Constructor University Bremen

GEE-2. Lab

Spring Semester 2024

Lab Experiment - Filter

Professor: Uwe Pagel

Students: Farid Mammadov and Naanlep Excel Tanko
Author: Farid Mammadov
Rotation 2
Date of execution: 14 March, 2024
Location: Room 54, Research 1, Constructor University, Bremen
Introduction:
In this experiment we’ll delve into designing circuits that could block a certain range of
frequencies while passing other frequencies. We’ll design these filters particularly using
capacitors to make so-called RC filters. There are 2 parts in this experiment: first with a
low-pass filter and second with a band pass filter, where we combine high and low-pass filters.
In the evaluation part we’ll draw various plots from measured and nominal values and then
compare and discuss errors.
Plots that we are particularly talking about are Bode and Nyquist plots. Bode plot is a graph
where we have magnitude in dB(which stands for change in ratio of output and input voltage
represented in logarithmic scale) as a function of frequency(frequency is also presented in
logarithmic scale, that is it increases by decade); additionally, we have Bode phase plot which is
similar to magnitude plot but instead we have phase as a function of frequency(again in
logarithmic scale). Apart from that we also have Nyquist plot which is a graph in Cartesian
system of coordinates where we take the real part of impedance on x-axis and imaginary part
on y-axis(we can also repeat this graph for complex conductance).
Filter is a circuit element designed to pass certain frequencies while blocking others. Filters are
commonly used to shape the frequency response of a circuit by allowing desired frequencies to
pass through while reducing unwanted frequencies.

Experimental Set-up and Results:

Starting the experiment with the first part we should design a RC low pass filter and illustrate its
characteristics on a Bode plot. We assemble the following circuit:

We have a generator as a source in sin mode and connect oscilloscope to input and output
using different channels to analyze phase shift and peak to peak change using measure
function. We input various frequencies to analyze frequency response.
 Frequency(Hz) Vin Vout Phase shift(deg)

50 10,1 10,1 0,36

100 10,1 10,1 -2,16

200 10,1 10,1 -3,17

500 10,1 10 -5,22

1000 10,1 9,84 -12,9

2000 10 9,2 -23,6

5000 10 6,8 -46,2

10000 10 4,24 -65,2

20000 10 2,32 -77,1

50000 10 0,952 -83,6

100000 10 0,488 -86,4

From the measured values above we can observe that relatively low frequencies are not
affected by filter and there is a little phase shift which is mainly caused by the nature of the
circuit. Since our filter consists of capacitor, the nature of the capacitor is such that at higher
frequencies its reactance gets smaller allowing more current to pass through which is causing
voltage to lag behind current which as consequence causes negative phase shift.

Second part of the experiment consisted of a band-pass filter which we assembled mainly by
combining low and high pass RC filters.
We connected the biggest R and C values to the low pass part(R2 and C2) and smallest ones to
the high pass part. Our overall circuit will look as follows:
In this circuit high frequencies will short C1 and C2 which will cause them to be sent into ground
(since C2 is shorted voltage across it Vout will be zero), and low frequencies will be basically
blocked by C1. In our circuit we assembled R1=8.2kΩ, R2=10kΩ, C1=1.5nF, C2=100nFand
again Vin will be powered by generator with 5Vpp, no offset, sine function settings.

Frequency(Hz) Vin Vout Phase shift(deg)

50 10,1 3 73,1

100 10,1 5,24 57,9

200 10,6 7,84 41,1

500 10,6 9,44 17

1000 10,6 9,76 1,44

2000 10,6 9,76 -2,6

5000 10,6 9,12 -19,9

10000 10,6 7,76 -41,1

20000 10,6 5,44 -56,9

50000 10 2,48 -74,4

100000 10 1,28 -80,9

As we can see from the table, since filter is band-pass, only mid-frequencies passed without
significant change in phase and voltage. But if we analyze border frequencies we’ll see that low
frequencies were significantly reduced(blocked); thus, their output voltage was lower and phase
shift was positive because of reactance of capacitor at low frequencies is high meaning it
doesn’t let current to pass through which in turn make current to lag behind voltage(=> positive
phase shift). At high frequencies we have similar situation to low pass filter’s response to such
frequencies.

Evaluation:
In part 1, as described in the introduction, we’ll draw Bode plots to describe the behavior of
circuit, and we’ll draw plots with measured and theoretical values.

To find dB, first I find magnitude(ratio) of transfer function which in our case will be ratio of
output and input voltages and then we convert that ration into dB by applying formula:
AdB=20*log10(Aratio) where Aratio=Vout / Vin
 Frequency(Hz) dB Phase shift

50 0 0,36

100 0 -2,16

200 0 -3,17

500 -0,086427 -5,22

1000 -0,226526 -12,9

2000 -0,724243 -23,6

5000 -3,349822 -46,2

10000 -7,452683 -65,2

20000 -12,69024 -77,1

50000 -20,42726 -83,6

100000 -26,2316 -86,4

Plots above are made from measured values, and indeed the graphs describe circuit letting only
low frequencies to pass and as frequency increases less voltage passes and more does voltage
lag behind current.
Now for theoretical values, where we actually used formulas for amplitude and phase using
nominal values such as R,C,f.

Where A is ratio of voltages and ⍵=2𝜋*f. After we find ratios we convert them to dB and have
the following table of values.
 Frequency(Hz) dB Phase shift

50 -0,000467 -0,593979

100 -0,001867 -1,18783

200 -0,007462 -2,374639

500 -0,046429 -5,918855

1000 -0,18281 -11,71401

2000 -0,689157 -22,52331

5000 -3,169762 -46,03303

10000 -7,242103 -64,25256

20000 -12,59995 -76,44231

50000 -20,3535 -84,49043

100000 -26,34397 -87,23883

Which gives us following plots:

We can see from tables and bode plots that dB values and phase shift values align almost
perfectly together

Now to calculate -3dB frequency and ⍵ we simply plug in the values of R and C in the formula
below

and to find ⍵ we simply use formula ⍵=2𝜋*f.

To get f and ⍵ from measured value we’ll analyze tables and look if phase shift is close -45° or
magnitude is close to -3dB.
After calculating f and ⍵ using formulas, we found that f=4822.877Hz and ⍵=30303.03 rad/s
And if we look into table(graph also) we find that 5000Hz has around -45° phase shift and
magnitude of -3,349822dB which are the closest values we have to ideal ones. Also, if we
calculate ⍵ for 5000Hz we get that ⍵=31415.927 rad/s which is relatively close to our theoretical
value.

To find gradient of A we must divide difference of first and last magnitude by difference of log
−7.452683+0.226526
first and log last frequency, so it’ll be grad A= 𝑙𝑜𝑔(10000)−𝑙𝑜𝑔(1000) = -7,226157 dB/decade.
Finally, we analyse limits of amplitude ratio in dB and phase for very high, very low or very close
to cutoff frequency. For f<<f-3dB we have that amplitude ratio will tend to 1 and phase will tend to
zero; ratio tends to 1 because for low pass, low frequencies don’t affect output voltage meaning
ratio will be 1 and log(1)=0, and since circuit doesn’t affect our voltage, input and output should
be same meaning there is no phase difference.
For f>>f-3dB amplitude ratio will tend to 0(from positive side) and phase will be close to -90°; ratio
will tend to 0 because Vout will be almost 0 at high frequencies which means that we have log of
very small positive number and if we plot standard graph of log we’ll see that for very small
positive number under log, log will go to minus infinity; additionally, phase will be -90° because
capacitor will just let through all the current due to low reactance at HF which will make voltage
lag behind current.
For f=f-3dB we have that amplitude ratio in dB will tend to -3 and phase to -45°; ratio tending to -3
1
is explained by the fact that at cutoff frequency amplitude ratio as scalar will equal whose log
2
equals -3dB, and phase difference of -45° is explained by the fact that if we analyze the transfer
function and represent it in complex number form we’ll have that this complex number has an
angle of -45°.

In part 2, we are asked to draw Bode magnitude and phase plots so we need to calculate
magnitudes in dB from given values:

Magnitude(voltage amplitude ratios) In dB

0,2970297 -10,544

0,5188119 -5,699802

0,7396226 -2,619796

0,890566 -1,006677

0,9207547 -0,717121

0,9207547 -0,717121

0,8603774 -1,306221

0,7320755 -2,708883

0,5132075 -5,794139

0,248 -12,11097

0,128 -17,8558

Again AdB=20*log10(Aratio) where Aratio=Vout / Vin

And now plotting this values in excel using the frequencies we had on log scale we get:
Now we need to draw bode plots for measured magnitude and phase shift. To find both of them
we’ll need to analyze the transfer function of the band pass filter(product of transfer functions of
low and high pass filters which themselves construct the band pass itself). H(jw)=
𝑅1*𝐶1 1
1 1 (1+ 𝑅2*𝐶2 )−𝑗(𝑤*𝑅1*𝐶1− 𝑤*𝑅2*𝐶2 )
1+𝑗𝑤*𝑅1*𝐶1
* 1 =
𝑅1*𝐶1 2 2 now we need to derive
1+ 𝑗𝑤*𝑅2*𝐶2 (1+ 𝑅2*𝐶2 ) +(𝑤*𝑅1*𝐶1−
1
)
𝑤*𝑅2*𝐶2
expressions for real and imaginary parts(I also did some cancellations):
𝑅1*𝐶1 1
(1+ 𝑅2*𝐶2 ) −(𝑤*𝑅1*𝐶1− 𝑤*𝑅2*𝐶2 )
Re{H}= and Im{H}= I
𝑅1*𝐶1 2 1 2 𝑅1*𝐶1 2 1 2
(1+ 𝑅2*𝐶2 ) +(𝑤*𝑅1*𝐶1− 𝑤*𝑅2*𝐶2
) (1+ 𝑅2*𝐶2 ) +(𝑤*𝑅1*𝐶1− 𝑤*𝑅2*𝐶2
)
should remind that R1=8.2kΩ, R2=10kΩ, C1=1.5nF, C2=100nF and we have values of w. Now
to get expressions for magnitude and phase shift we simply find magnitude and angle of this
2 2 𝐼𝑚
complex number. Magnitude= 𝑅𝑒 + 𝐼𝑚 and phase shift=arctan( 𝑅𝑒 )

Frequency Magnitude Magnitude in Phase shift Real part Imaginary

dB part

50 0,000211273 -73,503102 72,338007 6,41006E-05 0,00020131

100 0,000476011 -66,447663 57,415301 0,000256354 0,00040108

200 0,001293718 -57,763211 37,626358 0,001024636 0,00078983

500 0,006608688 -43,597695 15,443888 0,006370059 0,00175986

1000 0,025086251 -32,011285 4,6238453 0,025004606 0,0020223

2000 0,093180342 -20,613514 -4,236603 0,092925726 -0,0068837

5000 0,393096482 -8,1100168 -19,30416 0,370995403 -0,129951

10000 0,634468581 -3,9517976 -36,78618 0,50813058 -0,3799391

20000 0,527821989 -5,5502504 -56,64236 0,29023001 -0,4408657

50000 0,250285927 -12,031271 -75,30844 0,063476369 -0,2421029

100000 0,128315963 -17,834386 -82,53603 0,016668588 -0,1272287

Now plot:
From the plots we can see that phase shift plot and values align with experimental(measured)
values, but for magnitude plot we can still observe band pass behavior because plot peaks in
the middle; however, the values don’t align as much for low and middle frequencies.
We can see that indeed our circuit mostly lets middle frequencies to pass and blocks high and
low ones, and from phase plot we see that at low frequencies current lags behind voltage which
is the case because capacitors have large reactance but at higher frequencies current will lead
over voltage because there capacitors have close to 0 reactance letting current to just pass
through.

R1=8.2kΩ, R2=10kΩ, C1=1.5nF, C2=100nF

1 1
To find cutoff frequencies we use formula 2𝜋*𝑅*𝐶
and get that fH= 2𝜋*𝑅1*𝐶1 =12939,426Hz and
fL=159,155Hz from here we can directly find center frequency which will be geometric mean of fH
and fL so fcenter=1435.04667, also bandwidth will simply be difference of cutoff frequencies so
bandwidth=10416.24Hz. To find phase shift we should plug these values of frequencies to our
transfer function, get the complex number, and calculate the angle of this complex number.
0,0006491 + 0,000633326j at fL and 0,452739505 - 0,44173743j at fH and angles of these
complex numbers give us 44,29528695 and -44,29529684 degrees respectively.
Ideally, we should have gotten 45 and -45 degrees, so our values are very close.

If we look at the bode plot for measured magnitude to compare cutoff frequencies we’ll see that
at around -3dB the value of frequencies is indeed close to 200 and 10000. On our plot center
frequency is at around 1000Hz but we calculated it to be around 1435Hz, still close if we
consider on logarithmic axis.

To plot the Nyquist plot we’ll simply need to mark transfer function values for given frequencies
on a complex plane and then connect all these points. We already have the real and imaginary
part value in the table above and now we simply plot them on a complex plane. The values and
formulas used here can be found in the table above with values and formulas for Re{H} and
Im{H}.
Conclusion:
We had a chance to see how one could design filters, how they actually work(what physical
processes happen so that filters can block or let something to pass), how we can combine some
types of filters to get new filters with new properties. We also were able to use analytical
knowledge to calculate their expected behavior and compare them with measured ones using
plots. Finally, we found values that describe the type of filter we have and its properties(cutoff
frequency, bandwidth and center-frequency).

Overall, regarding the errors, only our theoretical values for magnitude of band pass filter are
odd, but we can still observe band pass behavior for both cases. Other values and plots align
and are quite close.

References:
1)file://01.0.generaleelab/01.2.generaleelab2/20240112-ch-211_manual.pdf
2)Filter Capacitor (learningaboutelectronics.com)