See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/289587614

Seeing How Fitting Process Works

Article · January 2016

Source: arXiv

CITATION READS

1 98

2 authors:

Vera Montalbano Mauro Sirigu

Università degli Studi di Siena Liceo Scientifico "Giovanni da Castiglione", Castiglion Fiorentino (AR), Italy
128 PUBLICATIONS   237 CITATIONS    1 PUBLICATION   1 CITATION   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Building a Foucault Pendulum at school View project

All content following this page was uploaded by Vera Montalbano on 15 January 2016.

The user has requested enhancement of the downloaded file.

 Seeing How Fitting Process Works
Vera Montalbano and Mauro Sirigu
University of Siena, Italy

Abstract
A common problem in teaching Physics in secondary school is due to the gap in terms of
difficulty between the physical concepts and the mathematical tools which are necessary to
study them quantitatively. Advanced statistical estimators are commonly introduced only a
couple of years later than some common physical topics, such as e.g. the electronic circuit
analysis. Filling this gap with alternative methods appears to be opportune, in order to let the
students reach a full comprehension of the issue they are facing with. In this work we use a
smartphone camera and GeoGebra to propose a visual method for understanding the physical
meaning of a fitting process. The time constant of an RC circuit is estimated by fitting the
discharge curve of a capacitor visualized on the screen of an oscilloscope.

1 Introduction
In order to engage students in understanding the physical world by constructing and using
scientific models to describe, explain, predict, design and control physical phenomena, a central
role is played by the fit of data measured in a physics experiment. For developing insight into the
structure of scientific knowledge by examining how models fit into theories and for showing
how scientific knowledge is validated, students must be engaged in evaluating scientific models
through comparison with empirical data (Hestenes, 1997). The discover of a physics law behind
the observations needs the definition of a set of parameters, whose number denotes the
assumed complexity of the model. In order to investigate and constrain the number and the best
estimates of these parameters, a fit of the empirical data is mandatory.
The concept of best fit of a sample of data, how the parameters in the proposed function change
to obtain the best fit, the uncertainties related to these parameters are difficult to explain to high
school students because of their lacking of advanced mathematical tools. Moreover, many
attempts of showing examples can be too abstract or remain obscure because of the use of
software which does not show in an explicit way the fitting process. Linear or polynomial
regression can be performed, for example, by MS Excel functions or by the Data Tool of the Open
Source Physics Project1. Anyway, with this kind of approach, the fitting process can be perceived
like a kind of obscure black box.
Therefore, it appears to be necessary the development of new teaching techniques which can
overcome the gap between usable and required mathematical tools. We propose an example of
how this can be easily obtained with a manual parameter optimization by using a Dynamic
Geometry Environment (DGE in the following) such as GeoGebra (Mariotti, 2002).
This effective and active way of teaching how a fitting process works emerged in a course
focused on physics lab didactics in pre-service education of physics teachers. The proposal was
designed during the data analysis of an experiment on the RC circuit and is therefore intended to
be tested with high school students struggling with electronics and electrical engineering.

2 The discharge of a capacitor

A typical exercise when dealing with electric circuit analysis is to derive theoretically and
measuring experimentally the electric potential discharge curve of a capacitor in an RC circuit.
When the circuit is supplied by a square wave generator, in the case that all other electric

1 http://www.opensourcephysics.org/webdocs/Tools.cfm?t=Datatool
1
components are ohmic and for negligible resistance of the cables, the flow of electric charges in
the circuit is governed by the first Ohm’s Law:

where dQ is the charge flowing in the circuit within the time dt, Q is the charge on each plate of
the capacitor of capacity C and R is the resistance of the resistor.
By assuming that at time t = 0 the circuit is closed and the capacitor fully charged, the
differential equation (1) has solution

where in the time constant of the circuit. By deriving and using again the first Ohm’s
Law, this equation provides

which shows the exponential decay of the potential difference between the plates of the
capacitor (see Figure 1).

Fig. 1 - The potential difference between the plate of a capacitor in an RC circuit during a discharge. In purple, the
geometrical interpretation of the time constant of the RC circuit.

By deriving the latter equation and remembering that at time , the potential derivative is
graphically represented by the tangent of the angle between the t-axis and the curve V(t) in ,
we can also show that  can be graphically represented as the abscissa of the intersection point
between the t-axis and the tangent to the discharge curve.
The time constant is a fundamental parameter in characterizing the circuit response to a signal
and can be experimentally measured by sampling the discharge curve of the capacitor.

3 Concept and implementation

The use of a DGE for obtaining and processing data is very useful in this context because it
allows to achieve multiple goals at the same time. It can: speed up the sampling of the discharge
curve; let the students deepen their competences in analytic geometry and show some
unexpected applications of it, such as a simple method to understand how fitting process works.
Moreover, it can be a captivating way of conducting a physical experiment, since it makes use of
multimedia tools, and can also strengthen the students’ computer expertness.

3.1 GeoGebra
GeoGebra2 is a powerful, easy to learn and open-source software that provides a dynamic
geometry environment. This tool is (or should be) extensively used in mathematics teaching
(Hall and Chambleeb, 2013; ) to introduce the geometrical construction and mathematical

2 http://www.geogebra.org
2
proofs and to provide visual and non-destructive tools for analytic geometry (Hohenwarter and
Fuchs, 2005; Baccaglini-Frank and Mariotti, 2010; Leung, Baccaglini-Frank and Mariotti, 2013).
It also implements data sheets, advanced statistical analysis tools (Prodromou, 2014) and 3D
modeling (Oldknow and Tetlow, 2008). Within the purpose of this work, GeoGebra has been
used to bypass mathematical difficulties and to obtain a visualisation of the physical meaning of
fitting.
The discharge curve of the capacitor can be easily displayed on the screen of an oscilloscope
connected in parallel. With a smartphone it is possible to take a picture of the screen and import
it into GeoGebra. This procedure allows to sample the curve in a high number of points in very
little time: instead of visually taking measures directly on the screen by also taking into account
the scaling factor in the time and potential axes, in a few clicks we can shift the picture in
GeoGebra in order to match the grid of the Cartesian plane and define the units on the x and y-
axis. By simply clicking on the curve, we can create points which will sample it very quickly. We
can also define analytically an exponential curve to be superimposed to the picture. The
parameters defining this exponential curve can be regulated in order to maximize the match
with the experimental one. This procedure defines a manual and visual approach to the fitting
process, and also allows an interdisciplinary approach to a physical problem, by implementing
analytical geometry tools and also reinforcing computer competences.
A fine tuning of the parameters can also be performed by defining sliders as we will better
explain in the following.

3.2 Materials and methods

As usual in TFA, low budget instruments and electrical components were used in order to
reproduce the material commonly available in secondary schools and understand the main
experimental difficulties that can be faced during a lab lesson. In particular, the frequency of the
square signal from the wave generator and the shape of the exponential discharge curve of the
ceramic capacitor can be obtained by sampling the output of an analog oscilloscope (a
GWInsteak GOS-622G with a 20MHz band width and a sensitivity of on both channels),
added to the circuit as shown in Figure 2. An accurate data reading requires at least a sample of
10 measures of the differential potential as a function of time. Due to the little screen, the spread
of the light trace, the scaling factors and the lack of a fine grid, these measures can be in some
case a bit difficult.

Fig. 2 - Scheme of the RC circuit with a square wave generator and an oscilloscope in parallel to the capacitor (on the
left) and the experimental setup (on the right).

3.3 Procedure
After setting up the circuit as shown in the previous section, we produced a square wave with
the generator and visualized it on the screen of the oscilloscope. On the second channel we
visualized the discharge curve of the capacitor. We changed the horizontal and vertical shifts and
3
regulated the scaling factors on the x and y-axes in order to show the discharge curve entirely
and as wide as possible, in order to minimize the uncertainty on the sampling.
We took a picture with a smartphone from a distance of about 3 meters, i.e. much greater than
the screen size, thus reducing the fish-eye effect that could affect the reliability of the
measures. The picture was imported into GeoGebra, moved and resized to make the oscilloscope
grid aligned to the Cartesian axes and the discharge curve begin at t = 0. The sampling of the
discharge curve can be done by simply clicking on several points of the curve. The coordinates of
the points in the Cartesian plane correspond to the measure of time and potential difference
respectively.
Finally, we superimposed to the graph an exponential function defined analytically as in (3) with
a free parameter . is set to match the starting value of the potential difference just before the
start of the discharge phase. The time constant is defined like an Action Object Tool of GeoGebra,
a Slider tool, which can be changed manually within a certain range and with defined steps. By
changing its value, it is possible to visually understand the effects of different time constants on
the time response of the circuit.
The  of the circuit is measured by choosing the value which best matches the experimental
curve, as shown in Figure 3. This curve can be compared to the best fit automatically generated
by GeoGebra with the function RegExp, which performs an exponential regression applied to the
list of points that sample the discharge curve.
The uncertainty on  can be evaluated by simply moving the slider and finding the upper and
lower approximation to the discharge curve.

Fig. 3. Analysis of the discharge curve of a capacitor with GeoGebra. The picture of the discharge curve was imported in
GeoGebra and aligned to the grid of the Cartesian plane. The blue points represent the sampling of the curve. The black
curve shows the exponential regression of the experimental points, while in red we show the curves corresponding to
upper and lower limit to the time constant of the circuit (the respective values are shown by the sliders on top of the
image). The black line is the tangent to the best fit curve at and the length of the segment represents the best
value for in .

Little discrepancies could be noticed between the values of the slider  and the abscissa of the
intersection point between the x-axis and the tangent to the discharge curve at . The
exploration of the causes of these discrepancies can be an excellent opportunity for a deeper
4
discussion on modelling, fitting and evaluation of uncertainties. Discrepancies could be
ascribed to the residual distortion and misalignment of the picture to the grid of the Cartesian
plane or to the choice of a model with only one parameter in the fit. In fact, the initial point of the
curve was fixed without uncertainties. How this fact can be taken into account in the fitting
process and how it changes the final results could be proposed as a further task after the data
processing.

3.4 Data and Processing

A first measure of the frequency of the square wave can be directly obtained by the generator:

The semi period of the oscillation is indeed shown also by the screen of the oscilloscope:

The uncertainty on the frequency is given by

From the semi period we derive , which leads to a second measure of the
frequency:

which is consistent with the first one.

The measure of the time constant of the circuit is obtained both manually (by moving the slider
to match the experimental discharge curve with the analytic exponential curve) and
automatically by the statistical tool of exponential regression provided by GeoGebra. The tool
returns the best values for the two parameters and of a function

With our sampling points we get and , which corresponds to an

estimated time constant .
To obtain an estimate of the uncertainty we define two exponential curves

By moving the sliders and we find the two curves that envelope the experimental one.
We thus get and t and define

We then finally obtain the best estimate for our time constant: .

4 Remarks and Conclusions

We provided an example of how the use of GeoGebra can be of help in sampling and analysing
data in a physics experiment. We obtained a measure of the time constant of a RC circuit by
importing a picture of the oscilloscope screen in GeoGebra, bypassing every difficulty related to
the lack of mathematical tools and then comparing the results with the automatic fitting tools
included in the software.
The method offers the opportunity to better explain the meaning of mathematical concepts such
as the derivatives, the differential equations and the statistical estimators both from a geometric
and a physical point of view.
This experiment can also be useful to stimulate curiosity and strategic and critical approach to a
physical experiment both reinforcing manual and computer abilities.
The use of GeoGebra can show how mathematics and geometry can be used in an unusual way to
solve a physical problem.
5
 The example of visualization of the fitting process described in the previous section reported the
context in laboratory in which the method emerged and was initially performed. Since the ease
and the range of situation in which it can be applied a more effective use of GeoGebra in data
analysis can be suggested. In the same context, a more sophisticated data fitting could involve
the fit of all the parameters of the model.
Moreover, the analysis of electric circuits is not the more effective moment in which introduce
the fitting process for at least two reasons. Students are introduced to electricity later in their
study in physics, usually after mechanics and thermodynamics. But many experiments in these
previous topics can be better understood if the fitting process is already introduced and well-
established. Furthermore, the fit of an exponential curve is not the best way for introducing the
visualization. Students are more accustomed to manipulating straight lines and interpreting
proportional laws. Thus, the next step in progress for implementing the visual fitting is to design
a learning path on an issue of mechanics and test it in class.

References
Baccaglini-Frank, A., & Mariotti, M. A. (2010). Generating conjectures in dynamic geometry: The maintaining
dragging model. International Journal of Computers for Mathematical Learning, 15, 225–253.
Hall, J., Chambleeb, G. (2013). Teaching algebra and geometry with GeoGebra: Preparing pre-service teachers for
middle grades/secondary mathematics classrooms. Computers in the Schools: Interdisciplinary Journal of
Practice, Theory, and Applied Research, 30, 12-29.
Hestenes, D. (1997). Modeling methodology for physics teachers. In AIP Conference Proceedings, 935-958. IOP
Institute of Physics Publishing LTD6.
Hohenwarter, M., & Fuchs, K. (2005). Combination of dynamic geometry, algebra and calculus in the software
system GeoGebra. In Computer Algebra Systems and Dynamic Geometry Systems in Mathematics Teaching
Conference 2004.
Leung, A., Baccaglini-Frank, A., & Mariotti, M. A. (2013). Discernment of invariants in dynamic geometry
environments. Educational Studies in Mathematics, 84(3), 439-460.
Mariotti, M. (2002). The influence of technological advances on students mathematics learning. In English, L.,
editor, Handbook of International Research on Mathematics Education, 695–723. Lawrence Erlbaum
Associates.
Oldknow, A., & Tetlow, L. (2008). Using dynamic geometry software to encourage 3D visualization and
modelling. The Electronic Journal of Mathematics and Technology, 2(1), 54-61.
Prodromou, T. (2014). GeoGebra in teaching and learning introductory statistics. Electronic Journal of
Mathematics & Technology, 8(8), 363-376.

View publication stats