Surprised about the niUSRP Fetech data

[HustLiliAn]

Hi,every one

I found some strange phenomenons about the Tx IQ data and Rx IQ data, for example, the Tx send the same IQ data each time, however, the Rx data of the I path and Q path is different from the Tx. Sometimes the Rx I path data is the same or negative as the Tx Q path data, sometimes the Rx I path data is also the same or negative as Tx I path, the relationship is different for each run process. Tx data is the same all the time. Please look at the picture

 

The picture show the Rx I path data is negative as the Tx I path data, and the Rx Q path data is negative as the Tx Q path data. There some other situations, I just wonder what is the reason lead to this difference. Thanks

[ErikL]

You are observing phase offset. On an IQ constellation plot it would look like a rotated constellation. Phase difference is caused by several factors. Phase differences in the Tx and Rx LO as well as the channel (distance of the antennas from one another). What you see is normal. The magnitude and phase response (if IQ are rectangular coordinates, amplitude and phase are the equivalent polar coordinates) should be similar on both Tx and Rx.

[Anthony_F]

As Erik mentions this phenomenom happens when the transmitter and receiver are not synchronized.

Mathematically, it is quite easy to "fix" the problem: Just multiply the complex waveform by e ^ (j * -theta), where theta is the angle your waveform has been shifted.

An example of this has been posted under the PSK Phase Shift Demo.  I would recommend playing with that for live signals. I've attached an "offline" version here with a sample waveform constant on the block diagram.

I suggest also generating a single tone on either I or Q and changing the phase to see the effects more radically.