beginners question
beginners question
(OP)
I'm wanting to recieve CDMA signals from satellite transmitters. I'd like to do it all in digital rather than analog. However my concern is obtaining corralation to despread an individual signal with the doppler effect from the satellite movement. I'd rather not perform the doppler search on the analog IF but if there is a way to perform this in the DSP that would be ideal. Now I've read how radar systems use "quarter turn" data to find the doppler. I'm not certain that it will work on satellites though.
Basically as I've read, in radar systems the first FFT is used for corralation with the broadcast carrier. Then past several DFT samples are FFT again across the bins to find the doppler of the return signal. I'm not completely clear on the full implementation of this as the article had no description of how the second FFT data was used.
I'm not certain that process would work. However I thought instead could I adjust the sample rate fed into the FFT comapred to the sample size to adjust the frequency of the corralation search instead?
If anyone has experience with this concept or has links to good information I can read I would greatly appreciate it.
Basically as I've read, in radar systems the first FFT is used for corralation with the broadcast carrier. Then past several DFT samples are FFT again across the bins to find the doppler of the return signal. I'm not completely clear on the full implementation of this as the article had no description of how the second FFT data was used.
I'm not certain that process would work. However I thought instead could I adjust the sample rate fed into the FFT comapred to the sample size to adjust the frequency of the corralation search instead?
If anyone has experience with this concept or has links to good information I can read I would greatly appreciate it.





RE: beginners question
The post processing approach you describe and what I describe here allow you to finish the demodulation at a slower rate. If you are going to adjust the frequency of your original sample rate (probably not trivial in itself) you'll need to do it in a closed loop system and deal with tracking errors (that's not trivial either).
These are the basics from low data rate PSK signals; it will get much trickier for CDMA. If you want more info, Google 'software radio'.
Z
RE: beginners question
How similar is your application to GPS?
RE: beginners question
I was wanting to see if its possible to use a type of sample adjustment from the already digital 2 bit sign magnatude (1 or -1) I/Q data time domain data.
Perhaps an oversampled, several time point averaging or liniarly dropping every Nth sample (or adding if enough oversampling is available) to shift the frequency of samples in time domain. Hopefully aligning sign/magnatude of I data of the actually doppler shifted carrier in the same time frame as the PRN spreading code.
Thereby always submitting the same number of samples into the FFT for frequency domain complex multiply with freq domain PRN sequence of fixed time length.
Time frame of 1-2 milliseconds depending on PRN code type such as L2C or L5.
That was my idea at least. I have no idea if it would work or not.
RE: beginners question
The above might not apply to GPS equipment used in other industries, such as precision positioning in agriculture, where they get into carrier tracking.