Hi Sverry,
there's a couple of things wrong here:
First of all, no, it can never be exactly the same!
There's random phase offset between RX and TX, unless you
use a USRP and a daughterboard that have deterministic
phase after tuning, and use timed commands for tuning
(which you don't). Also, the amplitude is not going to be
the same. Also, some noise will happen. That's basically
why we can't have infinitely fast internet everywhere –
your channel is never 100% known, so you can never fully
compensate it. Digital Comms :)
Then, you're using different RX and TX sample rates, so
the RX signal would at the very least have to look
differently "stretched" over the same number of samples.
Then: The signal going to the USRP must have |z| ⩽ 1.
Your signal has I and Q magnitude 25, and that makes for a
|z| = sqrt( I² + Q²) = sqrt(2)·25 > 1, so your signal
gets wrapped already at the TX. Multiply it by a factor of
1/(sqrt(2)·25), or better, much more.
Also, make absolutely, 100%, sure that you're not killing
your RX with your TX power. I don't know which USRP you're
using (you forgot to tell us, please try to put as much
relevant info into your problem descriptions as possible)
, but USRP frontends' RX LNSa are typically sensitive and
get damaged (with a few exceptions) at powers above
-15dBm. Most our TX frontends can put out something around
+20 dBm. That's more than a thousand times as much as
permissible for RX! So always, always use attenuators on
direct loop back.
Then, while this might work on an B2xx and on an X3xx,
not all devices support both the 5 and 8 MHz sampling rate
– a sampling rate MUST be an integer factor of the
device's master clock rate, and with 5 and 8, that means
the MCR must be a multiple (including 1) of 40 MHz. That's
the case for B2xx, because it has adjustable MCR and can
use 40 MHz directly, and for X3xx, because its MCR is
fixed at 5·40 MHz, but for the other devices, it won't
work (but you'd be notified in the shell window).
Best regards,
Marcus
On 03/16/2017 10:29 AM, Sverre
Turter Sandvold via USRP-users wrote:
_______________________________________________
USRP-users mailing list
address@hidden
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com