Say I transmit on an N210/WBX around 300 MHz and receive on a second N210/WBX that is on its own reference clock. I transmit two tones simultaneously, say 299.9 MHz and 300.1 MHz, and on the receive USRP, downconvert by its 300 MHz carrier to baseband and then take the FFT to compare where the receive USRP expects the two tones (-100 kHz and +100 kHz) and where they actually occur. I am doing this to get an idea of the local oscillator offset between the two radios.
I notice that this offset is not constant. For example, the -100 kHz tone will be at -100 kHz + 40 Hz, while the +100 kHz tone will be at +100 kHz + 50 Hz. Furthermore, if I transmit back from that receive USRP to the transmit USRP with the transmitter's RX2 set to 301 MHz, I have to digitally adjust the tone I send up 95 Hz from 301 MHz and then it will be demodulated to exactly at baseband on the transmit USRP. The amount of adjustment is simply a linear extrapolation based on the two offsets I measured (40 Hz at 299.9 MHz and 50 Hz at 300.1 MHz) going in the other direction from the transmitter USRP to the receiver USRP.
My question is if anyone understands the TX and RX chains, DDCs and DUCs, etc., well enough to explain the physical reason why two radios should not simply have a constant frequency offset at all frequencies, but instead have a linearly varying frequency offset. I would have thought that the difference in local oscillators between two radios would cause a constant frequency difference to occur no matter the carrier frequency.
Thanks,
Mike
