Oscilloscopes

From Nanoelektronikk

Jump to: navigation, search

Contents

Oscilloscopes

The aim for this page is to describe how to set up the oscilloscopes (mainly those in the course-lab), to make good measurments of different kind.

Agilent 54622D Mixed Signal Oscilloscope

First, the organization of the front panel:

File:Front_Panel.jpg

In addition to the soft buttons near the screen, there are 8 different panes and one general purpose knob. (The panes are as follows: Horizontal, Run Control, Measure, Waveform, Trigger, File, Analog and Digital)

Default settings

To achieve default settings, use the button auto-scale. Note that this will change many settings in the oscilloscopescope, such as averaging, gain, etc.. Thus it can be wise to check upon your settings after using auto-scale.

Fourier transform

To open be able to view a Fourier transform of the incoming data you have to use the Analog->Math button, however there are a few things that should be noted before doing so:

The scope will need to have constant temperature to give consistent results. This may be noticeable when measuring signal strength using FFT on weak signals (<-50dBV). When doing fine measurments, the scope should be in use approximately 30 minutes (measuring and calculating) before consistency can be achieved.

There are different settings that should be applied when attempting to determine signal strength or signal frequency. From here on, i will focus on how to set up the scope to be able to find signal strength.


Adjusting Volts/Div outside the math menu (Ch1 or Ch2 menu) does change the FFT readout. You will need to switch back and forth between math and Ch1/2 menu to set the correct v/div setting that gives the highest signal strength without going out of the boundries of the screen.

NOTE: The way these oscilliscopes display waveforms, having a sample rate that 
is too low for seeing the shape of the signal may conceal peak to peak values, 
thus it is easy to use too much gain and get clipping. Clipping will result 
in a too low FFT max readout. 

Adjusting gain should be done without averaging to avoid clipping of data. 

To avoid having erronous gain, adjust the sample frequency to a frequency that 
allows you to see the shape of the signal waveform along with the FFT, before 
adjusting gain. When you have achieved a gain that is suitable (as much as 
possible having the signal peaks within the window), you can (if desirable) 
change the sample frequency back to the desired rate. Using averaging on a 
external-triggered signal, the sample rate usually may be left high enough to 
view the signal waveform.  

When dealing with weak signals (too weak for the scope to trigger on), there are two ways of using the scope to collect data for FFT:

  • Recording with an external synchronization trigger
  • Recording without an external trigger

Before we divide into the two methods, we set up the general FFT behavior:

  • Set up the basic FFT settings:
    • Math->FFT->Settings->
      • Source: (Probe) 1
      • Sample rate: (Use the leftmost knob in the "horizontal" pane. The sample rate should be at least four times the signal frequency, however using 40-100 times the signal frequency will allow you to view and verify the signal waveform)
      • Span: (Typically 20% of the signal frequency)
      • Center: (The signal frequency)
      • More FFT->
        • Scale: (Typically 5dBV)
        • Window: Flat Top
        • Offset: Adjust until you can see the signal or the noise floor. (Remember how you got here, you will probably set offset several times)
Note that setting the basic settings only will leave 
the scope to its previous (random) settings..

For example switching between averaging and normal mode 
may cause the noise floor to jump as much as 30dB.  

To avoid this kind of behavior, make sure that you 
set all the desired features as described below...

Finding signal strength using FFT with an external trigger

Using an external trigger (if possible) will, after some time give the most stable and repeatable results, however if the goal is to identifying peak intensity in a noisy signal, the other approach will be more desirable. One should also be careful to ensure that one is measuring the desired signal, rather than noise from the trigger signal. The clue by using an external trigger is that the oscilloscope then allows us to do an averaging of up to 16383 data sets. Each data set contains 2048 samples in this scope. Note that using this setting is to ask the scope to average over 33.5 million samples. When sampling at 200kHz, gathering all these data will take at least 167 seconds, (slightly less than 3 minutes), however results will be presented on the fly on the screen- but it may need some time to settle.

Note also that if a lower sample rate is not required, using 40-100 times the signal frequency (or more) as sample rate will allow you to view both the signal waveform and the FFT simultaneously. This is highly recommended to avoid setting a gain that is too high (described in the general FFT section above). However, using this high sample rates, averaging is paramount to achieve a good FFT results. When adjusting gain, averaging should be turned off, to avoid clipping of noisy data.

Since results are presented before reaching the specified number of data-sets, the only need for less than maximum is when you want to record data that are constantly changing, in which the need for time to settle should be carefully planned. Thus the setup for FFT signal strengt using external trigger is as follows:

File:Front_Panel.jpg

  • Make connections as follows:
    • Connect the signal to be measured using the probe in channel 1
    • Connect the trigger pulse directly to channel 2
  • Set up the oscilloscope as follows:
    • Trigger->edge->
      • channel: 2
      • Select rising edge
    • Trigger->Mode/Coupling: Normal (you may have to adjust trigger level as well)
    • Waveform->Display: Deselect persistent
    • waveform->Acquire->
      • Select Averaging
      • Avgs = 16383
    • Measure->Quick Meas->
      • Source: Math
      • Select Measure Max (This will immediately show the peak value of the curve as long as it is shown in the window)
        • (Note: if measure max is not showing among the soft button choices, pushing "Quick Meas" should do the trick)

Use RunControl->Run/Stop twice before starting data collection to ensure that you don't read out a mix of current and previous data.

Finding signal strength using FFT without an external trigger

This oscilloscope does not make continuous FFT-sampling, it will only use 2048 samples before doing FFT, thus it cannot measure averages without a trigger. If the signal to be measured is strong enough to trigger the scope, we will of course be able to do averaging, however much of the time we use FFT when the signal is too weak for triggering, thus we cannot use averaging creatively. However, we still want to use the averaging setting, collecting one data set only. One could easily believe that this would be the same as normal mode, in which it is not. The setup described here will typically be good for finding peak signal strengt in a weak, noisy signal.

File:Front_Panel_persist.jpg

  • Make connections as follows:
    • Connect the signal to be measured using the probe in channel 1
    • leave channel 2 disconnected
  • Set up the oscilloscope as follows:
    • Trigger->edge
      • channel 1
      • Select rising edge
    • Trigger->Mode/Coupling: Select Auto level or Auto
    • Waveform->Display-> Select Persistent
    • Waveform->Acquire->
      • Select Averaging
      • Avgs = 1
    • Measure->Cursors Select Y1, then move the cursor using the knob to measure peak signal strength

To clear the display between measurements, do Waveform->Display->Clear

Personal tools