diff --git a/figs/apa95ml_5kg_PI_pdf_comp_huddle.pdf b/figs/apa95ml_5kg_PI_pdf_comp_huddle.pdf index bc021ab..9188194 100644 Binary files a/figs/apa95ml_5kg_PI_pdf_comp_huddle.pdf and b/figs/apa95ml_5kg_PI_pdf_comp_huddle.pdf differ diff --git a/figs/apa95ml_5kg_PI_pdf_comp_huddle.png b/figs/apa95ml_5kg_PI_pdf_comp_huddle.png index 4dd3d2d..52993eb 100644 Binary files a/figs/apa95ml_5kg_PI_pdf_comp_huddle.png and b/figs/apa95ml_5kg_PI_pdf_comp_huddle.png differ diff --git a/figs/huddle_test_pdf.pdf b/figs/huddle_test_pdf.pdf index 0ef9fc1..4c513b3 100644 Binary files a/figs/huddle_test_pdf.pdf and b/figs/huddle_test_pdf.pdf differ diff --git a/figs/huddle_test_pdf.png b/figs/huddle_test_pdf.png index 7b0fb87..ec5239a 100644 Binary files a/figs/huddle_test_pdf.png and b/figs/huddle_test_pdf.png differ diff --git a/figs/huddle_test_time_domain.pdf b/figs/huddle_test_time_domain.pdf index f4ff0ff..ce03bff 100644 Binary files a/figs/huddle_test_time_domain.pdf and b/figs/huddle_test_time_domain.pdf differ diff --git a/figs/huddle_test_time_domain.png b/figs/huddle_test_time_domain.png index 1805184..73f5341 100644 Binary files a/figs/huddle_test_time_domain.png and b/figs/huddle_test_time_domain.png differ diff --git a/index.html b/index.html index 50e9865..8c471da 100644 --- a/index.html +++ b/index.html @@ -3,7 +3,7 @@ "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> - + Test Bench APA95ML @@ -27,42 +27,42 @@

Table of Contents

@@ -70,26 +70,26 @@
-
+

setup_picture.png

Figure 1: Picture of the Setup

-
+

setup_zoom.png

Figure 2: Zoom on the APA

-
-

1 Setup

+
+

1 Setup

-
-

1.1 Parameters

+
+

1.1 Parameters

Ts = 1e-4;
@@ -98,8 +98,8 @@
-
-

1.2 Filter White Noise

+
+

1.2 Filter White Noise

Glpf = 1/(1 + s/2/pi/500);
@@ -111,13 +111,13 @@ Gz = c2d(Glpf, Ts, 'tustin');
-
-

2 Run Experiment and Save Data

+
+

2 Run Experiment and Save Data

-
-

2.1 Load Data

+
+

2.1 Load Data

data = SimulinkRealTime.utils.getFileScopeData('data/apa95ml.dat').data;
@@ -126,8 +126,8 @@ Gz = c2d(Glpf, Ts, 'tustin');
-
-

2.2 Save Data

+
+

2.2 Save Data

u = data(:, 1); % Input Voltage [V]
@@ -144,16 +144,16 @@ t = data(:, 3); % Time [s]
-
-

3 Huddle Test

+
+

3 Huddle Test

-
-

3.1 Time Domain Data

+
+

3.1 Time Domain Data

-
+

huddle_test_time_domain.png

Figure 3: Measurement of the Mass displacement during Huddle Test

@@ -161,8 +161,8 @@ t = data(:, 3); % Time [s]
-
-

3.2 PSD of Measurement Noise

+
+

3.2 PSD of Measurement Noise

Ts = t(end)/(length(t)-1);
@@ -173,12 +173,12 @@ win = hanning(ceil(1*Fs));
-
[pxx, f] = pwelch(y, win, [], [], Fs);
+
[pxx, f] = pwelch(y(1000:end), win, [], [], Fs);
-
+

huddle_test_pdf.png

Figure 4: Amplitude Spectral Density of the Displacement during Huddle Test

@@ -187,16 +187,16 @@ win = hanning(ceil(1*Fs));
-
-

4 Transfer Function Estimation using the DAC as the driver

+
+

4 Transfer Function Estimation using the DAC as the driver

-
-

4.1 Time Domain Data

+
+

4.1 Time Domain Data

-
+

apa95ml_5kg_10V_time_domain.png

Figure 5: Time domain signals during the test

@@ -204,8 +204,8 @@ win = hanning(ceil(1*Fs));
-
-

4.2 Comparison of the PSD with Huddle Test

+
+

4.2 Comparison of the PSD with Huddle Test

Ts = t(end)/(length(t)-1);
@@ -222,7 +222,7 @@ win = hanning(ceil(1*Fs));
-
+

apa95ml_5kg_10V_pdf_comp_huddle.png

Figure 6: Comparison of the ASD for the identification test and the huddle test

@@ -230,8 +230,8 @@ win = hanning(ceil(1*Fs));
-
-

4.3 Compute TF estimate and Coherence

+
+

4.3 Compute TF estimate and Coherence

Ts = t(end)/(length(t)-1);
@@ -248,14 +248,14 @@ Fs = 1/Ts;
-
+

apa95ml_5kg_10V_coh.png

Figure 7: Coherence

-
+

apa95ml_5kg_10V_tf.png

Figure 8: Estimation of the transfer function from input voltage to displacement

@@ -263,8 +263,8 @@ Fs = 1/Ts;
-
-

4.4 Comparison with the FEM model

+
+

4.4 Comparison with the FEM model

load('mat/fem_model_5kg.mat', 'Ghm');
@@ -272,7 +272,7 @@ Fs = 1/Ts;
-
+

apa95ml_5kg_comp_fem.png

Figure 9: Comparison of the identified transfer function and the one estimated from the FE model

@@ -291,12 +291,12 @@ In the next section, a current amplifier is used.
-
-

5 Transfer Function Estimation using the PI Amplifier

+
+

5 Transfer Function Estimation using the PI Amplifier

-
-

5.1 Comparison of the PSD with Huddle Test

+
+

5.1 Comparison of the PSD with Huddle Test

Ts = t(end)/(length(t)-1);
@@ -313,7 +313,7 @@ win = hanning(ceil(1*Fs));
-
+

apa95ml_5kg_PI_pdf_comp_huddle.png

Figure 10: Comparison of the ASD for the identification test and the huddle test

@@ -321,8 +321,8 @@ win = hanning(ceil(1*Fs));
-
-

5.2 Compute TF estimate and Coherence

+
+

5.2 Compute TF estimate and Coherence

Ts = t(end)/(length(t)-1);
@@ -340,14 +340,14 @@ Fs = 1/Ts;
-
+

apa95ml_5kg_PI_coh.png

Figure 11: Coherence

-
+

apa95ml_5kg_PI_tf.png

Figure 12: Estimation of the transfer function from input voltage to displacement

@@ -355,8 +355,8 @@ Fs = 1/Ts;
-
-

5.3 Comparison with the FEM model

+
+

5.3 Comparison with the FEM model

load('mat/fem_model_5kg.mat', 'Ghm');
@@ -364,7 +364,7 @@ Fs = 1/Ts;
-
+

apa95ml_5kg_pi_comp_fem.png

Figure 13: Comparison of the identified transfer function and the one estimated from the FE model

@@ -372,12 +372,12 @@ Fs = 1/Ts;
-
-

6 Transfer function of the PI Amplifier

+
+

6 Transfer function of the PI Amplifier

-
-

6.1 Compute TF estimate and Coherence

+
+

6.1 Compute TF estimate and Coherence

Ts = t(end)/(length(t)-1);
@@ -390,11 +390,11 @@ The coherence and the transfer function are estimate from the voltage input of t
 

 
 

-The coherence is very good as expected (Figure 14).
+The coherence is very good as expected (Figure 14).
 

 
 

-The transfer function show a low pass filter behavior with a lot of phase drop (Figure 15).
+The transfer function show a low pass filter behavior with a lot of phase drop (Figure 15).
 

 
 

@@ -406,37 +406,42 @@ The transfer function show a low pass filter behavior with a lot of phase drop (
 

 
 
-

+

 
PI_E505_coh.png
 

 
Figure 14: Coherence

 

 
 
-

+

 
PI_E505_tf.png
 

 
Figure 15: Estimation of the transfer function from input voltage to displacement

 

 
 

-The delay can be estimated as follow:
+The delay can be estimated as follow (in ms):
 

 

-
finddelay(u, um)*Ts
+
finddelay(u, um)*(1000*Ts)
 

 

 
 
-0.0004
+0.4
 

+
+
+

+This most probably corresponds to a FIR filter.
+

Author: Dehaeze Thomas

-

Created: 2020-07-24 ven. 13:16

+

Created: 2020-07-24 ven. 15:48

diff --git a/index.org b/index.org index 6e249b5..1f9d00c 100644 --- a/index.org +++ b/index.org @@ -108,13 +108,17 @@ ** Load Data :noexport: #+begin_src matlab - load('./mat/huddle_test.mat', 't', 'u', 'y'); + load('./mat/huddle_test.mat', 't', 'y'); +#+end_src + +#+begin_src matlab + y = y - mean(y(1000:end)); #+end_src ** Time Domain Data #+begin_src matlab :exports none figure; - plot(t, y) + plot(t(1000:end), y(1000:end)) ylabel('Output Displacement [m]'); xlabel('Time [s]'); #+end_src @@ -136,7 +140,7 @@ #+end_src #+begin_src matlab - [pxx, f] = pwelch(y, win, [], [], Fs); + [pxx, f] = pwelch(y(1000:end), win, [], [], Fs); #+end_src #+begin_src matlab :exports none @@ -385,7 +389,7 @@ In the next section, a current amplifier is used. hold off; set(gca, 'XScale', 'log'); set(gca, 'YScale', 'log'); xlabel('Frequency [Hz]'); ylabel('ASD [$m/\sqrt{Hz}$]'); - legend('location', 'northeast'); + legend('location', 'southwest'); xlim([1, Fs/2]); #+end_src @@ -590,11 +594,12 @@ The transfer function show a low pass filter behavior with a lot of phase drop ( #+RESULTS: [[file:figs/PI_E505_tf.png]] -The delay can be estimated as follow: +The delay can be estimated as follow (in ms): #+begin_src matlab :results replace value - finddelay(u, um)*Ts + finddelay(u, um)*(1000*Ts) #+end_src #+RESULTS: -: 0.0004 +: 0.4 +This most probably corresponds to a FIR filter.