[Discuss-gnuradio] Questions about Resampling for MPSK

[Tracey Bernath]

Ok, I've been banging around on this one for a while, and I'm gonna swallow my pride and ask for input.

I have a sample of a CDMA Pilot channel 50, 1.9325G, decimation 16, so 6.5Msps.  I want to run that through to extract the QPSK constellation, which I can feed to the short PN to extract the pilot sequence.  I figured this was a good place to start since the W0 Walsh code simplifies the math a bit.

Using usrp2_fft, there is a noticeable 'bump' at the 1.9325G, which is line with the fact it should be 6-8db 'louder' than the other channels.  I captured that to a file, and used the throttle to feed it back at 6.5Msps to a resampler.  I've tried the basic Costas loop, using the 2*pi*target/sampling fmin/fmax, with alpha in the range 0.000025 through 0.0001 with beta set to a*2/4.  I fed the output to a constellation plot, and it looks like teenage acne at its worst.  I waned to tackle the MPSK Receiver block, but I'm had a hard time finding enough info on all the parameters.

I'm about to go dust off old MATLAB code to run some known data through this, but I'd appreciate any and all insights....
Thanks,
TMB