You are not logged in.
Hello all together,
I'm currently working with some WARPv3 boards and the OFDM Reference Design v18. To get a feeling of the performance of the boards i did some measurements with two boards, connected via coaxial cable and series attenuation of 50 and 75 dB. I used BPSK and half-rate coding and measured the receive power at the destination board by calculating it from the RSSI values and measured the packet statistics.
As it can be seen from the PER statistics in PER-plot; there are some packet losses (not detected by packet detector) in the range from -85 dBm to -70 dBm, but no misses in the range between -90 dBm and -85 dBm. Above -70 dBm again no packets are missed by the detector.
Do you have an idea why the packet detection shows this behavior?
Thanks&Regards
Offline
I'm not sure how to explain this behavior; I don't recall seeing this when using the OFDM Ref Design (a *long* time ago), but that's no guarantee it didn't happen. You could try modifying the autocorrelation packet detection parameters to see if you can affect the false negative rate in that SNR region. These params are set in warpphy.c.
Offline
Thanks for your response murphpo! I played with some values like the detection threshold and also the autocorrelation detector ones, without an effect. Do you have a specific value in mind?
Offline
You should definitely see an effect from changing the autocorrelation detector parameters (correlation threshold and energy threshold). For example you should be able to directly control the minimum SNR at which the detector fires by raising the energy threshold.
Offline