You are not logged in.
Hi Rice,
I have made some modifications on the PktBuffer_CRC(tx part) and Data_buffer(rx part). After my modification, the receiver cannot detect the incoming packet when simulated in Simulink. I tracked the signal and found that the long correlator cannot detect the incoming packet. Moreover, this phenomenon only happens occasionally and when I run the simulation for the second time, it can detect the incoming packet again. Since I only modified the payload processing part, it should not disturb the packet detection part which is only related to the short training sequence and long training sequence(preambles). Could you kindly point out the possible reason?
Thanks,
Tan
Offline
If you run the simulation with an unmodified PHY model, do you see the same behavior? All of the parameters set in C code in hardware are set by the .m init scripts in simulation; you can change these (like the correlation threshold) if needed.
Offline
I didn't see this issue when I run the simulation for the original design. The interesting thing is when I change
BER_RAM_init_values for the original design, the incoming packet cannot be detected for the long correlator sometimes. I know BER_RAM_init_values is irrelevant at all when it is not at FlexBERMode. So what is going wrong? For my modified design, the incoming packet can be detected for the first simulation. When I ran it for the second time, the first incoming packet cannot be detected; but the second incoming packet can be detected.
I'm pretty sure the problem is in the long correlator.
For my modification, I just added some logic between the scrambler and QAM mapper to transform the payload before QAM mapping in the tx. In rx, I did the reverse staff. My logic works fine and the BER value in the rx part shows no packet error as long as the incoming packet can be detected. So could you tell me the possible reason for this? I think the payload or scrambling has nothing to do with the long correlator in rx. Is this correct? Also, how to decide the optimum correlation threshold if I have to modify it?
Thanks,
Tan
Offline
tzhant32 wrote:
BER_RAM_init_values is irrelevant at all when it is not at FlexBERMode
The BER_RAM_init_values are also used in simulation for the BER calculation (different from the flexBER mode, which is only useful in hardware).
You can look at the long correlator output via the 'Rx Input' scope in the OFDM Rx block. During the preamble, you should see two tall spikes rising well above the noise floor. You want to set the threshold so that only these spikes exceed it.
Offline