RFX2400+USRP buffers

Hello everyone,

I am trying to implement a simple FSK system to measure bit error rate
(BER), and packet loss rate (PLR) using:

1-Linux
2-USRP motherboard
3-RFX2400 daughter card

I want to implement an echo back system where one Transceiver (say
Master)
will transmit a small frame (let say 48 byte, and important thing is
that I
neccessarily need small payload) to other transceiver (say Slave). The
Salve
returns the frame immediatelly without any processing to Master and
Master
measures BER and PLR.

The problem which I am facing is that I am unable to transmitt such
small
packets. I have noticed that if I have a large file with let say few
Kbytes
contents then it is transmitted (partially, the end of file is always
missing…I guess it is stuck somewhere in buffer???) but if I have
such
small packets as I metioned earlier then most of the time the packets
are
lost. If the packet is received then it is totally corrupt (say almost
100%
BER) unlike large files which are received in very good quality (But
important and notable point is that even the large files are 100%
courrupt
at the end). This makes me wonder why last few hundered bytes are always
courrupt??? (for small file the whole file lies in that category of last
few
hundered bytes)

I have tried a lot but unable to solve the problem! Now can any one help
me
to understand what can be the problem? Is it somewhere in the buffers?
Is it
some sort of synchronization problem??? and which buffers (of
USRP/RFX2400/GNURadio) are involved in such flowgraph. The flow graph
with
actual names of blocks used by me is given below (I have attached
complete
code if some one want to have a look):

Transmit path flowgraph:

gr.file_source -> gr.simple_framer -> gr.bytes_to_syms ->
gr.interp_fir_filter_fff -> gr.frequency_modulator_fc ->
gr.multiply_const_cc -> usrp.source_c

Receive path flowgraph:

usrp.source_c -> gr.fir_filter_ccf -> gr.quadrature_demod_cf ->
simple_correlator -> gr.file_sink

Best Regards

Kaleem Ahmad

http://www.nabble.com/file/p19415046/fsk_rx.py fsk_rx.py
http://www.nabble.com/file/p19415046/fsk_tx.py fsk_tx.py

View this message in context:
http://www.nabble.com/RFX2400%2BUSRP-buffers-tp19415046p19415046.html
Sent from the GnuRadio mailing list archive at Nabble.com.

Any ideas or comments will be helpful for me and I am anxiously
waiting!!!.

Kaleem Ahmad

kaleem ahmad wrote:

I want to implement an echo back system where one Transceiver (say Master)
packets are lost. If the packet is received then it is totally corrupt
actual names of blocks used by me is given below (I have attached complete
usrp.source_c -> gr.fir_filter_ccf -> gr.quadrature_demod_cf ->
simple_correlator -> gr.file_sink

Best Regards

Kaleem Ahmad

http://www.nabble.com/file/p19415046/fsk_rx.py fsk_rx.py
http://www.nabble.com/file/p19415046/fsk_tx.py fsk_tx.py


View this message in context:
http://www.nabble.com/RFX2400%2BUSRP-buffers-tp19415046p19430641.html
Sent from the GnuRadio mailing list archive at Nabble.com.

Dear All,

I have tried different packet sizes and I found that 2048 bytes data at
the
end is always missing:

With 128 byte packet size 17 last packets were not received
With 256 byte packet size 8 last packets were not received
With 512 byte packet size 4 last packets were not received
With 1024 byte packet size 2 last packets were not received

Can some one give me any clue??? Actually one possible solution can
be
to send 2048 bytes dummy data at the end of every packet, but the
problem is
I want to send small packets of not more than 100 bytes…so it is large
overhead for me…So if someone can give me some clue about where this
data
is being lost then it will be a great help???

Best Regards

kaleem ahmad wrote:

I am trying to implement a simple FSK system to measure bit error rate
to Master and Master measures BER and PLR.
bytes are always courrupt??? (for small file the whole file lies in that

Kaleem Ahmad

http://www.nabble.com/file/p19415046/fsk_rx.py fsk_rx.py
http://www.nabble.com/file/p19415046/fsk_tx.py fsk_tx.py


View this message in context:
http://www.nabble.com/RFX2400%2BUSRP-buffers-tp19415046p19433136.html
Sent from the GnuRadio mailing list archive at Nabble.com.

Even I am facing the problem of packet loss.
I have tried with all the packet sizes.
Can u suggest what is the packet size to be used?


View this message in context:
http://old.nabble.com/RFX2400%2BUSRP-buffers-tp19415046p27188383.html
Sent from the GnuRadio mailing list archive at Nabble.com.

On Thu, Sep 11, 2008 at 01:59:39AM -0700, kaleem ahmad wrote:

(BER), and packet loss rate (PLR) using:

category of last few hundered bytes)
Kaleem, until the inband code is completely sorted out, you’ll need to
ensure that the number of samples generated for your burst always ends
up as a multiple of 512 bytes on the USB. We use 16-bit I & Q on the
USB, so if you’re using complex baseband, you need to make sure that
you’re sending a multiple of 128 samples. This includes any frame
synchronization header, the payload, crc, etc.

Eric

On Sat, Jan 16, 2010 at 03:06, amarnath_iitm
[email protected] wrote:

Even I am facing the problem of packet loss.
I have tried with all the packet sizes.
Can u suggest what is the packet size to be used?

http://gnuradio.org/redmine/wiki/1/ReportingErrors

Johnathan

This forum is not affiliated to the Ruby language, Ruby on Rails framework, nor any Ruby applications discussed here.

| Privacy Policy | Terms of Service | Remote Ruby Jobs