Uhd_fft and rx_ascii_art_dft die after 30 seconds

I am running ubuntu 10.04 in vmware and connecting to an n210 via a
gigabit ethernet USB dongle. When I run either uhd_fft.py or
rx_ascii_art_dft all works fine for exactly 30 seconds and then the data
display stops updating. If I rerun the app it works again for 30 seconds
and stops.

In wireshark the data looks good coming from 192.168.10.2 to port 49156
until it hangs then pc sends packet on 35433 and usrp reponds on 49152
and it just goes back and forth forever.

there’s plenty of disk space and memory and no messages in dmesg.
doesn’t seem to matter what sample rate I choose. I assume it has
something to do with vmware but just guessing. Any ideas? Thanks

On 30/10/11 11:27 PM, Clark P. wrote:

I am running ubuntu 10.04 in vmware and connecting to an n210 via a gigabit
ethernet USB dongle. When I run either uhd_fft.py or rx_ascii_art_dft all works
fine for exactly 30 seconds and then the data display stops updating. If I rerun
the app it works again for 30 seconds and stops.

In wireshark the data looks good coming from 192.168.10.2 to port 49156 until it
hangs then pc sends packet on 35433 and usrp reponds on 49152 and it just goes
back and forth forever.

there’s plenty of disk space and memory and no messages in dmesg. doesn’t seem
to matter what sample rate I choose. I assume it has something to do with vmware
but just guessing. Any ideas? Thanks


I would suspect some combination of VMWare and your USB Ethernet
dongle. USB-2.0 operates at
480Msps maximum, which is just shy of half of the required Gigabit
rates supported by
real Gigabit ethernet cards.


Principal Investigator
Shirleys Bay Radio Astronomy Consortium



Principal Investigator
Shirleys Bay Radio Astronomy Consortium
http://www.sbrac.org

Some more information on this. I’ve figured out that after the 30
seconds the VM network stack (I guess) throws an ICMP destination port
unreachable packet which looking at the usrp firmware will kill the
stream. There’s no reason for the ICMP message since hundreds of packets
to port 4140 precede the message.

I guess its a long shot but maybe someone out there has seen this with
ubuntu 10.04 in vmware? Thanks