Full duplex and half duplex doesnot work

GNU Radio
1

Hi all,
Before i start telling about my problem, i want to mention that i have
already read the post named “help: cannot send after transport endpoint
shutdown” which is very close to my problem but i could not solve it.

I’m trying to transmit from the usrp port A, and then make it receive
at
the same port using “Auto T/R” which i expressed by typing “-v” (isn’t
it
correct) after the benchmark.
but some time it works and most of the times its not (not stable), when
its
not working it shows me the following error:
.onur@vecon3:~/Documents/abbasi/programmitx.py -f 1200000000 -m dbpsk -v
–tx-amplitude=0.08 -T A

gr_fir_ccf: using SSE

Modulator:
bits per symbol: 1
Gray code: True
RRC roll-off factor: 0.35
Tx amplitude 0.08
modulation: dbpsk_mod
bitrate: 100kb/s
samples/symbol: 2
USRP Sink: A: Flex 1200 Tx MIMO B
Requested TX Bitrate: 100k Actual Bitrate: 125k
Warning: failed to enable realtime scheduling
…onur@vecon3:~/Documents/abbasi/programmirx.py
-f 1200000000 -m dbpsk -v

gr_fir_ccf: using SSE

Demodulator:
bits per symbol: 1
Gray code: True
RRC roll-off factor: 0.35
Costas Loop alpha: 1.50e-01
Costas Loop beta: 5.62e-03
M&M mu: 0.50
M&M mu gain: 1.00e-01
M&M omega: 2.00
M&M omega gain: 2.50e-03
M&M omega limit: 0.01

Receive Path:
modulation: dbpsk_demod
bitrate: 100kb/s
samples/symbol: 2
usrp_open_interface:usb_claim_interface: failed interface 2
could not claim interface 2: Device or resource busy
usrp_basic_rx: can’t open rx interface
Traceback (most recent call last):
File “mybenchmark_rx.py”, line 112, in
main()
File “mybenchmark_rx.py”, line 101, in main
tb = my_top_block(demods[options.modulation], rx_callback, options)
File “mybenchmark_rx.py”, line 45, in init
self.rxpath = usrp_receive_path.usrp_receive_path(demodulator,
rx_callback, options)
File
“/home/onur/Documents/abbasi/programming/examples/digital/usrp_receive_path.py”,
line 68, in init
self._setup_usrp_source(options)
File
“/home/onur/Documents/abbasi/programming/examples/digital/usrp_receive_path.py”,
line 73, in _setup_usrp_source
self.u = usrp_options.create_usrp_source(options)
File
“/home/onur/Documents/abbasi/programming/examples/digital/usrp_options.py”,
line 88, in create_usrp_source
gain=options.rx_gain,
File
“/home/onur/Documents/abbasi/programming/examples/digital/generic_usrp.py”,
line 144, in init
_generic_usrp_base.init(self, **kwargs)
File
“/home/onur/Documents/abbasi/programming/examples/digital/generic_usrp.py”,
line 63, in init
except: raise Exception, ‘Failed to automatically setup a usrp
device.’
Exception: Failed to automatically setup a usrp device.
onur@vecon3:~/Documents/abbasi/programming/examples/digital$

ON THE OTHER HAND:
when i try the full duplex (use a slot from block A as transmitter and
slot
from block B as receiver) it also doesnot work. it will not be able to
setup
the interface.

i appreciate your help,
regards,

–
View this message in context:
http://old.nabble.com/Full-duplex-and-half-duplex-doesnot-work-tp27226209p27226209.html
Sent from the GnuRadio mailing list archive at Nabble.com.

2

On Tue, Jan 19, 2010 at 06:00:43AM -0800, abbasi9999 wrote:

not working it shows me the following error:
From the backtrace below, it appears that you already have the source
open, or you’re running something else that has the source open, or
you opened it earlier and didn’t close it, or still have a live python
to the earlier instance.

Half-duplex vs full-duplex is orthogonal to this issue. Likewise,
Auto T/R has nothing to do with the error below. With either full or
half duplex you would most likely open the source and sink once each
at the beginning of the program, then go about your business.

Eric

3

Eric B. wrote:

On Tue, Jan 19, 2010 at 06:00:43AM -0800, abbasi9999 wrote:
From the backtrace below, it appears that you already have the source
open, or you’re running something else that has the source open, or
you opened it earlier and didn’t close it, or still have a live python
to the earlier instance.

about that I’m running something NO.
However, YES the port is open because i dont know how to close it. I
thought
the benchmark_{rx,tx}.py close it self after finishing transmitting or
receiving if i added the parameter " -v ".

From other post i saw that 2+Ctrl will reset the usrp, I’ve tried it but
its
no use.
Another post said that to free the usb resources i should write the
command
“modprobe -vr ehci_hcd”, but it also did not work.

So how to free the usb resources or how to close the already opened usb
(opened by benchmark_rx,tx.py program).

regards,

Eric B. wrote:

Half-duplex vs full-duplex is orthogonal to this issue. Likewise,
Auto T/R has nothing to do with the error below. With either full or
half duplex you would most likely open the source and sink once each
at the beginning of the program, then go about your business.

Eric

As i said before, should not the parameter “-v” do the trick, or you
meant
that it do not close the usb resources that it already opened. therefore
i
have to do it manually?

best regards,

View this message in context:
http://old.nabble.com/Full-duplex-and-half-duplex-doesnot-work-tp27226209p27236110.html
Sent from the GnuRadio mailing list archive at Nabble.com.

4

The problem has been resolved.
thanks a lot Eric

abbasi9999 wrote:

So how to free the usb resources or how to close the already opened usb

at the beginning of the program, then go about your business.

Eric

As i said before, should not the parameter “-v” do the trick, or you meant
that it do not close the usb resources that it already opened. therefore i
have to do it manually?

best regards,

–
View this message in context:
http://old.nabble.com/Full-duplex-and-half-duplex-doesnot-work-tp27226209p27297831.html
Sent from the GnuRadio mailing list archive at Nabble.com.

5

Hi again,

Now I’m just trying to make sure that Auto T/R is working probably.
Maybe
when i was using two daughter boards i got the first problem. Now I
just
want to transmit from one daughter board then receive from the same
one.
When i do this after changing the transmitter into receiver it cannot
receive any thing?? it doesn’t even shows me any errors.

I’ve checked the “self.u.set_auto_tr(True)” command which written in
usrp_transmit_path and delete it.
after that i tried the same procedures and got the same result (NO
RECEIVING)

So, is there any idea to make sure that “self.u.set_auto_tr(True)” is
working probably? (do i have to use supporting code to make it works).

In some post (
http://old.nabble.com/help%3A-cannot-send-after-transport-endpoint-shutdown-to21700320.html#a21721364
http://old.nabble.com/help%3A-cannot-send-after-transport-endpoint-shutdown-to21700320.html#a21721364
) it was mentioned that
"…And benchmark_{rx,tx} are already set to do that.
See line 73 in receive_path.py and line 99 in transmit_path.py (line
numbers are from the trunk.) "
but I’ve check these lines and it has nothing to do with Auto T/R (i
found
Auto T/R only in “usrp_transmit_path”, it is not even exist in
“usrp_receive_path” )

best regards,

Eric B. wrote:

correct) after the benchmark.
Auto T/R has nothing to do with the error below. With either full or
Modulator:
…onur@vecon3:~/Documents/abbasi/programmirx.py
M&M mu gain: 1.00e-01
usrp_basic_rx: can’t open rx interface
line 68, in init
“/home/onur/Documents/abbasi/programming/examples/digital/generic_usrp.py”,
ON THE OTHER HAND:

Discuss-gnuradio mailing list
[email protected]
Discuss-gnuradio Info Page

–
View this message in context:
http://old.nabble.com/Full-duplex-and-half-duplex-doesnot-work-tp27226209p27240300.html
Sent from the GnuRadio mailing list archive at Nabble.com.

6

How did you resolve your problem? I have the same problem and reading
your
post I cannot understand the solution of your one.

Thanks in advance

blwfsoj wrote:

its not working it shows me the following error:
bitrate: 100kb/s
Gray code: True
modulation: dbpsk_demod
File “mybenchmark_rx.py”, line 45, in init
File
except: raise Exception, 'Failed to automatically setup a usrp
regards,

–
View this message in context:
http://old.nabble.com/Full-duplex-and-half-duplex-doesnot-work-tp27226209p30338716.html
Sent from the GnuRadio mailing list archive at Nabble.com.