I have a confusing issue. I am running my flow graph feeding data from a
source vector (repeat set to False). The data after processing is being
passed to UHD source. The file executes and data gets sent over the
channel. But even after the python file has been executed and returnds,
the
data is being sent by the USRP (until i power it off). I tried using
flowgraph start(), stop(), wait, sleep commands, to time and control it,
but the same issue persists.
Uzair B.
Could you describe what you mean by “data gets sent over the channel”,
and exactly what your flow-graph looks like. You say that you’re
passing data to a UHD source, but clearly, that’s not possible.
Also, the vector source/sinks were never intended for production use,
but
only for QA testing.
Unless there’s ethernet frames “keeping the pump full”, you shouldn’t
see anything out of the transmitter, except for LO leakage, after you
stop sending data.
Which versions of UHD are you using? What is your hardware? How is it
set-up? Which version of Gnu Radio are you using? What is your
platform? (Ubuntu, Fedora, Windows, etc).
–
Marcus L.
Principal Investigator
Shirleys Bay Radio Astronomy Consortium
I have two USRP1 (each connected to a laptop), separated 30cm apart.
Each
have a single Antenna connected to RFX2400 daughter card. I am running
Ubuntu LTS version. I downloaded GnuRadio+UHD from your (Marcus L.)
script 2 months ago.
Below are the details of my GnuRadio version:
MAJOR_VERSION=3
API_COMPAT=4
MINOR_VERSION=2
MAINT_VERSION=git
Now I am not sure how to find the information of my UHD drivers. (I
installed using your script, with no other changes, so it will be
exactly
the same, as in the script). I am sending some samples (all zeros or
ones,
using a vector source) to a UHD sink. The data does get sent, and
received
not once, but more than once, in the UHD source (connected to the other
laptop). And even after my code stops, and I try to capture data again,
using the UHD source, I get similar data, until after a 3-4 minutes
later
(when its just noise again) or I have switched off the power supply of
my
USRP1 that is a UHD sink.
My flow graph is pretty simple. UHD sink is in the tx.py file ( http://old.nabble.com/file/p33045005/tx.py tx.py ), and UHD source in
the
rx.py ( http://old.nabble.com/file/p33045005/rx.py rx.py ) file. I am
attaching those. To make sure I get all the samples, I run the rx.py
file
for many symbols (as compared to the number of symbols being sent).
Yes, right now, I am just using my vector sources and sinks for initial
testing to get an idea, I plan to shift to a file source.
installed using your script, with no other changes, so it will be exactly
rx.py ( http://old.nabble.com/file/p33045005/rx.py rx.py ) file. I am
attaching those. To make sure I get all the samples, I run the rx.py file
for many symbols (as compared to the number of symbols being sent).
Yes, right now, I am just using my vector sources and sinks for initial
testing to get an idea, I plan to shift to a file source.
UB
OK, so heres an idea. Rather than having people here debugging your
code, why don’t you try a very simple experiment.
On the TX side, construct a simple flow-graph that emits a single tone
at your chosen frequency–you should be able to whip that
up in GRC in about 30 seconds.
On the RX side, use uhd_fft.py to observe that tone.
Then, stop the TX side. Do you still see the tone?
–
Marcus L.
Principal Investigator
Shirleys Bay Radio Astronomy Consortium
No I don’t see the tone/spike. Infact, even with my own code, if I just
tx
and connect the source USRP to uhd_fft.py file, I dont find any peak at
the
frequency range.
Surely, this clears that the sink USRP could not be transmitting, but
then
why do I get similar repeated samples in my source USRP 40-60 seconds
even
later after my tx flowgraph has been executed. I am not sure, but is
because
my rx flowgraph uses a head block, so the required number of samples get
filled, and the remaining, are filled in the buffer of the USRP1, that
when
I run my rx flowgraph again, I get those samples… (doesn’t make sense,
i
think to have such a functionality, so maybe this is not the case).
Please help me find out the issue, as I keep getting the same samples
printed on my matplotlib graph even after tx has been executed
UB
Marcus D. Leech wrote:
Then, stop the TX side. Do you still see the tone?
Marcus L.
Principal Investigator
Shirleys Bay Radio Astronomy Consortium http://www.sbrac.org
(please discard the just previous post, i had made)
I tried your suggestion Marcus. I don’t find any peak with the GnuRadio,
when I construct a flow graph and stop the flow-graph. But when I do run
my
tx script and even after python returns call to the system, the
uhd_fft.py
keeps getting the peak at centre frequency, that means the tx USRP keeps
on
sending the data, for a loong time… Please HELP !
Uzair B.
Marcus D. Leech wrote:
Now I am not sure how to find the information of my UHD drivers. (I
UB
Then, stop the TX side. Do you still see the tone?
No I don’t see the tone/spike. Infact, even with my own code, if I just
tx
and connect the source USRP to uhd_fft.py file, I dont find any peak at
the
frequency range.
Surely, this clears that the sink USRP could not be transmitting, but
then
why do I get similar repeated samples in my source USRP 40-60 seconds
even
later after my tx flowgraph has been executed. I am not sure, but is
because
my rx flowgraph uses a head block, so the required number of samples get
filled, and the remaining, are filled in the buffer of the USRP1, that
when
I run my rx flowgraph again, I get those samples… (doesn’t make sense,
i
think to have such a functionality, so maybe this is not the case).
Please help me find out the issue, as I keep getting the same samples
printed on my matplotlib graph even after tx has been executed
UB
Marcus D. Leech wrote:
Then, stop the TX side. Do you still see the tone?
Marcus L.
Principal Investigator
Shirleys Bay Radio Astronomy Consortium http://www.sbrac.org