Why modulation and demodulation with PXI-5412 and PXI-5660 was wrong?

[phonglt03]

Hi all.

I've use PXI-5412 to modulate OQPSK signal and use PXI-5660 to demodulation it but the result was wrong. I attach some pictures to illustrate the problem.

Could you help me, please?

Thanks

-----------------------

LuongTanPhong

[Timothy_S]

LuongTanPhong,

I'll need to see if I can test these examples on my end which could take a bit of time to get all of the hardware together, but for now could you explain to me a bit more about the hardware setup that you are using?

What PXI chassis are we using, and are we running on an internal controller or a PC? Is the 5412 and the 5660 setup the only hardware in the system? How are we connecting the signal from the output of the 5412 to the input of the 5660? Is anything happening to the signal in between the two? Let me know what you can about the configuration and setup while I look in to this issue. Thanks.

[phonglt03]

Dear Timothy S,

Thank you for your concerned.

I use PXIe-1065 chassis, PXIe-8106 controller with  5412 and 5660 were setup on. I used 5412 to transmit OQPSK signal at 10.7 MHz with 24k symbol rate and used 5660 to demodulate the signal (I used example "MT niFGen PSK Signal Generation.vi" in FGEN and "MT ni5660 PSK Demod.vi" ni5660 directories of Modulation Toolkit). All applications were run on PXIe-8106 with a cable connect 5412 and 5660 directly.

 When applications are running, the waveform recovered on IQ Constellation (demodulation) has the same shape with the transmiter but the bits out are wrong (their phase difference, right?)

Could you help me to fix the error, please?

Thanks.

LuongTanPhong

[Timothy_S]

While I'm looking in to the particulars on those examples (I should be able to get the hardware together tomorrow), could you see if you're seeing the same odd behavior with any other settings (other symbol rates/parameters, etc) as well as try some of the other Modulation examples just to confirm if we're seeing an issue with the one set of examples or with the hardware itself. Let me know what you can do. Thanks.

[phonglt03]

I've test with other settings (symbol rates, frequencies, modulation types) with user-defined pattern and I found that:

- BPSK, QPSK, OQPSK: the shape (IQ constellation) on transmiter and receiver are similar but different on their phase.

- pi/4 QPSK, PSK8, PSK16: their shapes are difference.

LuongTanPhong

[Timothy_S]

I've checked my system (with a 5660 and 5412) with the two examples and the setup you're using and I am actually getting similar results from what you're seeing. I'll be looking in to the issue a bit more in depth with these examples to see if I can locate the cause of it. In the mean time, would it be possible for you to try other examples (so any of the other modulation toolkit examples/VIs) to see if we can get any other modulation schemes to show correctly on the hardware?

[Timothy_S]

LuongTanPhong,

The main question we'll want to look at is whether or not the actual bits being transmitted were being demodulated successfully on your end. Using PSK for with two different units (one for generation, and one for acquisition), seeing a phase offset or difference is not an uncommon occurrence. What we're more interested in is the integrity of the transmitted data, so make sure that we check that first. If we're still having issues, or we want to keep the units in phase (to see similar constellation plots), we would want to try using the same reference clock for both generation and acquisition systems. The easiest one to use for this case would be the 10MHz backplane clock in the PXI chassis. Without a common reference clock, the 5412 and 5660 are going to use different clocking sources and may be out of phase naturally.

[phonglt03]

Dear Timothy S,

Thank you for your feedback.

In fact, I used USRP2 (with LFRX module) on LabView to demodulate OQPSK signal and PXI-5660 to modulate signal for it. When I compera the results of USRP2 (bits out) with 5660, I found that there are differences between them. I also find that they're not the same phase, the phase of 5660 is stable during the phase of USRP2 is constantly change. 

My utimate goal is to compare the results between the modulation and demodulation rather than their phase.

Could you help me to fix issues, please?

Best Regards.

Luong Tan Phong