Bug in benchmark_tx/rx/loopback?


#1

I have noticed that in benchmark_tx/rx (as well as in
benchmark_loopback), there is always
one packet lost in the transmission (always the last).
It is either not transmitted at all (getting out of the USRP)
or is not received. This behavior is persistent regardless of packet
size/data-rate/modulation format.

Has anyone looked into this before?

Achilleas


#2

Achilleas A. wrote:

I have noticed that in benchmark_tx/rx (as well as in
benchmark_loopback), there is always
one packet lost in the transmission (always the last).
It is either not transmitted at all (getting out of the USRP)
or is not received. This behavior is persistent regardless of packet
size/data-rate/modulation format.

Has anyone looked into this before?

I think it may be due to the amount of data needed by some of the
processing blocks to decode the whole packet. I would be that it
detects the preamble, but does not completely decode the packet because
it needs a couple more samples to decode the last bit or two. Maybe
check if it at least detects the preamble.

  • George

#3

Hi,

I have noticed it as well. If not the last packet, then the whole list
of packets received will be wrong. As in, even when using the same set
up, and the same configuration, the experiment will work this time but
maybe not the other time. I’m quite puzzled about this too. Could it be
the antenna problem?

And also, as i am very unfamiliar with gnuradio, but am asked by my
supervisor, whether there is a way of acknowledging the end of
transmission by the receiver to the transmitter? Like for example, to
send a straight 0s back to the transmitter?
Regards,Joreen> Date: Mon, 8 Dec 2008 23:22:03 -0500> From:
removed_email_address@domain.invalid> To: removed_email_address@domain.invalid> Subject: Re: [Discuss-gnuradio]
bug in benchmark_tx/rx/loopback ?> CC: removed_email_address@domain.invalid> > > >
Achilleas A. wrote:> > > > I have noticed that in
benchmark_tx/rx (as well as in > > benchmark_loopback), there is always>

one packet lost in the transmission (always the last).> > It is either
not transmitted at all (getting out of the USRP)> > or is not received.
This behavior is persistent regardless of packet > >
size/data-rate/modulation format.> > > > Has anyone looked into this
before?> > > > I think it may be due to the amount of data needed by
some of the > processing blocks to decode the whole packet. I would be
that it > detects the preamble, but does not completely decode the
packet because > it needs a couple more samples to decode the last bit
or two. Maybe > check if it at least detects the preamble.> > - George>

_______________________________________________> Discuss-gnuradio
mailing list> removed_email_address@domain.invalid>
http://lists.gnu.org/mailman/listinfo/discuss-gnuradio


#4

I have noticed something similar, except in my case its always a set
of packets, not just one. The higher the bandwidth I use, the more
packets I see lost (always towards the end).

-Ritesh

On Mon, Dec 8, 2008 at 11:16 PM, Achilleas A.