Thread Subject:

Determining Frequency Different/Drift

Hello,
The problem that I am trying to tackle is determining frequency/phase difference between two oscillators over time. My two oscillators are GPS locked oscillators (Trimble Mini-T), which each output a 10MHz signal. Looking at both signals on an oscilloscope (Tek MSO2024) I can see that there is indeed some drift between the two over time. They will move in and (slightly) out of phase over periods of time, but the changes are slow.

I have logged data (only 10,000 samples, if you can help me figure out how to take more, that'd be great!) using tmtool in the instrument control toolbox. I took essentially three data sets: some arbitrary phase difference between them, completely inphase, almost completely out of phase.

I have taken these data sets, calculated the FFT of each one, and plotted the results. In each case I get FFTs which indicate that the frequency is precisely 10MHz, for both oscillators, no matter what I do.

Any suggestions of how I could make this measurement? Or perhaps I'm doing something wrong? I will post my code below. I'd like to take longer datasets, but I cannot figure out how to change the "MaxNumberPoint" variable to 100,000. I've tried changing everything in the driver file and I can get my time samples (length(X)) up to 100,000, but never my actual data (X).

At any rate, any help would be much appreciated! Thanks!

Code:
close all; clear all; clc;
%% Load Captured data from oscilloscope
a = pwd;
d1 = load([a '\data\dataSet_1.mat']);
d2 = load([a '\data\dataSet_2_HandOnOscillator.mat']);
d3 = load([a '\data\dataSet_3_Overcorrection_InPhase.mat']);

Y1 = {d1.Y1, d2.Y1, d3.Y1}; %Channel 1 (Oscillator A)
Y2 = {d1.Y2, d2.Y2, d3.Y2}; %Channel 2 (Oscillator B)

t = d1.t1; %Only need one time set, they are all identical

%Define sampling frequency
Fs = 1e9; %Tek MSO2024 samples at 1GS/s


%% Generate FFT of sinusoids
%generate zero-pad data, making length a power of two

for n = 1:length(Y1)
    nfft = 2^(nextpow2(length(Y1{n})));

    %Take FFT
    fft_y1{n} = fft(Y1{n}, nfft);
    fft_y2{n} = fft(Y2{n}, nfft);

    %Take the positive portion of the FFT
    [fft_y1_pos{n},freqRange{n}] = positiveFFT(Y1{n},Fs);
    [fft_y2_pos{n},freqRange{n}] = positiveFFT(Y2{n},Fs);

    %Plot the FFT
    subplot(3,1,n)
    plot(freqRange{n}/(1e6), abs(fft_y1_pos{n}));
    hold on;
    plot(freqRange{n}/(1e6), abs(fft_y2_pos{n}), 'r');
    grid on;
    xlabel('Freq (MHz)');
    xlim([0 20]);
    hold off;
end

"Chris" wrote in message <jvvb9c$m85$1@newscl01ah.mathworks.com>...
> Hello,
> The problem that I am trying to tackle is determining frequency/phase difference between two oscillators over time. My two oscillators are GPS locked oscillators (Trimble Mini-T), which each output a 10MHz signal. Looking at both signals on an oscilloscope (Tek MSO2024) I can see that there is indeed some drift between the two over time. They will move in and (slightly) out of phase over periods of time, but the changes are slow.
>
> I have logged data (only 10,000 samples, if you can help me figure out how to take more, that'd be great!) using tmtool in the instrument control toolbox. I took essentially three data sets: some arbitrary phase difference between them, completely inphase, almost completely out of phase.
>
> I have taken these data sets, calculated the FFT of each one, and plotted the results. In each case I get FFTs which indicate that the frequency is precisely 10MHz, for both oscillators, no matter what I do.
>
> Any suggestions of how I could make this measurement? Or perhaps I'm doing something wrong? I will post my code below. I'd like to take longer datasets, but I cannot figure out how to change the "MaxNumberPoint" variable to 100,000. I've tried changing everything in the driver file and I can get my time samples (length(X)) up to 100,000, but never my actual data (X).
>
> At any rate, any help would be much appreciated! Thanks!
>
> Code:
> close all; clear all; clc;
> %% Load Captured data from oscilloscope
> a = pwd;
> d1 = load([a '\data\dataSet_1.mat']);
> d2 = load([a '\data\dataSet_2_HandOnOscillator.mat']);
> d3 = load([a '\data\dataSet_3_Overcorrection_InPhase.mat']);
>
> Y1 = {d1.Y1, d2.Y1, d3.Y1}; %Channel 1 (Oscillator A)
> Y2 = {d1.Y2, d2.Y2, d3.Y2}; %Channel 2 (Oscillator B)
>
> t = d1.t1; %Only need one time set, they are all identical
>
> %Define sampling frequency
> Fs = 1e9; %Tek MSO2024 samples at 1GS/s
>
>
> %% Generate FFT of sinusoids
> %generate zero-pad data, making length a power of two
>
> for n = 1:length(Y1)
> nfft = 2^(nextpow2(length(Y1{n})));
>
> %Take FFT
> fft_y1{n} = fft(Y1{n}, nfft);
> fft_y2{n} = fft(Y2{n}, nfft);
>
> %Take the positive portion of the FFT
> [fft_y1_pos{n},freqRange{n}] = positiveFFT(Y1{n},Fs);
> [fft_y2_pos{n},freqRange{n}] = positiveFFT(Y2{n},Fs);
>
> %Plot the FFT
> subplot(3,1,n)
> plot(freqRange{n}/(1e6), abs(fft_y1_pos{n}));
> hold on;
> plot(freqRange{n}/(1e6), abs(fft_y2_pos{n}), 'r');
> grid on;
> xlabel('Freq (MHz)');
> xlim([0 20]);
> hold off;
> end


Hello Chris,

I'm writing you because I am a new user of instrument control toolbox and I am trying to connect my oscilloscope tek MSO2024 to tmtool with no good results, I downloaded the driver from this link:

http://www.mathworks.com/matlabcentral/linkexchange/links/3053-using-tektronix-msodpomdo-3000-and-4000-oscilloscopes-from-matlab


I'm trying to create the .mdd for the driver 'Tkdpo2k3k4k' by using makemid() and I get this error:

>> makemid('Tkdpo2k3k4k')
Error using makemid (line 139)
File does not comply with the VXIplug&play sub file format.

>>makemid('Tkdpo2k3k4k', 'ivi-c')
Error using makemid (line 139)
File does not comply with the VXIplug&play sub file format

Could you please tell me where can i download the driver for this oscilloscope??

I'm sorry I can not help you with your problem.

Thanks in advance