Happy New Year!

Hi everyone,

and a happy new year to you all! The last year was quite eventful for us
in GNU Radio land, and I’m excited to see where this year will take us.
Now that we have an API that’s not going to change any time soon, I’m
looking forward to a more stability, which hopefully means more
out-of-tree modules that work for a longer period of time.

We’ll be seeing some recurring events in 2014, such as:

  • The GNU Radio Conference
  • Google Summer of Code
  • Hackfests etc.

Possibly the first big event of the year will be FOSDEM in Brussels,
where GNU Radio is a big part of the Software-Defined Radio track.
Admission is free, so if you’re in central Europe, make sure not to miss

Our developer calls will also continue as usual; the first call is
January 16th, the usual time.

On a personal note, as you might have guessed from my email adress, I am
now an employee of Ettus R… This means I can continue my job as
Community Manager for GNU Radio, which I am very happy about.
My old email address at KIT will become obsolete soon, so please purge
that from your address books.

All the best,



I’m trying to send FM modulated packet data using the message burst
source and it’s mostly working.

I am able to see the tx_sob and tx_eob tags using the tag_debug block
hooked up to the same final block the USRP sink is (ie, the tags are not
getting lost somewhere along the chain).

When I examine the output file, the actual data for the bursts do indeed
start and end at the offsets listed in the tag_debug output, and there
are no samples in between.

If I receive the transmitted burst with a second USRP and plot the data,
you can see the received waveform at http://imgur.com/ZNuNWjt

The tx_sob and tx_eob are pointed out in the diagram, but there is a
long period of just a blank carrier at the end of the data being added
by the USRP N210. This works for some of the receivers, but is not
interoperable with the COTS hardware I’m working with.

I’ve attached the output of the block debug and my USRP instantiation
below. You can see that the tx_sob and tx_eob are listed, but that it
still gives an underflow indication (I thought the tx_eob was supposed
to prevent that).

How can I get rid of this extra energy the USRP is adding to the end?


The output of the tag debug is this:

Using Volk machine: avx_64_mmx_orc
– Opening a USRP2/N-Series device…
– Current recv frame size: 1472 bytes
– Current send frame size: 1472 bytes
Tag Debug: print_tags
Input Stream: 00
Offset: 0 Source: message_burst_source1 Key: tx_sob Value: #t
Offset: 3875 Source: message_burst_source1 Key: tx_eob Value:


My USRP instantiation is:

# TX Sink
self.uhd_usrp_sink_0 = uhd.usrp_sink(
self.uhd_usrp_sink_0.set_center_freq(cfreq, 0)
self.uhd_usrp_sink_0.set_gain(15, 0)
self.uhd_usrp_sink_0.set_antenna(“TX/RX”, 0)

    self.connect((self.unclip, 0), (self.uhd_usrp_sink_0, 0))