Hi Marcus,Thanks for your reply :)
-----------------------Original Message-----------------------From:
MarcusMüller<marcus.muel...@ettus.com>To: Hojoon Yang
<omnibu...@kaist.ac.kr>,usrp-users@lists.ettus.comSent date: 2018-06-04
20:37:04 GMT +0900 (Asia/Seoul)Subject: Re: [USRP-users] Complex signal
representation?
Hi Hojoon,
well, the point is that USRPs deal with *complex baseband*, so, of
course, all the samples are complex :)
Q1
So, yeah, this is all due to complex signal representation. To
understand this, there's no way around understanding complex equivalent
baseband, but don't fret, any communications basics textbook I've ever
seen does that!
What happens is that your cosine (cos x = 1/2(e^{jx}+e^{-jx}) )
undergoes a random phase change; that's perfectly normal and not a
problem at all, once you've gotten used to understanding signals in
complex baseband.
Q2
basically the same as Q1: These two *shouldn't* have the same
amplitude. All is good once you consider the shape of the complex
baseband representation of a phase-shifted cosine. It will probably
help if you, as I did above, consider the real cosine as the sum of two
complex sinusoids and consider what happens to each one of them when
mixed up and down with a quadrature mixer with a fixed phase offset.
Again, this is absolutely textbook stuff, and since you're writing from
a university address, I'd recommend you use the literature that is
recommended in the communications basics lecture. If there's no such
lecture at your university, GNU Radio does have a recommended reading
page:
https://wiki.gnuradio.org/index.php/SuggestedReading
Best regards,
Marcus
On Mon, 2018-06-04 at 20:07 +0900, Hojoon Yang via USRP-users wrote:
> Hi all,
> I'm using two B210 USRPs which are connected each other through
> coaxial cable with a 30dB attenuator.
> I generated a simple sinusoidal wave using MATLAB(i.e. cosine wave,
> fc=100) and saved it as complex float binary form.
> The reason I stored the samples as *complex float" is that there is
> no real float type in the UHD.
> I have seen that uhd always take the samples as complex!
> Anyway, I stored the samples in the following way: samples(4bytes)
> 0(4byte) samples(4bytes) 0(4bytes) samples(4bytes) 0(4byte)...
>
> Now, I ran tx_samples_from_file example on the one USRP and
> rx_samples_to_file on the other USRP.
> I expected that there will be only In-phase samples in the received
> file and all zero data in the quadrature samples. Because I send it
> such way.
> However, this figure is what I get: both In-phase and quadrature
> exist.
> I know it is not a bug at all, but don't know why this happens. Is
> this related with complex signal representation?
>
> Q2. why they are different? red one is quite pretty, but blue one is
> not. blue one is real and red one is imag.
>
> Both USRPs are connected to an external 10 MHz.
>
> Thanks!
>
> _______________________________________________
> USRP-users mailing list
> USRP-users@lists.ettus.com
>
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
_______________________________________________
USRP-users mailing list
USRP-users@lists.ettus.com
http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
