[USRP-users] Fwd: Re: USRP's B210 sluggish start of transmission
perper at o2.pl
Fri Oct 6 04:40:10 EDT 2017
W dniu 27.09.2017 o 07:06, Piotr Krysik via USRP-users pisze:
> Hi all,
> I narrowed down when the issue occurs.
> When using separate sides of B210 for transmitting (i.e. TX/RX port on
> RF A side) and receiving (i.e. RX2 port on RF B side) everything is fine and B210 starts transmitting like i.e. X310.
> The problem appears when transmitting and receiving on the same side of
> B210 and it is dependent on what you connect to the TX port. The 300us
> transition appears if I don't connect anything and measure the TX-RX
> leakage or when VERT900 antenna from Ettus is connected.
> I tried also to connect a power splitter to the TX port and an antenna
> to one of the splitter's ports - with the same result. But as soon as I
> connected the 30dB attenuator to another port of the splitter, the
> problem immediately goes away.
>> The attenuator also has a path (resistive 50 ohms) to ground....an
>> antenna with a small pad on it (say 1 db) will probably work fine as well.
>> Kent Torell
> So you might be right here Kent that it was the path to the ground in the attenuator that
> changed how the output behaves. It is funny thought that the problem
> only appears when RX2 port on the same side of B210 is also used :).
As for my current application transmitting bursts and receiving
continuous signal with use of all B210 RF ports is not required, I can
avoid the problem. So I don't have currently motivation to pursue the
The problem isn't solved though. There are of course applications in
which it is required to use all RX and TX ports and it is convenient to
transmit burst. If someone needs it I can provide the codes that I used
for tests (they are not secret, I just don't want to lose time on
preparing them in case there is no such person). What I have is a GNU
Radio block that adds tx_time and length tags to a signal, with user
defined burst length and gap length. From the tagged stream UHD sink is
able to produce bursts instead of continuous signal. I have also a GRC
application that presents the problem.
More information about the USRP-users