Keysight U8903A User manual

Keysight Technologies
U8903A Audio Analyzer
Creating a Custom Filter for the
U8903A Audio Analyzer
Demo Guide

2
This demonstration will teach you how to create your own custom lter for the Keysight Technologies, Inc.
U8903A Audio Analyzer by using the MATLAB software on your computer.
Table of Contents
Demonstration Objectives 2
Step 1: Matlab fdatool
(lter design and analysis tool) 3
Step 2: Create Filter 4
Step 3: Export Filter
Coefcient 7
Step 4: Settling Time (Delay) 9
Step 5: Transfer to .juf le 11
S tep 6: Upload .juf le to
U8903A Audio Analyzer 13
Conclusion 14
Ordering Information and
Related Literature 15
Demonstration objectives
Steps to create a custom lter
Follow the steps below to create your own custom lter for the U8903A.
Steps
1. Matlab fdatool (lter design and analysis tool)
2. Create Filter
3. Export Filter Coefcient
4. Settling Time (Delay)
5. Transfer to .juf le
6. Upload .juf le to U8903A Audio Analyzer

3
Matlab fdatool (Filter Design and Analysis Tool)
Step 1
Figure 1. MATLAB command window
Figure 2. Filter Design and Analysis Tool
Instructions
1. Start MATLAB program (Figure 1)
2. In the command window, type “fdatool” to start Filter Design and Analysis Tool in MATLAB (Figure 2)

4
Create lter
Step 2
Figure 3. “Set quantization parameter” icon
Figure 4. Selecting “Single-precision oating-point” and “Return to the Design Filter page”
Instructions
1. Click on “Set quantization parameter” (Figure 3)
2. Select “Single-precision oating-point” because the architecture of the U8903A is 32 bits (Figure 4)
3. Return to the Design Filter page (Figure 4)

5
Instructions
4. Select the “Response Type” e.g. Lowpass, Highpass, Bandpass, and Bandstop (Figure 5)
5. Select the “Design Method” e.g. IIR or FIR (Figure 5)
6. Select the “Frequency Specications” you require e.g. Fpass (1 and 2), Fstop (1 and 2) (Figure 5)
Requirements for Frequency Specications:
– Low Bandwidth = 78125 Hz
– High Bandwidth = 312500 Hz
7. Select the “Magnitude Specications” e.g. Wpass (1 and 2), Wstop (1 and 2) (Figure 5)
8. Press the “Design Filter” button (Figure 5)
Figure 5. Selecting “Response Type”, “Design Method”, “Frequency Specications”, “Magnitude Specication”, and Press “Design Filter” Button

6
Instructions
9. The “Order” number (see Figure 6) must be less than 256 in order to meet the architecture of the U8903A Audio Analyzer. Changing
the “Design Method” will change the order number.
Figure 6. Order number must be less than 256

7
Export lter coefcient
Figure 7. Exporting Filter Coefcient Information to a le
Instructions
1. Go to “File”, then select “Export...”
2. At “Export To”, select “Coefcient File (ASCII)”
3. Click on “Export” button, and Save to *.fcf le
4. The Filter Coefcient Data File will then be displayed
Export lter coefcient information to a le (gure 7)
Step 3

8
Figure 8. Export Filter Coefcient as an object for analysis
Instructions
1. Go to “File”, then select “Export...”
2. At “Export To”, select “Workspace”
3. At “Export As”, select “Object”
4. At “Variable Names”, for “Discrete Filter”, enter “Hd”, then click the “Export” button
5. The Hd object is then in the MATLAB window
Export lter coefcient as an object for analysis (Figure 8)

9
Settling time (Delay)
Step 4
Knowing the settling time of your lter
You can know the settling time for your lter by creating a 20 Hz Sine Wave for testing the lter you just created using the
MATLAB software
Instructions
In the MATLAB command window
1. Type in the code below:
– fs = 78125
– f = 20
– n = 1:65536
– xn = sin(2*pi*f*n/fs)
These objects (fs, f, n, and xn) are now created in Matlab (Figure 9):
– fs (Frequency Sampling): 78125 for low bandwidth 312500 for high bandwidth
– f(Frequency): 20 Hz
– n(Points): 1 to 65536
– xn (20 Hz Sine wave)
2. Type in the code below to plot the 20 Hz sine wave:
– plot(xn,’DisplayName’,’xn’,’YDataSource’,’xn’);gure(gcf);
A 20 Hz Sine wave graph (Figure 10) is created.
Figure 9. MATLAB command window and objects
Figure 10. 20 Hz Sine wave

10
Step 4
Applying lter to the 20 Hz Sine wave
Instructions
In the MATLAB command window
1. Type in the code below to apply lter to 20 Hz sine wave (Figure 11) :
–l=lter(Hd,xn);
Objects lter is now created in MATLAB
2. Type in the code below to plot the result:
–plot(l,’DisplayName’,’l’,’YDataSource’,’l’);gure(gcf);
A 20 Hz Sine wave graph with the lter applied (Figure 12) is created.
3. By zooming into the graph you are able to search for the result (Figure 12)
Result after applying lter with 20 Hz Sine wave:
–Noise level will be ne after 4000 samples
Figure 11. MATLAB command window and objects
Figure 12. Zooming into the graph to see in-depth result