How to achieve digital up conversion and digital down conversion in an OFDM transceiver system

Started by amitjonak 4 years ago5 replieslatest reply 4 years ago496 views

Hello again!

I want to achieve wideband frequency hopping by digitally up converting my OFDM signal and then digitally down converting the signal at the receiver to baseband. I tried to do simple up conversion by multiplying the I and Q components of the OFDM baseband signal by a carrier and add it up to form the real passband signal. However, this doesn't give me a up-converted spectrum. So my question is how to frequency translate the OFDM signal here. Why is the above method not working with an BPSK+OFDM system but works for a general BPSK transceiver system?

I would be happy if you could suggest any reference that illustrates these methods of doing digital up and down conversion on an OFDM signal and the required filtering techniques that is needed to achieve the purpose.  

Thank you!

[ - ]
Reply by dgshaw6March 31, 2020

Hello Amit,
Sorry to be too basic if this is such.
Just a quick question as to whether you are carefully choosing a sampling rate that can handle the real only signal within the half sampling rate.
Are you sure you have a sampling rate at least twice the highest frequency at the top edge of the OFDM signal?

[ - ]
Reply by amitjonakMarch 31, 2020

Yes sir, I used a sampling rate of 40MHz for the baseband OFDM signal and I used an interpolation factor of 20 to achieve the output sampling rate of 800 MHz with center frequency of 100 MHz. Actually the problem I am facing is this. If I use just a in-built block of Digital Up Converter (DUC) and use the above specs, it works for me. But if I just try to up convert by manually, it doesn't show frequency translation. So my question is how is up conversion different in an OFDM signal than a normal PSK modulated signal?

[ - ]
Reply by dgshaw6March 31, 2020

Should not be any different if you are starting with time domain samples.  If you up-convert the baseband to 800 MHz (interpolated by 20) for the real and imaginary independently, you should have a good base-band signal.
It should then just be a matter of choosing the real samples of a complex modulation by 100 MHz sampled at 800 MHz.

All of your carrier shifts are done on time domain base-band samples?

[ - ]
Reply by amitjonakMarch 31, 2020

Yes sir all carrier shifts are done on time domain baseband samples. I am still trying with different specs to see if I made any error in the method

[ - ]
Reply by napiermMarch 31, 2020


I would suggest breaking up the problem into simple steps.

1st: interpolate.  In the past I've use a sharp (long) half band filter to double the sample rate and give me the excess sample bandwidth needed for other steps.  Also it introduces very, very little in-band frequency ripple.

2nd: rate-match.  In real systems the symbol rate is determined by the link (external clock, RX recovery, steering, etc.).  There needs to be some way to go from the symbol domain to the clock domain of the D2A converter.  Having the signal band-limited to less than Fs/4 at this step is very helpful for a Farrow filter.

3rd: interpolate up to the D2A clock rate.  Easier now that you have excess BW.

4th: rotate.  The is a multiply of the complex base-band signal by the complex signal Fc, the center of your IF or RF signal.  Analyze the complex signal at this point and see that it is correct.

5th: mix to real.  Signal = real(complex signal) - imag(complex signal);

Have fun!