
2∆ , and 99.6% inside the even bigger interval − 3∆ , +3∆ .
6 How do we Estimate the Error on an indirectly measured

quantity
This is the case whenever the quantity in question, which is not measured directly, depends on
other quantities (or variables) which we can measure directly, i.e., they are of the type treated
above. The error on this functional quantity (that is a function of other quantities) is estimated
by propagating the errors of the directly measured variables to the functional quantity.
Assume that is a function of the independent variables . This is symbolically written

as . This means that the value of is readily calculated from the measured value
of It can be then shown9 that the error on is given by10:
9
The error on the quantity is where . We can expand the
function about the point in Taylor series:
we then get . If
the variables are independent, as we have assumed all along this document, then terms like
give, on the average, zero (because we would have as many positive
contributions as negative ones). We thus find , which is Eq. 4.
9
(4)
where denotes the derivative of with respect to while maintaining all the
other variables constant (it is called the partial derivative of with respect to . Equation (4)
says that knowing the errors on the error on can be easily calculated. Let’s look at
few examples:
1.
where is a constant (i.e. depends neither on nor on ). This is the simplest form of . We have here
only one variable and the derivative is quite simple: . Substituting this result in Eq. 4 we
get:
or
Example: we measure the radius of a circle to be . What is the circumference of the circle
and its error?
Answer: The circumference is . The error on is, using ,

. thus .
2.
where and are constants. We have now two variables and and the partial differential with respect to
is (remember, when differentiating with respect to , we
treat the other variables as constant).
In a similar way
The substitution of these derivatives in (Eq. 4) gives
(5)
Example: An object is composed of two parts of mass and . What is the mass
of the compound object and its error?
Answer: The mass of the compound object is . The error on is,

applying Eq. 5, given by
3.
10
Here and mean the same thing (i.e. error on )
10
where is constant (i.e. is proportional to the product of two variables and ).
Then and . Replacing these derivatives in Eq.

4 we get:
We can put this equation in an easy to remember form. We divide by and get:
(6)
Example: We have a rectangular surface of length and width .

What is the area of this surface and its error?
Answer: The required area is given by . The error on is, applying Eq. 6,
given by 1.8 So we have

.
4.
where and are constants. Note that now depends on only one variable.
Then . Replacing in Eq. 4 and dividing by , as done in the previous example, one gets
Or
(7)
Example: If we measure the radius of a disk to be (1.000 ± 0.001)m, what is the error on the area of the disk?
Answer: First the area is . The error is given by
, ie,
5.
Then . As before, substituting in Eq.4 and

dividing by we get (check it!):
(8)
11
Note that this formula is similar to that of case 2 (see Eq. 6).
Example: A cube of volume of (27.00 ± 0.05) and mass (250.0 ± 0.1)g. What is the error on its density
?
Answer:
. Therefore .
6.
where and are constants. The derivative with respect to is . Substituting

in Eq.4 we get:
Dividing by we find:
then
Note: if , we can write this as . We then can apply the formula above with
replaced with .
7.
where and are constants.
. Replacing in Eq.4 and dividing by we get:
Then
8.
where and are constants.
12
and thus, using Eq.4, we get
Similarly one can show that if then
9. One can continue in a similar way with other kind of functions.
7 Significant Figures
Significant figures in a given number are the figures that we know reliably, i.e. we are not
(totally) uncertain about them. Take for example the following result of some measurement:
. The error tells us that we are quite sure about the first three figures
( and a bit less sure about the fifth one (that is 4). But we are totally unsure about what
comes on the right side of 4; it could be any digit between 0 and 9. So we have in all 4
significant figures: . This means that even if the result was given to us in another form
, we should round it and keep only 4 figures (
), because the error tells us there really are only 4 significant figures.
7.1 Counting significant figures

The significant figure on the far left is the most significant one. The significant figure on the far right is
the least significant figure. All the digits between the most and least significant digits, including zeros,
are significant. The zeros on the left or right of a number may be problematic and we need clear rules to
remove any ambiguity in counting significant figures. We have essentially two cases:
1. Zeros on the left are not significant:
Example: 012 and 0.12 both have two significant figures. 0.001 has only significant figure (the
zeros on the left do not count!).
The reason for this rule is quite obvious for non-decimal number. We can discard the zeros
without changing the value or the precision of the considered number. For example 012 is the
same as 12 and thus both numbers should be considered as having two significant figures.
The rational for a decimal number (like 0.001) is more subtle. The zeros are not considered
significant because one can change the unit (i.e. use a smaller unit) and recast the number in a
non-decimal form (example: and this 0.001 has
only one significant figure). In that case the zeros just inform us about the used scale. (Do not
confuse this with errors. We are assuming that 0.001 expresses the result of a measurement and
not the value of whatever error).
13
2. Zeros on the right, usually, are not significant for a non-decimal number but they are significant
for a decimal number:
Examples: 1200 has 2 significant figure (by the first part of this rule). 0.10 has also two significant
numbers (the second part of the rule).
The rational of the first part of this rule is as follow: assume you are given the number 1234 and
are asked to round it to two significant figures (because of some error, we are not sure about the
last two figures 3 and 4). What would you do? You cannot of course write 1234 as 12, since this
totally changes the value of the number, but you can write it as 1200 and state that the zeros on
the right are not significant. Thus 1200 would have only two significant figures. The zeros in this
case are just place holders (telling us where are the position of units, tens, hundreds, etc.). they
are not the real number zero (like in 0, 1, 2, 3,…etc.).
Note: it may occur that a zero on the right (of a non-decimal) number is significant (i.e. it is really
the number zero and the zero place-holder). In this case a decimal point should be added at the
right end of the number (if the zero in 120 is significant then write it as 120.).
The situation is different for a decimal number (like 0.10). A zero, if it is there, is significant. If it is
not significant we should not write it in the first place (write 0.1 instead of 0.10). Therefore zeros
on the right of a decimal number are significant (you cannot add or drop them just like that! They
represent precision and precision costs money!).
Examples:
a. 15734 has 5 significant figures

b. 71.090 has 5 significant figures
c. 0.0051 has 2 significant figures
d. 00063 has 2 significant figures
e. 04560 has 3 significant figures
7.2 Operations with significant figures

One has to be careful to not gain precision just by doing arithmetic operations. For example if one does
the operation 25/12 using a calculator one gets on the screen 2.0833333. One should not just copy this
number as it is given by the calculator, because it pretends far more precision than contained in the
original numbers 25 and 12. Instead one must round the result to express the original precision and
report in this specific case 2.1. When operating with numbers, a distinction is made between addition
one side and multiplication or division on the other side. In the latter case the number of significant
figures of the least precise factor is of interest while in the former case it is the position of the least
significant digit which is of importance. The following makes things clearer.
1. Multiplication and Division: when multiplying or dividing two numbers the result should
retain a number of digit equals to the number of significant figures of the least precise factor. This
rule can be understood by considering error propagation as follow: take for example the operation 8 × 8.
Since we are not given an error on both factors (8 and 8) we assume an error equal to a unit of the least
14
significant figure i.e. 1 (we then have 1 and ). The error on the product is
(see error propagation above). This means that . On the other

side, following the actual rounding rule, we can directly write . This makes sense because
the number 60 means an error equal to 10 which is of the same order of the one we just calculated using
error propagation. In other words, is the same results as . Other rules, e.g. the rule for
addition, can be motivated using the same reasoning.
Examples11:
 8 × 8 = 60  both factors have one significant figure, so should do the result (60 has only one
significant figure).
 8 × 8.0 = 60  the first factor (8) has one significant figure and the second factor (8.0) contains two
significant figures. The result should, as stated above, retain only one significant figure.
 8.015 × 8.0 = 64  the first factor (8.015) has four significant figures and the second factor (8.0)
contains two significant figures. The result should retain two significant figures.
 8 / 2.0 = 4  the first factor (8) has one significant figure and the second factor (2.0) contains two
significant figures. The result should retain one significant figures.
 8.6012/2.0 = 4.3  the first factor (8.6012) has five significant figures and the second factor (2.0)
contains two significant figures. The result should retain two significant figures.
2. Addition: when adding (or subtracting) two numbers the result should be rounded to
the position of the least significant digit of the least precise number.
Examples12:
1. 1 + 1.1 = 2  1 is significant to the ones place, 1.1 is significant to the tenths place. Of the two, the least
accurate is the ones place. The answer cannot have any significant figures past the ones place.
2. 1.0 + 1.1 = 2.1  1.0 and 1.1 are significant to the tenths place, so the answer will also have a number in
the tenths place.
3. 100 + 110 = 200  100 is significant to the hundreds place, while 110 is significant to the tens place.
Therefore, the answer must be rounded to the nearest hundred.
4. 100. + 110. = 210.  100. and 110. are both significant to the ones place (as indicated by the decimal), so
the answer is also significant to the ones place.
Note: From the previous examples one sees that one could lose precision when operating with two
numbers with different precision (number of significant figures). That is why it’s important to measure
the involved quantities to the same precision.
11
These examples are borrowed from Wikipedia (http://en.wikipedia.org/wiki/Significance_arithmetic)
12
These examples are copied from Wikipedia (http://en.wikipedia.org/wiki/Significance_arithmetic). More
examples are listed there.
15
8 Types of Error, Precision and Accuracy
We distinguish two types of error: random and systematic. We also define the physical meaning
of the precision and accuracy.
8.1 Random errors

This is the type of errors we described its estimation in the previous sections. This kind of error
causes the result of repeated measurements to be randomly13 distributed around a mean value
(usually standing for the true value). Its origin lies in the inherent limitation of any apparatus to
produce the same value in all measurements. As we have seen above, see Eq. 3, the magnitude
of this error can be reduced by repeating the measurement many times.
8.2 Systematic errors

This type of errors tends to shift systematically the mean value of the measurements towards
either smaller or larger values than the true value. In other words, the results of repeated
measurement will be all smaller than the true value or all larger. This means than we cannot
reduce this error by repeating the measurement. Usually a systematic error can be detected14
by analyzing carefully the experiment and by comparing the actual result with other similar
experiments. Once detected and estimated, the systematic error should be corrected for in the
measurements by either adding or subtracting its value to or from the best estimate. If the
actual error causes the measurements to be larger than the true value, then we should subtract
it, otherwise add it.
8.3 Precision and accuracy

Precision and accuracy have different meaning in physics although they are interchangeable in
everyday language uses. The precision is related to the random error: The smaller this error is
the more precise the measurement will be.
The accuracy, on the other side, describes how far is the measured value from the true value.
The accuracy is rather related to the systematic error. Figure 1 sketches the difference
between precision and accuracy: the center of circles represents the true value and the dots
represent repeated measurements.
13
It often happens that the distribution is Gaussian.
14
Systematic errors are more difficult to detect than random errors.
16
Figure 5: accuracy versus precision
9 Comparing Two Measurements15

If two quantities and are exactly known (i.e. no errors is attached to them) then they are
equal if their difference is equal to zero:
(9)
But if it happens that we know the two quantities only approximately, i.e. an error is attached
to each of them, then the difference alone will not tell us if they are equal or not. For example,
the two quantities may be equal in reality but the measured values could produce a non-zero
difference. Think of two pieces of wood which do have the same length but, because of the
finite precision of any measurement, they are found to have different length. The equation
above gives then a non-null difference.
We need therefore an alternative way to test if two quantities are equal. We obviously have to
make use of the errors which we denote and . The error on is given by (see example 2
in the error propagation section):
(10)
We could then say that the measured quantities and are equal if their difference is within
the error
15
Comparing two measurements can be a very serious and difficult business. Here we propose a rather naïve and
intuitive method.
17
Or to have mathematically well behaved quantities
This can be written as:
(11)
where , pronounced chi-squared, is a random variable in the sense that if we repeat the
measurement of and and calculate every time, then obtained values will be randomly
distributed and follow a chi-square distribution. The mean value of the distribution is equal to
the number of degree of freedom for the situation at hand. In our case, Eq. 11, we have one
degree of freedom. Because of statistical considerations, we will say that two measurements
and are equal if and different otherwise i.e. we use 3 and not 1 as in Eq. 11 as follow:
The value 3 is chosen because if really equals , then by doing several measurement on
them, only less than 5% of the measurements will produce a greater than 3. Thus if it
happens that your is less 3, then you are almost sure that and are equal. (To improve
our confidence in the comparison outcome we need to either do more measurements and/or
improve a measurement precision).
Remark: If the prescription just given suggests to you that two values are different, but which
by other means you know are equal, then you may want first to make sure that you have not
underestimated your errors and second look for any possible systematic error that you may
have ignored in your measurements.
10 Fitting Procedure
Fitting data means finding the theoretical curve which describes the data. Most of the fitting
you will do in this lab, if not all, are linear. This means that you assume that your data (or
measurements) are represented by a straight line. The fitting operation provides you with the
parameters of this line, namely, the slope and intercept. You can do that very easily by using
Excel. The theory behind the linear fitting, used by Excel, is described in Appendix B. you are
encouraged to have a look at it.
18
Measurement and Errors Analysis
Pre-Experiment (1)
1 Aim of the experiment

The aim of this session is to introduce you to measurements and error analysis.
2 Introduction
The material needed for this lab has been explained in detail in the introductory chapter
“Measurement and Error analysis”.
3 Significant figures
1. How many significant figure does each of the following numbers have:
a. 13948 …………
b. 71.090 …………
c. 0.0051 …………
d. 00063 …………
2. Round the previous numbers to at most three significant figures.
a. …………….
b. ……………..
c. ……………..
d. ……………..
3. A series of measurement gave us the following numbers (in some unit). How would you
report the final result (i.e. best value and its error) ? (Keep one significant figure in the
error).
17.47468 17.47392 17.47616 17.47499 17.47257
4. Write properly the results of the following operations:

a. 456 + 38938 =
b. 0.4565 + 12.4 =
c. 34 x 124 =
d. 2346 / 233 =
19
e. 123.5 / 94 =
4 Error Propagation
In general, if is a function of and : , the error on is given by
Applying the above formula, what is the error on dues to the errors on and in each of
the following cases? You should write the derivation steps. (Hint: see “Data and Error Analysis”
document).
1. , and are constant.

2. , is constant.
3. , is constant.
5 Application of Error Propagation

1. Measure the dimensions of the parallelepiped given to you (length , width and
height ) and estimate their error.
2. Calculate the volume of the parallelepiped and its error
3. Measure the mass of the parallelepiped and estimate its error.
4. Calculate the density of the parallelepiped and its error .
20
The Gauss distribution
Pre-Experiment (2)

You investigate in this experiment the Gauss distribution (also called the Normal distribution)
which, in many cases, represents the random errors distribution.
2 Introduction
The Gauss distribution of the random variable is described by the following function:
(1)
Where and are the mean and the variance, respectively. Thus the Gaussian distribution
has two independent parameters ( and ). Figure 1 shows graphical representations of the
Gaussian distribution for different values of the mean and the variance.
Figure 1: Gaussian distribution16
One can easily see, either from Eq. 1 or from the graphs in Fig. 1, that the Gaussian distribution
is symmetric around the perpendicular axis passing through its mean.
16
See http://en.wikipedia.org/wiki/Normal_distribution
21
One reason for the importance attached to the normal distribution is that it is very often used
to describe the distribution of the errors of a measurement. What does that mean? Assume we
want to measure the length of a table and, to do so, we ask 100 students to take the
measurement with a same ruler. Because each measurement has inevitably an error, the
responses of the 100 students will be different. The question is then how are their results
distributed (here the X-axis represents the length of the table and the Y-axis the number of
students)? The response to this question is that the distribution should approach the normal
distribution with a specific mean and variance and , respectively: About 68% of the
students will report a value which lies in the interval . (We expect the rest to
record a value outside this interval). If we take the wider interval , then we
expect that about 99.7% of students would have their measurement inside that interval. This
statistical interpretation of the error is illustrated in figure 2.
Figure 2: Statistical interpretation of the error
The mean of the distribution will be our best estimate of the length of the table. The error on
that estimate is given by where is the number of measurements (here the number of
students). Notice that the error on the mean is given by and not by the standard deviation
(which is the error of one measurement). Thus the more measurements we take, the less the
statistical error is.
The symmetry propriety of the normal distribution has a simple significance now: there is no
reason why the number of students who measure a length of the table greater than the mean
by certain amount ( , say) should be much different from the number of students who get
a length smaller than the mean by the same amount ( There should be no bias neither to
the right side nor to the left side of the mean.
I this lab we will do a similar measurement experiment. We will measure the number atoms of
a radioactive substance which decay in fixed amount of time (30 second). But instead of calling
for 100 students to take the measurements, we will repeat the experiment 100 time (which is
equivalent). The electronic equipments make this task particularly easy.
22
3 Equipment needed
 Radioactive source: this is a Beta source which emits electrons when the atoms disintegrate
(beak up) .
 Geiger tube: It detects the particles (electrons) emitted by atoms.
 Counter: It counts the number of particle detected in a fixed period of time.
4 Experimental procedure
1. The setup up should be already connected for you (consult the instructor if it’s not).
2. Make sure that the time interval for counting is set to 30 sec.
3. Press the “Count” button and wait until the “Stop” light comes on. Record then the
displayed number in the table below.
4. Press the reset button and do step 3 again. Stop when you have accumulated 100
reading.
Table 1:Gaussian distributed counts
Trial Number Counts

1
2
…
99
100
5 Data Analysis
1. Enter the counts ( ) of table 1 in an Excel sheet (use one column which we
call counts column).
2. Make a histogram with the counts in table 1. Here is the procedure:
a. Create the bin column as follow:
i. Calculate the bin width
ii. Fill in the bin column. It should have 12 entries with following values:
Bin number bin column

0
1
2
… …
11
b. On the Excel menu, click on “DataData Analysis”. A small window will pop up. (If
you don’t have Data Analysis package, you should install it:
23
i. Click the Microsoft Office Button , and then click Excel Options.
ii. Click Add-Ins, and then in the Manage box, select Excel Add-ins.
iii. Click Go.
iv. In the Add-Ins available box, select the Analysis ToolPak check box, and then click OK.
 Tip If Analysis ToolPak is not listed in the Add-Ins available box, click Browse to locate it.
 If you get prompted that the Analysis ToolPak is not currently installed on your computer,
click Yes to install it.
v. After you load the Analysis ToolPak, the Data Analysis command is available in the Analysis
group on the Data tab.
See also Excel help: http://www.java2s.com/Tutorial/Microsoft-Office-Excel-
2007/0200__Data-Analysis/AddInstallDataAnalysisAddin.htm).
c. Select “Histogram” on the pop up window and press “ok”.

d. Now you enter the histogram input data in the specified fields:
i. Input range: here you enter counts column
ii. Bin range: here you specify the bin column
iii. Chart output: tick this box to create a histogram
e. Press enter. This will create a histogram (in a new sheet).
3. Fit17 the histogram with a Gaussian as follow:
a. Calculate the mean and the standard deviation of the counts column (Use Excel
AVERAGE and STDEV functions, respectively).
b. Create the Gaussian column: For each entry in the bin column, calculate the
expected frequency by using the Excel function:
where is the number of counts (in table 1). The mean and standard deviation
were calculated in the previous item. represents the Gaussian (Normal)
distribution:
17
Properly speaking the obtained curve is not a fit to our data. It seems that Excel does not have the possibility to
fit data with Gaussian curve. The actual recipe is the closest one can get!
24
Bin number Gaussian column
0 )
… ...
b. Now you have to draw this column on the same histogram previously obtained. You
can do that as follow:
i. Right click on the chart of the histogram (on the Excel sheet) and select “Select
data”.
ii. Click on “Add” (under “Legend Entries (Series)).
iii. Enter “Gaussian fit” or any other title in the “Series name” field.
iv. Enter the Gaussian column in the “Series values”. You have to erase the default
content first.
v. Press ok
vi. You need to change the display of the Gaussian fit: right click on the
corresponding histogram and select “Change Series Chart Type”. Select “XY
(scatter)” with a smooth line.
5. You should obtain a figure like this:
5. Comment on how well your measurements are represented by a Gaussian distribution

(i.e. comment on the quality of fit).
6. What is your measurement of the mean number of decays in 30 sec and its errors?
7. What is the mean number of decays per one second.
6 Further questions
1. What is the relationship between the normal distribution and the process of
measurement?
2. What’s a histogram? (make sure you understand what it is and how it’s constructed!).
3. What’s a Poisson distribution (and by which function is represented)?
4. What is a binomial distribution (and by which function is represented)?
25
Ohm’s law
Experiment (3)

The purpose of this experiment is to verify Ohm’s law and find the equivalent resistance of
series and parallel connections.
2 Introduction
Ohm’s law, which is empiric, relates the electric potential drop across a conductor to the
current passing through:
Since in this experiment we will set a value for the potential and measure the produced current,
we re-write the previous relation as follow (so that is a function of the independent variable
):
(1)
This is, of course, a straight line equation with as slope.
3 Equipment needed
Power supply, resistors, multimeters.
4.1 Part 1: measurement of
1. Connect the circuit shown in figure 1 using one resistor (call it )
26
Figure 1: Measurement of resistor
2. Set the power supply voltage to the values shown in the table given below and record
the corresponding currents. In all cases, you should not let the current exceed 2 Amps to
protect the resistors. Do not forget to record the errors on and and specify the units.
(volts) (amperes)
2
3
4
5
6
7

1. You could use the same previous procedure to measure the resistance and its error of
the second resistor. But instead of doing so, use the multimeter to measure the
resistance of . (No table is needed in this part).
2. Estimate .
4.3 Part 3: measuring the total resistance when and are in series
1. Connect the two resistors in series as shown in figure 2.
Figure 2: and are in series
2. Repeat steps 2 of part 1 for this new circuit and fill in a new table (you need 5 or 6
measurements). Remember that In all cases, to protect the resistors, you should not let
the current exceed 2 Amps.
4.4 Part 4: measuring the total resistance when and are in parallel
1. Connect the two resistors in parallel as shown in figure 3.
27
Figure 3: and are in parallel
2. Repeat steps 2 of part 1 for this new circuit and fill in a new table (you need 5 or 6
measurements). Remember that In all cases, to protect the resistors, you should not let
the current exceed 2 Amps.
5 Data Analysis
1. Plot versus (i.e. on the Y-axis and on the X-axis), using Excel, and draw the error
bars (Error bars represent the error on the Y-axis). Do not forget the units and titles.
2. Fit linearly your data and get the slope and the intercept.
3. Using the Excel function linest get the errors on the slope and intercept.
4. Does your data represent the linear model given by Eq.1? (look at the fit quality variable
).
5. Plot your data, error bars, fit line on a graph paper this time (only for this part).
6. Calculate and its error ( ) using the slope and it’s error.
5.2 Part 3: measuring the total resistance when and are in series
1. Follow steps 1-5 in part 1.
2. Calculate the total series resistance and its error .
3. Using the values of and found in part 1 and 2, calculate the theoretical total
resistance for two resistance in series and its error .
4. How does compare with . Are they equal? (use the method).
5.3 Part 4: measuring the total resistance when and are in parallel
1. Follow steps 1-5 in part 1
2. Calculate the total series resistance and its error .
3. Using the values of and found in part 1 and 2, calculate the predicted
(theoretical) and its error .
4. How does compare with . Are they equal? (use the method).
6 Further questions
1. How does compare with and taken individually? (i.e. is
bigger/smaller than or ?). Does it make sense to you? Explain.
28
2. How does compare with and taken individually? (i.e. is
bigger/smaller than or ?). Does it make sense to you? Explain.
3. Does Ohm’s law apply for a capacitor in DC current?
29
The Fuse
Experiment (2)

You investigate in this experiment the dependence of the fusing current of a copper wire on its
diameter. The fusing current is the current at which the wire blows out and thus stops
conducting.
2 Introduction
It is proposed, see footnote18, that the fusing current depends on the wire diameter in the
following way:
(1)
where and are constants which depend on the wire’s material.
Your task in this experiment is to verify if that relation holds and, If so, estimate the constants.
We can put Eq. 1 in a more useful (linear) form by taking the natural logarithm of both sides:
(2)
If we consider as the independent variable and as the dependent variable then Eq.2
is the equation of straight line where represents the slope and the intercept You will
use this form to analyze the data.
18
The heat produced per second (this is the power) by a current (in amps) flowing through a wire of
length , diameter and resistance (in ohms) is equal joule/second (this is the watt). This causes
the wire’s temperature to rise to an equilibrium value which depends on the emissive power of the wire.
The emissive power of the wire is proportional to its lateral area and the temperature gradient
between the wire and the surrounding . At equilibrium we then have: the
input power = the emissive power, or . Now the resistance of a wire of cross section area
is given by where is the resistivity of the wire’s material. Thus the
equilibrium equation can be written as which is equivalent to or
.
30
3 Equipment needed :
Power supply, set of wires, rheostat, ammeter
1. Take a wire from the provided set and measure its radius using the micrometer. Do not
forget to estimate its error.
2. Use that wire and reproduce the circuit shown in the figure below. The rheostat, which is
a changeable resistor, should be at its highest value (check with the instructor). This is so
because we want to start the smallest current.
Figure 1: Fuse circuit
3. Slide the movable part of the rheostat as to decrease its resistance. Note that the current
increases.
4. Continue to increase the current slowly until the wire burn out. Record the value of the
current when this happens and estimate its error.
5. Repeat steps 1-4 for other wires from the same set (i.e. the same diameter) and fill the
following table.
6. Repeat steps 1-5 for other diameters.
Set No Wire’s diameter (mm) Fusing current (Amps)
1
2
3
4
5
5 Data Analysis
1. Calculate the error on using the standard deviation method. Compare the result with
your estimation in item 4 above. Take the highest value as error.
31
2. Plot versus , using Excel, and draw the error bars. Do not forget the units and
titles.
5. Does your data represent the model given by Eq.2 (and thus Eq.1)? look at the fit quality
variable .
6. Plot your data, error bars, fit line on a graph paper this time.
7. Calculate , , and .
8. The theoretical value of is (see footnote 1), how does your measurement compare
with this value. (use the method).
6 Further questions
1. Does the length of the wire has any effect on the fusing current? Explain.
2. Find out, form the internet, which materials are used in normal fuses.
3. Are there any other models (function) that describe the relation between the fusing
current and the wire dimensions?
32
Dielectric Constant
Experiment (3)

The purpose of this experiment is to measure the dielectric constant of an insulator (a
dielectric).
2 Introduction
The dielectric constant of a given insulator is defined as the ratio of the electric permittivity of
the material to the electric permittivity of vacuum :
Note that has not units. It follows that the dielectric constant of vacuum is 1 (
). The dielectric constant of air is very close to 1 ( ). In this experiment we will
simply assume that .
The capacitance of a two parallel plates capacitor is given by19:
(1)
where is the plate area, the distance between the two plates and the electric permittivity
of the insulator inserted between the plates.
19
The capacitance of a conductor is defined as . For two parallel plates, with distance in
between, we have , where is the constant electric field between the plates (see
Giancoli). in the last equation is the charge density of one plate and equals . Putting all this together
we get .
33
3 Equipment needed
Parallel plate capacitor, capacitor meter.
4.1 Part 1: measurements with plastic
1. Connect the circuit shown below and switch on the power. Experiment with reading the
capacitance on the capacitance-meter.
Figure 1: measuring capacitor circuit
2. Insert one of the plastic sheets provided to you and read the capacitance off the meter.
Read also the distance between the plates and fill in table 1.
3. Add other sheets, once a time, record the corresponding capacitance, read the total
distance between the two plates and complete the table.
Table 2
Number of sheets Distance Capacitance

1
2
3
4
5
6
4.2 Part 2: measurements with air

1. Repeat steps 1-3 of part 1 but this time without using plastic. Try to use the same
distances you recorded in table 1.
2. Fill in table 2.
Table 3
Distance Capacitance
34
5 Data analysis
1. Do steps 1-4 for part1 and part2:
2. Plot versus , using Excel, and draw the error bars. Do not forget the units and titles.
5. Does your data represent the model given by Eq.1? (look at the fit quality variable ).
6. Plot your data, error bars, fit line on a graph paper this time (do this for part 1 only).
7. You can easily see that , where and are the
slopes obtained for part 1 part 2, respectively. Calculate and .
8. The textbook value of for plastic is (see Giancoli page 638). How does your result
compare with this value. (use the method). Do they agree?
6 Further questions
1. Explain why the capacitance of parallel plate capacitor is proportional to the area of the
plate. It’s not proportional to the volume, why?
2. What’s the effect of inserting a dielectric between the plates on the capacitor’s
capacitance? (i.e. does it increase or decrease?)
3. By what factor does the capacitor then change?
4. Could you explain why this happens?
35
Electric Conduction Through Semiconductors
Experiment (4)

In this experiment you will investigate the dependence of the resistance of a semiconductor
wire on its temperature.
2 Introduction
The behavior of the resistance of a semiconductor wire in function of temperature can be
modeled by the following relation20:
(1)
Where is the absolute temperature and and are constants to be determined by the
experiment.
To use Eq. 1, we need to write it in a linear form. We can do so by taking the natural logarithm
of both side. We find:
(2)
This is a linear equation if we consider and as dependent and independent variables,

respectively. In this form represents the slope and the intercept.
3 Equipment needed
Semiconductor resistor, Ohmmeter, hot plate, oil bath, thermometer.
20
The exponential factor is reminiscent of the Boltzmann carriers distribution at absolute temperature (~
36
4.1 Part 1: heating up
1. Make sure the circuit below is connected correctly and that you can read reasonable
temperature and resistance.
2. Switch on the heater and read the resistance at the values of temperature given in the
table below. Do not forget to estimate the error on both the temperature and resistance.
Table 4
Temperature Resistance
60
70
80
90
100
110
120
4.2 Part 2: Cooling down

1. Switch off the heater and put the oil bath (including the semiconductor wire and
thermometer) off the hot plate.
2. Record the resistance at the temperatures listed in the table below. Do not forget to
estimate the error on your readings.
Table 5
Temperature Resistance
120
110
100
90
80
37
70
60
5 Data analysis
Do steps 1-6 for both tables:
1. First you need to convert your Celsius scale to the absolute (Kelvin) scale by using the
well known formula .
2. Calculate the error on which will be needed to draw the error bars.
3. Plot versus using Excel and draw the error bars. Do not forget the units and titles.
6. Does your data represent the model given by Eq.2 (and hence Eq. 1)? Look at the fit
quality variable .
8. Calculate the experimental values of , , and .
9. How do and of part 1 compare to and of part 2. (use the method). Do they
agree?
6 Further questions
1. What does represent?
2. You can use the semiconductor wire as a thermometer (called thermistor)! You can do
that by measuring the resistance and then find the corresponding temperature using
Eq.1. Assume that you have read a resistance of 250 Ohms (we ignore the errors), what
is the temperature and its error that you should report?
38
RC dc Circuit
Experiment (5)

The purpose of this experiment is to study the charging and discharging of a capacitor and
measure the time constant.
2 Introduction
Figure 1 shows the charging up circuit of a capacitor. The power supply is initially off and the
capacitor is uncharged (there are no electric charges on the plates). In other words both plate
are at the same potential which we will take zero.
Figure 1: capacitor charging up circuit
When one switches the power on, a current is set up and the capacitor’s plates start
accumulation charges. This current, which pass through the resistor R, is proportional to the
potential drop across the resistor (Ohm’s law). At , this voltage is equal , the power
supply value (at , the potential across the capacitor is zero). But as time goes on, a
difference of potential between the capacitor’s plates builds up (because now there are an
amount of electric charges on each plate). This causes the current to diminish21 and become
zero when the capacitor voltage is equal to the power supply voltage .
Thus it’s clear that the current decreases with time until there is no current at all. The precise
way the current changes is given by the following equation
21
You can see that by noticing that, because of the capacitor voltage, the potential across the resistor has now
diminished from its initial value and thus the current too has diminished (The power supply voltage is constant
during time of course).
39
(1)
Where is the initial current (i.e. the current at t=0) and is the time constant of the RC
circuit. This equation can derived by applying Kirchhoff law to figure 1 (see footnote22). To
exploit Eq.1, you need to cast it in a linear form. This is done by applying natural logarithm on
both side:
(2)
If we consider as the independent variable and as the dependent variable then Eq. 2 is a
straight line equation with and as slope and intercept, respectively.
3 Equipment needed
Power supply, capacitor, resistor, ammeter.
4.1 Part1: charging up the capacitor
1. Measure the values of R and C and estimate their errors (put on units).
2. Connect the circuit shown in figure 1. (if in doubt talk to the instructor).
3. Switch on the power supply and increase its voltage until the current shown by the
ammeter is slightly above 60 . Start your stopwatch when the current is exactly 60 .
4. Record the time when the current has diminished by 10 and fill the table given below.
Do not forget to estimate the error on and .
60 0
22
By definition, the capacitance of a capacitor is where is the potential across the capacitor and is the
charge on one plate. Thus the voltage across the capacitor is . The voltage across the resistor is given by
Ohm’s law . Applying Kirchhoff voltage law to figure 1, we get . Differentiating both sides of this
equation with respect to time, we obtain ( and are all constants and ): . This can be
written as and integrated to , where is the current at and is equal (at Kirchhoff
law gives which means ).
40
50
40
30
20
4.2 Part2: discharging the capacitor
1. The capacitor being now charged up (from the previous step), you can discharge it by
removing the power-supply as the following figure shows:
Figure 6: discharging circuit
2. Start the time (stopwatch) when you reconnect the circuit and read the current at that
moment. (you need to inverse the polarity of the ammeter so that its needle move to
the right side. At discharging the current is in the inverse direction to the case where
the capacitor is charging up).
3. Read the time when the current decreases by 10 and fill the following table. Do not
forget to estimate the error on and .
5 Data analysis
Do steps 1-5 for both tables:
2. Plot versus , using Excel, and draw the error bars. Do not forget the units and
titles.
5. Does your data represent the model given by Eq.2 (and thus Eq. 1)? Look at the fit
quality variable .
41
7. Calculate the experimental values of , , and .
8. How does of part 1 compare with of part 2. (use the method). Do they agree?
9. Calculate the theoretical value of and its error using the measured values, in step 1, of
and and their errors.
10. Compare the theoretical value of with the experimental one (take the mean obtained at
parts 1 and 2 for the latter). (use the method). Do they agree?
6 Further questions
1. What’s the role of the resistor in the charging/discharging circuits?
2. What would be the value of the time constant if the resistance was only 1 ? (Note how
small it is! Could you do this experiment in such conditions).
3. The capacitor is used in some application to measure time. Could you explain how this
could be done.
4. Could you cite briefly other uses of the capacitor.
42
Earth Magnetic Field
Experiment (6)

The purpose of this experiment is to measure the horizontal component of the Earth magnetic
field.
2 Introduction
It is well known that the Earth possess a magnetic field whose poles almost coincide with the
geometrical North and South. To measure the horizontal component of this field, we use the
field generated by a circular coil (as explained below). The magnitude of this later field at
distance on the axis of the coil is given by:
(1)
Where is the number of turns in the coil, its radius, the current passing through and
the permeability of free space which equals . This formula can be derived
from Biot-Savard law which gives the magnetic field produced by an element of current (see
footnote23).
23
Biot-Savard law says that the magnetic field due an element of current is given by (see
th
Giancoli, P743-744, 4 Ed.): (see Figure below).
The sum of the vertical components of all element contributions is zero (the contribution of an
element is cancelled by a diametrically opposite one). Hence comes only from the parallel
components and is along the coil axis. Its magnitude is given by:
43
Eq. 1 reduces a much simpler form if we consider, as we will do, the field only at the center of
the coil:
(2)
To obtain an accurate measurement one has to make sure that the compass lies totally in the
plane of the coil and at its center (See Fig.1). The coil magnetic field is then perpendicular to the
compass magnet.
Figure 1: Compass lying in the plane (and center) of the coil
The net effect of the two fields (Earth and coil) is to orient the compass in the direction of their
sum, see Fig. 2.
Figure 2: Earth and coil magnetic fields effect on the compass magnet
We can see, from Fig.2, that we have the following equation:
(3)
Which we write as:
(4)
. Now is given by Biot-Savard law mentioned above. In our case it gives
. Replacing this in the equation of we get:

which is Eq. 1.
44
Where is constant in this experiment. This equation can be put in a linear form if we
take and as the dependent and independent variable, respectively:
The slope and intercept of the corresponding straight line are then and zero, respectively.
3 Equipment needed
Power supply, coil, compass, ammeter.
1. Measure the radius of the coil and estimate its error:
2. Record the value of the number of turn in the coil .

3. Make sure the compass needle is in the center of the coil.
4. Rotate the coil so that the compass magnet (not the indicating needle) is in the plane of
the coil. This way the magnetic field generated by the coil is perpendicular to the
horizontal component of the field generated by the earth (see Fig. 1 and 2).
5. Set the current at the values listed in the table given below and measure the
corresponding needle deviation angle (the current values may be change if they are not
suitable for your setup). Estimate the error on and (explain how you do that).
0.10
0.15
0.25
0.40
1.00
5 Data analysis
45
2. Plot your data, on the Y-axis and on the X-axis, using Excel and draw the error
bars. Do not forget the units and titles.
5. Does your data represent the model given by Eq.4? Look at the fit quality variable .
6. Plot your data, error bars, fit line on a graph paper this time.
7. Calculate and its error (consider also the error due to the error on ).
8. Compare your measurement with the world best value?
6 Further questions
1. How would one proceed to measure the vertical component of the earth magnetic
field.
2. Aurora is a well know phenomena occurring at the earth polar regions. Look up the
phenomena on the internet and give a very brief explanation.
46
Oscilloscope
Experiment (7)

This experiment introduces you to a very important measurement tool: the Oscilloscope.
2 Introduction
The Oscilloscope displays, on its screen, the real time values of differences of potential
(voltages). Most Oscilloscopes have two channels and can therefore show simultaneously two
voltages. This feature makes them very handy in understanding and debugging the behavior of
different electronic circuits. The main component of an Oscilloscope is the Cathode Ray Tube
(CRT). The CRT can be divided in three basic sections:
1. The electron gun which produces a focused electrons beam.

2. The deflection part where the beam is deflected in horizontal and vertical direction. The
vertical deflection, usually, represents the measured voltage and the horizontal
deflection represents time.
3. The screen which visualizes the position of the beam (beam spot). The screen has also a
grid allowing the reading of the beam spot position (representing voltage versus time for
example).
In what follows we describe briefly the generation of electrons and their journey all the way to
the screen. Figure 1 show schematically the main components involved in this process:
1. Thermionic electrons are emitted from a hot filament pictured, in red, on the far left of
the figure. These electrons are accelerated towards a grid set at a positive potential. The
value of the potential can be increased or decreased, by knob 1 (see figure below24), with
the effect of attracting more or less electrons and thus affecting the intensity of the
beam.
2. The electrons beam, for instance still diffuse, is then focused by a set of anodes. The
potentials on these anodes are such that the electric fields push the electron to form a
24
These knobs are of course on the front side of the Oscilloscope. You be asked later in this experiment to identify
their functionalities.
47
narrow beam (the effect is similar to a focusing lens in optics). The potential of the
central anode can be changed, by knob 2, to make the focusing optimum.
Figure 7: Main components of an Oscilloscope
4. The so obtained narrow beam passes then between the plates of a first capacitor. This
capacitor, which is connected to the signal one wants to measure, deflects the beam up
and down (i.e. vertically) depending on whether the applied voltage is positive or
negative, respectively. It also constitutes the first channel (an Oscilloscope usually has
two channels. The second one is not shown on the figure). The signal can be coupled to
the capacitor either by direct coupling DC or alternate coupling AC. It can also be
coupled to the ground (i.e. zero voltage). One can switch between these three modes
using knob 5. Using a voltage divider one can choose, using knob 6, to take only a small
part of the input signal. This is useful when the input signal is too high and can thus
deflect the beam well outside the screen. Finally, one can shift the input signal up and
down by applying to the capacitor a constant voltage. This is done with knob 3 (using a
voltage divider again).
5. The capacitor mentioned in the previous item will shift the beam vertically, so one will
see a vertical line on the screen (Because of the usually high frequency of the signal the
eye (and maybe even the glowing phenomenon on the screen) can not follow up the
beam spot and one actually sees a segment of a line instead of just a spot). We need
therefore a second capacitor to deflect the beam horizontally. However, the voltage
responsible of this deflection is generated by the oscilloscope itself using an electronic
system called ramp generator (it can be otherwise but it’s not important for us here).
The ramp generator produces uniformly changing voltages called sweeps (see figure 2-
c). Each sweep causes the beam spot to travel form one side of the screen to the other.
48
Here also one can shift to left or right (i.e. horizontally) the input signal by applying,
using knob 4, a constant voltage. There is however a difficulty here: When the second
capacitor, through the ramp voltage, moves the beam from one side of the screen, say
the left side, and then reach the right side, it has to bring the bean back to the left side
and repeat the process again and again. During the return time the beam is impeded
from the reaching the screen and therefore no curve are drawn (the beam is
blackened). The difficulty is that once on the left side the actual value of the input
signal may differ from the one which was used in the previous sweep. The result is that
the new curve will not be superposed to the previous one. The user may see then many
non-stable curves! To solve this problem we need to make sure that the ramp
generator always produces the ramp voltage at a same specified value of the input
voltage. This is done through the trigger system which is explained next.
6. The trigger system helps stabilizing the displayed curves. It does that by comparing the
input signal to a pre-set value (called trigger level value and set by the user using knob
7) and not generating a pulse signal, see Fig. 2-b, only when the input signal is equal to
the trigger level value. This pulse signal is fed to the ramp generator (see figure 1) and
causes it to start generating the ramp voltage (see figure 2) for a fixed time. This time
can be modified by the user to his/her convenience using knob 8. The benefit is that we
have, in this way, a stable display but this is so only at the price that some parts of the
signal are not plotted and are therefore lost. The displayed parts of the curve,
corresponding to the period where the ramp voltage is increasing, are pictured on
figure 2 by a bold green line. The lost parts of the signal are drawn with a thin blue line.
You can check out the effect of the trigger system by turning left and right the trigger
level knob (knob 7).
Figure 8: trigger system function
49
We now describe how we can use the Oscilloscope to measure a couple of quantities.
1. How to measure frequencies and amplitudes from the screen:
a. First of all, the displayed signal should be stable and enlarged to ease the
measurement process.
b. Then note that the frequency, which is the number of cycles per second, is
where , in seconds, is the period (that is the minimum amount of time
in which the motion repeats itself). The frequency is then measured in 1/seconds
also called hertz.
c. Make sure that the (sine) signal is symmetric about the horizontal time axis
(check this out by selecting the ground as input signal and shift the
corresponding horizontal line to coincide with the zero volt line).
d. The period is directly measured from the Oscilloscope screen:
where is the
distance, on the screen, spanned by one period (see figure 3) and is
the scale used on the time axis (Time/div knob setting). The frequency is then
e. Similarly one can measure the amplitude

. Here the
refers to the distance from zero volt to the maximum of
the signal (see Fig. 3). In this case also one has to make sure that the signal is
symmetric about the horizontal time axis. Alternatively one can measure the
distance from peak-to-peak and divide by 2:
. It is preferable to
use this second method because it reduces the relative error.
Figure 9: Period and Amplitude
2. How to measure the difference in phase between two signals:
a. First, make sure that the (sine) signal is symmetric about the horizontal time axis
(check this out by choosing the ground as input signal and shift the corresponding
horizontal line to coincide with the zero volt line).
b. The phase difference can be calculated using where is the period
of either signals and the time distance between the two signals (see Figure 4). It
50
was explained previously how one would proceed to measure the period. The time
can be measured in a similar way:
where refers to the distance between the two signals crossing
the same point, say zero volt, in the same direction (see Fig 4). Note that and
thus where you can read the frequency off the generator directly.
Figure 10: phase difference measurement
3 Equipment needed:
Oscilloscope, AC and DC generators, multimeter, BNC connectors25.
4.1 Familiarize yourself with the Oscilloscope
1. Link the names of the knobs listed in the table below to the set of numbers in figure 1.
You can help yourself by re-reading the introduction above.
Knob name on the Oscilloscope Knob number on figure 1

Y position shift
X position shift
focus
intensity
Trigger level
AC/DC/G switch
screen
Volt/div knob
Time/div knob
2. Setup a sine wave of about 1000 Hz on the generator and make it visible on the screen.
Change the amplitude on the generator and watch how the curve changes on the
oscilloscope. You can do the same for the frequency.
3. Try and play with other functional forms (e.g. triangular signal, etc.)
25
See http://en.wikipedia.org/wiki/BNC_connector
51
4. Turn left and right the trigger level knob and make sure you understand its effect on
the display. Watch, in particular, how the left end of the curve moves up and down
when you play with the trigger level.
4.2 Measurement of the frequency

1. Select the sinusoidal signal on the generator.
2. Vary the frequency on the function generator to take on the values in table 1. The
amplitude should be fixed in this part.
3. For each frequency, on the function generator, measure using the oscilloscope the
corresponding period . You can do that by reading the number of divisions
spanned by one full period and multiply it by the time scale . (This is explained in the
introduction).
4. Estimate the error on the generator frequency and the number of division. The error
of the scale is 5%.
5. Calculate the frequency and fill in table 1.
Generator Number of Time scale Period Frequency

frequency division
1000
3000
5000
7000
9000
4.3 Measurement of the amplitude

1. Select the sinusoidal signal on the generator.
2. Set the generator frequency to about 2 kHz.
3. Vary the amplitude on the generator so that you read, on the multimeter, the voltage
values listed in the table below. For each value read the corresponding amplitude on
the Oscilloscope. Do not forget to estimate the errors on your measurements.
Voltage RMS value on multimeter Voltage Amplitude value on Oscilloscope

(volts) (volts)
2
4
6
8
10
52
4.4 Calibration of the voltage on the Y axis
1. In this section you use the DC generator (instead of the AC generator). Do the proper
connections (talk to the instructor if in doubt).
2. Set a fix voltage value, 5 volt say, on the DC generator (use the multimeter to read this
value).
3. Tweak the calibration knob (the small knob on the top of the Volts/div knob) so that the
reading on the Oscilloscope gives you exactly 5 volts. This procedure is called
calibration: it consists to define a new scale, here the oscilloscope Volts/div scale, by
comparing it to a well established scale, here the multimeter scale).
4. Checking out the calibration: set exactly 8 volt on the multimeter and see if you can read
8 volts on the Oscilloscope without changing the calibration (i.e. without changing the
calibration knob position).
5 Data Analysis
5.1 Measurement of the frequency
1. Calculate the error on the period and then that on the frequency .
2. Plot, using Excel, versus and draw the error bars. Do not forget the units and
titles.
5. What are the expected values of the slope and intercept?
6. How do your measurements compare with these values? (Use the method).
5.2 Measurement of the amplitude

1. Plot, using Excel, versus and draw the error bars. Do not forget the units and
titles.
4. The theory tells us that for a sine wave we do have . How do your
measurements compare with the theory? (Use the method).
6 Further Questions
 Prove that for a sine wave we do indeed have .
53
Transformer
Experiment (8)

The aim of this experiment is to study some of the properties of the transformer.
2 Introduction
The transformer is used to step up or down an AC voltage. It is a device that is present almost
everywhere around us. Its functioning is based on Faraday’s law which says that a change of the
magnetic field flux trough a conducting wire loop creates a difference of potential given by26:
(1)
Where is the flux of the magnetic field through the wire loop of area ( the area
vector with magnitude and direction perpendicular to the area). Please note that it is the
change of the flux which creates the voltage difference, and that this can be achieved in two
very different ways: either by changing or by changing . In a transformer it is that changes
and it is done by the so called primary coil (see Fig. 1).
Figure 1: A transformer with its input and output coils and iron core
An ac current is sent in the primary coil which creates27 a changing magnetic field. A secondary
coil, placed nearby, sees the changing the flux and thus a voltage difference is created across
the ends of coil (this is faraday’s law as explained above). The role of the (laminated) iron core
is to make sure that almost all the magnetic field lines created by the primary coil pass through
26
See Giancoli chap. 29, P758.
27
Due to Biot-Savard law.
54
the secondary coil. If this is so then the flux per one loop is the same in both coils (primary
and secondary). This means that (using Faraday’s law):
and
Where and are the number of turns in the primary and secondary coils, respectively.
Diving one equation by the other we get:
(2)
This relation tells us how the secondary voltage is related to the primary voltage and shows
how one could by judicious of the number of turns in both coils step up or down a given input
voltage.
A similar relation can be obtained for the currents28:
3 Equipment needed
Power supply, transformer, rheostat, oscilloscope, ammeters.
4.1 Part 1: In this part we verify Eq.2: Change and measure .
1. Read off the transformer the number of turns in the primary and secondary coils and
.
2. Connect the circuit shown in the figure below. When connecting the oscilloscope leads
make sure you respect the order of inputs of the coils: if, say, the black lead of channel 1
is connected to the E end of the primary coil then the black lead of channel 2 must be
connected to the E end of the secondary coil. (similarly for the end A of the coils). If not
sure talk to the instructor.
28
Assume conservation of power in the transformer and replace in Eq.2.
55
Figure 2: Transformer circuit without load
3. Using the rheostat, which is used as voltage divider, set an input voltage of 1 volt
(read the value from the oscilloscope).
4. Read, from the oscilloscope, the value of the output voltage and fill the table below.
Estimate the error on your measurements.
Table 6:Vs vs Vp
Input voltage Output voltage

1
2
3
4
5
6
3. Measure the phase difference between the primary and secondary voltage.
4.2 Part 2: In this part we study the efficiency of the transformer in function
of a resistive load.
1. The efficiency of the transformer is defined as the ratio of the input power ( )
and the output power ( ): . We need therefore to measure the
input and output currents as well as the input and output voltages.
2. Connect the circuit shown in figure 2. Note that the rheostat, if there is one, has been
removed and that two ammeters, to measure the currents, and a resistive load have
been introduced. (Check it out the instructor).
56
Figure 3: Transformer circuit with a load
3. Give the load resistance the values listed in table 2, measure the input and output
currents and voltages and fill the table. Estimate the error on your measurements.
Table 7: Transformer efficiency
5
20
40
80
120
140
160
200
280
5 Data analysis
5.1 Part 1:
1. Plot, using Excel, versus for the data in table 1. Do not forget the units and titles.
4. Does your data represent the model given by Eq.2? Look at the fit quality variable .
5. Plot your data, error bars, fit line on a graph paper this time (only for this part).
6. Calculate the experimental and its error (using the slope and its error)
7. Calculate the predicted (theoretical) values of (use the values of and you have
read at the beginning of part 1).
8. Compare the measured value to the predicted value? Do they agree?
9. Does the phase difference measured at the end of part 1 makes sense to you? Explain.
57
5.2 Part 2:
1. Plot, using Excel, versus for the data in table 2. Do not forget the units and titles.
2. Find, from the graph, the resistance at which the efficiency is maximum and estimate the
error on it.
6 Further questions
1. Explain why the transformers are important in transporting electricity through very long
distances.
2. Explain briefly how the transformer is used in a car ignition mechanism (spark plug).
3. Given the circuit in figure 4, for which value of the resistive load the dissipated power,
by this load, is maximum? Hint: Calculate the power dissipated by the load in function of
, and and set the derivative with respect equal to zero ( and are
constant).
Figure 4: Voltage source with a resistive load
4. Why it is difficult to apply the same method to the transformer (in part 2 above)?
58
Low and High Pass Passive Filters
Experiment (9)

The purpose of this experiment is to study the frequency response of an RC and RL circuits.
2 Introduction
In this experiment we use an ac input voltage of the form where is the
maximum amplitude, the angular frequency and the phase at . The sinusoidal voltage (and
current) can be seen as the component of a rotating vector (also called phasor) with angular
frequency (see figure 1).
Figure 1: a sinusoidal voltage can be seen as the Y component of a rotating vector
Consider now the circuit in Fig. 2. The output voltage is given by:
(1)
since . This very recurrent formula is called the voltage divider rule.
59
Figure 2: R-R circuit
The gain is defined as the ratio between the maximum amplitude of the output and input
voltages:
For the circuit in Fig.2, we have (using Eq. 1):
(2)
We notice two important things from Eqs. 1-2 (you will check this quickly using the
oscilloscope):
 and are in phase (i.e. the angles inside the sine functions are equal at all times
(see Fig. 3). Thus the phase difference is zero ( )).
 The gain is constant, and doesn’t depend on the frequency of the input voltage .
The different voltages in figure 2 can represented by the following phasors diagram:
Figure 3: Phasors diagram of the R-R circuit. All voltages are in phase.
We next turn our attention to the RC circuit in figure 4.
60
Figure 4: RC circuit. A high pass filter
It can be shown, see footnote29, that if we take the current to be , then the
voltage across the capacitor is given by:
The voltage across is of course (Ohm’s law). We can clearly see that the
two voltages and are not in phase, i.e. the argument of cosine functions are not equal. In
fact leads by an angle of (or ). This is illustrated in figure 5 which also shows
(the sum of and : ).
Figure 5: RC phasors diagrams. leads by . leads by .
The maximum amplitude of is found using Pythagoras theorem (see Fig. 5):
The gain is then given by:
29
or . We divide both side of the last equation by and get where we used the current
definition . We get by integration: (because is constant). Now, for a sinusoidal
current . The integration gives where we
used . By analogy to Ohms law, is proportional to the impedance of the capacitor.
61
(3)
We can draw the following important conclusions:
 (which is ) and are not phase. Their phase difference

30
depends in fact on . ( and are the output and input voltage phases,
respectively).
 The gain does depend on the frequency of the input voltage .
 At low frequency, i.e. , the gain is zero. This means that the output voltage is
zero and thus at these frequencies no voltage passes to the next stage of the circuit.
Only at high frequencies a voltage will pass. Thus the name High Pass Filter.
Consider Now the LR circuit in figure 6.
Figure 6: LR circuit. A low pass filter
The voltage across the coil is given by (see footnote31):
We see here that the voltage across the coil leads the voltage across the resistor by (or
). This is shown in figure 7.
30
The phase difference can be calculated from the phasors diagram in figure 5: . Thus
.
31
The drop of potential across an inductor (coil) is given by Faraday’s law where is the electromagnetic
flux (through the inductor). The induced magnetic field is proportional to the current (Biot-Savard law) and so is
the flux. . Hence . For a sinusoidal current, , we get
. By analogy to Ohm’s law, is proportional to the impedance of the coil.
62
Figure 7: RL phasors diagram. lags by . lags behind .
In a similar way as we did for the RC circuit, we get the gain for the LC circuit:
(4)
The following properties hold for an LR circuit (to be compared for those of an RC circuit):
 (which is ) and are not phase. Their phase difference

32
depends in fact on .
 The gain does depend on the frequency of the input voltage .
 At high frequency, i.e. , the gain is zero. This means that the output voltage is
zero and thus at these frequencies no voltage passes to the next stage of the circuit.
Only at low frequencies a voltage will pass. Thus the name Low Pass Filter.
3 Equipment needed
AC power supply, resistors, capacitor, coil, oscilloscope.
4.1 Part 1: R-R circuit: measure the gain and phase difference.
1. Connect the R-R circuit shown in figure 2. Choose
2. Connect channel 1 of the oscilloscope to display and channel 2 to display the voltage
across . (Mind the ground!). Consult the instructor before switching the power on.
3. Choose the sinusoidal signal on the functional generator.
4. Vary the frequency ( ) as requested in table 1, measure the quantities

displayed there ( , , gain, phase difference ) and fill the table in. Do not forget
to estimate your errors. The phase difference can be find by the measuring the
32
The phase difference can be calculated from the phasors diagram in figure 5: . Thus
.
63
difference in time (from the oscilloscope) between the two signals ( and ) and
the period of either signal. Then
5. Do the gain and phase difference values make sense to you? (Hint: refer to the theory
above.).
Table 8: R-R circuit
100
10000
6. You can that the Gain, and phase difference, do not change with the change of
frequency. You can easily check this by eyes by varying the frequency on a wide range.
4.2 Part 2: RC circuit: measure the gain and phase difference.

1. Connect the RC circuit shown in figure 4. Choose and
2. Connect channel 1 of the oscilloscope to display and channel 2 to display the
voltage across . (Mind the ground!).
4. Vary the frequency ( ) as requested in table 2, measure the quantities
displayed there ( , , gain, phase difference ) and fill the table in. Do not forget
to estimate your errors.
Table 2: RC circuit
100
1000
3000
6000
10000
15000
4.3 Part 3: RL circuit: measure the gain and phase difference.

1. Connect the RL circuit shown in figure 6. Choose and read the value of from
the coil.
2. Connect channel 1 of the oscilloscope to display and channel 2 to display the voltage
across . (Mind the ground!).
3. Vary the frequency ( ) as requested in table 3, measure the quantities displayed
there ( , , gain, phase difference ) and fill the table in. Do not forget to estimate
your errors. (Do not forget to measure at the same time as and ).
64
Table 3: RL circuit
100
500
800
1200
2000
5000
5 Data analysis
5.1 Part 1: RR circuit
1. Plot, using Excel, versus and vs (on separated graphs) for the data in table 1. Do
not forget the units and titles.
2. Calculate and plot the error bars on the same graphs.
3. Is your Data consistent with the theory (for this part)? Explain.
5.2 Part 2: RC circuit

1. Plot, using Excel, versus and vs (on separated graphs) for the data in table 2. Do
not forget the units and titles.
3. Plot the theoretical gain (Eq. 3) (i.e. plot versus ). Use the values of and of your
circuit and values of listed in table 2 above.
5.3 Part 3: RL circuit

1. Plot, using Excel, versus and vs (on separated graphs) for the data in table 3.
Do not forget the units and titles.
3. Plot the theoretical gain (Eq. 4) (i.e. plot versus ). Use the values of and of your
circuit and values of listed in table 2 above.
6 Further questions
1. show that the gain for an RL circuit is given by .
2. Comment on the names low pass and high pass and make sure you understand their
meaning
65
RLC Resonant Circuit
Experiment (10)

The purpose of this experiment is to study the resonance properties of an RLC circuit.
2 Introduction
In the previous lab experiment (Experiment 7), we have seen that an RC circuit passes high and
stops low frequencies. The inverse was true for an RL circuit. One may then expect that the
combination of R, L and C would produce a circuit that passes only a range (or band) of
frequencies. Figure 1 illustrates qualitatively this idea.
Figure 1: A high pass plus a low pass filters produce a band pass filter
In this experiment we test and study the above proposition. For that we build the circuit shown
in figure 2:
Figure 2: RLC circuit
The corresponding phasors diagram is presented in figure 3. (You may need to revisit the
previous experiment to see how the different voltages are calculated).
66
Figure 3: Phasors diagram for the RLC circuit.
It’s easy to relate the amplitude of to those of , and using Kirchhoff voltages law
and figure 3. Kirchhoff law says that at each instant . Then
where means the maximum amplitude of . We calculate this in two steps:
1. because they are back-to-back. This is represented by the black

arrow on Fig.3.
2. We see that the result (phasor) of the previous step makes an angle with the .
Thus we can use Pythagoras theorem to get the maximum amplitude of :
. We know that, see previous lab, ,
and . Hence .
We are now ready to calculate the gain and phase difference for and RLC circuit:
 Gain:
(1)
 Phase difference: we have from Fig. 3 . thus
(2)
From the above formulas we can conclude the following RLC band filter properties:
 The gain depends on the frequency and shows a resonance feature (as was expected in
figure 1). The resonant angular frequency , for which the gain is maximum, is given by
solving the following equation (note that this is the condition for the gain to be
maximum):
67
which gives
or (3)
 The phase difference at resonance is zero ( ) (this is so because

=0). This feature provides us with a way to find from the oscilloscope: this
is the angular frequency at which and are in phase.
 The phase difference changes sign at resonance: it is negative at
frequencies below and positive for frequencies above . This is because at low
frequencies the RLC circuit acts like an RC circuit and hence the phase difference is
negative ( lags behind ). At high frequencies the RLC circuit behaves like an RL
circuit and therefore the phase difference is positive ( leads
3 Equipment needed
AC power supply, resistor, capacitor, coil, oscilloscope.
1. Connect the RLC circuit shown in figure 2. Take and . is fixed (by
the geometry of one given coil).
2. Connect channel 1 of the oscilloscope to display and channel 2 to display the
voltage across . (Mind the ground!). Consult the instructor before switching the power
on.
4. Vary the frequency ( ) and localize the resonant frequency (Remember that
at the input and output voltages are in phase. That is how you should find ).
5. Being at , read , and fill the first row of the table below.
6. Vary the frequency as requested in table 1, measure the quantities displayed there ( ,
, gain, phase difference ) and fill the table in. Do not forget to estimate your
errors. The phase difference can be find by the measuring the
difference in time (from the oscilloscope) between the two signals ( and ) and
the period of either signal. Then . Be careful about the fact that
changes sign at .
Table 9: RLC circuit
68
5 Data analysis
1. Plot, using Excel, versus and vs (on separated graphs) for the data in the table
above. Do not forget the units and titles.
3. Plot the theoretical gain and phase difference on the same graphs in item 1. Use Eqs. 1,2
and the values of , and used in your circuit. (use the same frequencies as in table 1
above).
4. Find the resonant frequency from the graphs and its error.
5. Compare your measured value of with the theoretical one (see Eq. 3).
6 Further questions
1. What is the maximum gain you have got in this experiment.
2. What is the maximum gain expected from the theory (see Eq. 2). (Remember that the
maximum gain occurs at the resonant frequency)
3. Why your measured maximum gain is much smaller than the expected (theoretical) one?
4. Can you think and explain what happens at resonance.
69
References
[1] Data Reduction and Error Analysis for the Physical Sciences, P. Bevington & D Robinson, 3ed
[2] An Introduction to Error Analysis: The Study of Uncertainties in Physical Measurements, J. R.

Taylor, 2nd ed.
70
Appendix A: How does a Vernier caliper work?
1 Construction of the Vernier caliper

The Vernier caliper, Fig. 1, has two scales: the main scale and the Vernier scale (which is
movable). It is more precise than a simple ruler (see paragraph 5 below). The trick is in the fact
that the Vernier scale is graduated in such a way that 10 of its graduation cover only 9
graduations of the main scale (this is true only for the caliper shown in Fig. 1, other types have
different segmentation. See the note by the end of this document). In other words, if a
graduation of the main scale covers 1 mm then a graduation from the Vernier scale covers only
mm (i.e. less than a millimeter). Because of this, when measuring nothing, the only
graduations on the Vernier scale which align with another on the main scale are the first and
the last one (i.e. 0 and 10).
Figure 11: Vernier caliper with its two scales
2 How to read the Vernier Caliper

Assume we are measuring a distance as shown on Fig.2. One can clearly see on the main (fix)
scale that this distance is between 1mm and 2mm. So, from the main scale, equals at least 1
mm. We can then use the Vernier scale to read the first decimal as follow. We find the first
graduation on the Vernier scale that coincides with one on the main scale. We can see that the
3rd graduation does so (see Fig. 2). The Vernier caliper reading rule says that the required extra
distance is to be read as mm (i.e. we divide the number of graduations on the Vernier scale,
here 3, by 10). The total distance is then mm mm. (Note that this is true
only when we use mm as unit).
71
Figure 12: Measuring an length with the Vernier
3 Rational Explanation
From Fig. 3 below we can see that where is easily read, from the main scale,
to be 1mm. To read we need to use both the main and the Vernier scale. We can see from
the figure that , where should be read from the main scale and from the
Vernier scale. It’s clear that mm and mm (remember that a graduation o the
Vernier scale is mm). Therefore mm and finally
mm (as found previously).
Figure 13: illustration of the way the Vernier works
4 Important Note
Not all Vernier calipers are constructed as described in paragraph 1. Perhaps the main
difference is in how the Vernier scale is related to the main scale. But the same rules and
explanations given above can be applied for the other calipers as well. Especially, one gets a
measurement in three steps:
1. One gets a rough value by reading the main scale (in mm).
2. One then gets the precision reading by finding the coincidence graduation, reading its
position on the Vernier scale and dividing the result by 10 (in mm).
3. Add together the results of the two previous steps.
72
5 How to estimate the error
We can estimate the error committed when using a Vernier caliper in the following way. We
notice first that the error comes in because we are unable to locate exactly the first graduation
on the Vernier scale that coincides with one on the main scale (see Fig. 4). Things being so we
can fairly assume that, in most cases, we would miss the coincidence graduation by at most one
graduation (on the Vernier scale). But we know that one graduation reading on the Vernier
scale corresponds to 0.1 graduation on the main scale (remember that we have to divide by 10
to translate a reading on the Vernier scale to a reading on the main scale (see paragraph 2).
Therefore the quoted error should be 0.1 x the smallest graduation on the main scale. Compare
this error with the one which we should quote if we have used the main scale only (i.e. a
normal ruler). This should be the smallest graduation. The Vernier caliper improves thus the
precision by 10 times.
figure 4: A case where it's difficult to pick up only one coincidence graduation
6 For practice
http://www.members.shaw.ca/ron.blond/index.html
73
Appendix B: Least-square fit
We will use the method of Least Square to fit our data to a straight line. The aim is to find the
line that represents best our measurements. For this we have to minimize the following
quantity:
(12)
where and are our measured data and uncertainties which correspond to the
independent measured variables . is the predicted (theoretical) value at . (We
will drop in what follows the scripts and ). As one can see the , pronounced chi-
square, quantity is the sum of the square of the distances, called residuals (see Fig. 1), between
the measured data points and the theoretical straight line . Therefore
minimizing the means requiring the residuals to be as small as possible and the straight line
which satisfies this requirement is the best fit line33. The minimization is, of course, with respect
to the two free parameters, and of the theoretical line.
33
Here is a deeper reason, given by the Maximum Likelihood method, why we do minimize the . We assume first
that the measurements are normally distributed (i.e. Gaussian) with Means situated on the best fit line
and Standard Deviations given by the measurements error. With this assumption, the
probability for making the observed measurement is . This follows from
the statistic interpretation of the errors (with the Gaussian distribution as parent). The probability for making all
the observed ’s is, following probability multiplication rule, the product of the individual probabilities
. The best fit line is the one that
gives the maximum probability This is the Maximum Likelihood method. But we can see that maximizing
, i.e. maximizing the exponential function, is equivalent to minimizing the
74
465
y = 9.4857x - 158.52
R² = 0.9535
455
445
435
Residuals
425
415
405
59 61 63 65 67
Figure 14: Data, best fit and residuals are shown. Errors bars are equal to 5 (i.e. )
At the minimum the partial derivatives should be equal to zero:
These two equations can be rearranged as
75
These are two linear simultaneous equations in the two unknown parameters and All the
other quantities are our measured (known) quantities. Finding the parameters and , by
solving these equations, means that we have found the best fit straight line . For
34
simplicity, we assume that all the quantities have the same error The solution is then
given by the following equations:
(13)
(14)
(15)
Note that the solutions do not depend on the error . This is true only when the error is
common to all as we have assumed. In calculating the parameters and , one may first
accumulate the four sums , , and and then insert them in these
equations.
Now we turn to calculating the errors on the fit parameters. Since and depend on the
which have the common error , we can calculate the errors on and by using the error
propagation method35 (we assumed that have negligible errors):
34
This is the case for our labs.
35
It means that the error on and come from the fact that we have errors on See also Eq.4.
76
We have used in this derivation and the fact that , ie the indexes
and are just dummy variables which we can label them as we wish.
Thus the error on (the intercept) is given by:
(16)
with a similar calculation, we get the error on (the slope):
(17)
Fit Example:
Suppose we have the following measurements (this is the data used in figure 1 above):
Table 10
X Y
60 413
61 415
62 433
63 436
64 453
65 456
77
We would like to calculate the slope , the intercept and the error on them using these
measurements and the formulae developed above. First, as suggested, let us calculate the
needed sums (as we see on the table, the number of measurement is equal to 6). You can
use Excel to do these summations:
Then the determinant is (see Eq. 15):
The intercept is (see Eq. 13):
The slope is (see Eq. 14)
Note that the values obtained here for and are the same values as those found by Excel (see
Fig. 1). Note also that we kept the same number of significant figures as done in Excel (Fig.1) to
compare our results with the result there.
We can continue on and calculate the errors on and . Note, on the graph of Fig.1, that the
common error (this is the error bar on the y axis):
The error on the intercept is (see Eq. 16):
78
Then
The error on the slope is (see Eq. 17):
Thus and
Note: The values of the errors and obtained here are not equal to the
36
ones obtained with Excel ( 1.0468 65.45) . The reason is that Excel does not use the
error stated in the error bars (i.e. ) but uses an error calculated from the spread of the
data points themselves using the following formula:
(18)
Let us calculate this in our case and try to find Excel results. The following table shows the
calculation needed steps: Starting from the values of and , we calculate the values of
, then , and finally .
60 413 410.622 2.378 5.655

61 415 420.108 -5.108 26.089
62 433 429.593 3.407 11.605
63 436 439.079 -3.079 9.481
64 453 448.565 4.435 19.671
65 456 458.050 -2.050 4.204
From Eq. 18 and the last column in the table above we have:
36
See the tutorial on how to calculate the errors using Excel.
79
We use this value in evaluating as follow:
1.095782
Which means
For we have:
Then
We can see that we have obtained, with the new , the
80

PHYS194 Manual PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

PHYS194 Manual PDF

Uploaded by

Copyright:

Available Formats

Physics-II Lab Manual

Physics for Engineering

© Abdelouahab Abdesselam (abdesselam12@gmail.com)

Measurement and Error Analysis............................................................................. 3

Measurement and Errors Analysis ........................................................................ 19

The Gaussian distributions .................................................................................... 21

Ohm’s law .............................................................................................................. 26

Appendix A: How does a Vernier caliper work? ..................................................... 71

Appendix B: Least-square Fit ................................................................................. 74

Phys-194 is a laboratory experimental course designed to accompany the theoretical

1 How do we Report a Measurement

2 The Meaning of Error

Figure 1: the true could be anywhere within the uncertainty interval

3 Why is it Important to Know the Error

Figure 2: using the errors in comparing two measurements

4 How do we Estimate the Error on a Directly Measured Quantity

4.1 Analog Devices:

4.2 Digital Devices

4.3 The Standard Deviation method.

c. Finally one calculates the error on the mean5.

4.4 Number without error

1. 45678 has an error 1 .

5 Statistical interpretation of the error

Figure 3:Gaussian distribution of repeated measurements

6 How do we Estimate the Error on an indirectly measured

Assume that is a function of the independent variables . This is symbolically written

function about the point in Taylor series:

contributions as negative ones). We thus find , which is Eq. 4.

Answer: The circumference is . The error on is, using ,

The substitution of these derivatives in (Eq. 4) gives

Answer: The mass of the compound object is . The error on is,

Then and . Replacing these derivatives in Eq.

Example: We have a rectangular surface of length and width .

given by 1.8 So we have

Answer: First the area is . The error is given by

Then . As before, substituting in Eq.4 and

where and are constants. The derivative with respect to is . Substituting

where and are constants.

. Replacing in Eq.4 and dividing by we get:

where and are constants.

Similarly one can show that if then

9. One can continue in a similar way with other kind of functions.

7.1 Counting significant figures

1. Zeros on the left are not significant:

a. 15734 has 5 significant figures

7.2 Operations with significant figures

(see error propagation above). This means that . On the other

8.1 Random errors

8.2 Systematic errors

8.3 Precision and accuracy

9 Comparing Two Measurements15

This can be written as:

1 Aim of the experiment

17.47468 17.47392 17.47616 17.47499 17.47257

4. Write properly the results of the following operations:

1. , and are constant.

5 Application of Error Propagation

1 Aim of the experiment

Figure 1: Gaussian distribution16

Figure 2: Statistical interpretation of the error

Table 1:Gaussian distributed counts

Trial Number Counts

Bin number bin column