modulation-demodulation synchronization between 5672 signal generator and USRP-2930

[irad]

Hello,

I am trying to get a 5672 signal generator and a USRP 2930 to communicate using a PSK modulation.
There are two useful NI examples that almost do the job. On the transmitter side, which is the 5672 signal generator,
I am using "MT RFSG Generate PSK". On the receiver side (USRP 2930), I am using "niUSRP EX PSK Rx". The transmitter
and the receiver are connected with a short cable, so it's possible to make the signal to noise ratio very high.

At the moment, I am sending a user defined input sequence instead of a PRBS because that makes it easier to check
results without writing additional code. The sequence gets demodulated correctly, but the bits often get flipped
because of the phase ambiguity due to lack of synchronization as explained in the help file for the PSK demodulation
block. To fix this problem, I slightly modified the two VIs to incorporate the block for generating synchronization
parameters. This block is present in the "MT PSK transceiver (one shot)". The VIs with the modifications are attached.
In the transmitter file, the only change is the addition of the synchronization sequence to which the user defined
message sequence is concatenated. In the receiver file, "MT generate synchronization parameters" now appears. It takes
the synchronization bits defined on the transmitter side.

Unfortunately, the result is still the same as before. What else need to be done to make this work? In "MT PSK transceiver
one shot", there is an implicit frame synchronization. It is a simulation VI where the demodulation part consumes the array
produced by the modulation part. This is not the case in the setup that I have. Even if I set data rates to make the modulator
and demodulator work with the same array sizes in every iteration, the expected location of synchronization bits is still
not known.

Regards,

Ivan

[Wan-L]

Hello Irad,

are you refering the filter delay issue mentioned in the MT Demodulate PSK.vi help file?

[irad]

Hi Wan, 

I am talking about incorrect mapping of symbols in the receiver. For example, 00 could be interpreted as 11 in case of QPSK where there are two bits per symbol. 

The help file says that the absence of synchronization parameters could cause this problem. 

I don' t think filter delay is a problem for me. I send repetitive sequences and I don't mind losing a few bits in the beginning. 

[Wan-L]

Try taking just a "single shot" instead of acquiring continuously, it helps rule out the filter delay possibility. In this other DF it looks the person is having a similar issue to you. http://forums.ni.com/t5/RF-Measurement-Devices/PSK-Demod-VI-doesn-t-demodulate-the-enitre-message/m-...

You might need to try the recommendations there if you need to acquire continously.