Re: [Discuss-gnuradio] OFDM benchmark - premature termination

[Tom Rondeau]

Katy Qian wrote:
> Hi,
>
> For the majority of my tests (ex. below), with benchmark_ofdm_*, the
> receiver terminates prematurely with:
> ...
> ok: True         pktno: 7501     n_rcvd: 4537    n_right: 4263
> ok: True         pktno: 7502     n_rcvd: 4538    n_right: 4264
> ok: True         pktno: 7503     n_rcvd: 4539    n_right: 4265
> Exception in thread Thread-1:
> Traceback (most recent call last):
>   File "/usr/local/lib/python2.4/site-packages/gnuradio/gr/gr_threading_24.py",
> line 446, in __bootstrap
>     self.run()
>   File "/usr/local/lib/python2.4/site-packages/gnuradio/blks2impl/ofdm.py",
> line 294, in run
>     self.callback(ok, payload)
>   File "./benchmark_ofdm_rx.py", line 168, in rx_callback
>     (pktno,) = struct.unpack('!H', payload[0:2])
> error: unpack str size does not match format
> uOuOuOuOuOuO
> ...
>
> It seems a certain combination of fftlength, occupied tones, and cp
> length allows the program to run to completion (FFT/OT/CP = 64/28/54
> and 40/20/32 works but 256/160/128 ends early). Does anyone know the
> "magic" behind these sets of values, or can explain something I have
> overlooked?
>   

This is an error that occurs when the packet is corrupted in a way that 
ends up with an empty payload, usually due to a bad value for the packet 
length in the header.

> Also, is there a way to control / define exactly which subcarriers are
> being used to transmit the data?
>   

No. You have to hack the C++ code in gr_ofdm_mapper. Others have been 
looking to do the same thing, but I'm not sure how far they have gotten.

> Any help is appreciated, thank you in advance.
>
> Katy
>