Re: [Discuss-gnuradio] USRP GNU radio receiver

[Nick Foster]

The USRP has a DC offset removal filter in the FPGA. A carrier sent at 0Hz is exactly the same as DC offset, and will be removed. Zero-IF receiver architectures (like the USRP) intrinsically have DC offset imperfections. The filter can be disabled in the USRP Source "FE Options" tab. It is not good practice to design a system which depends on DC in a zero-IF receiver, because the DC offset error will change over frequency, temperature, time, etc.

On Thu, Jul 18, 2019 at 4:29 AM Simona Sibio <address@hidden> wrote:

Hi all,

Thank you very much for your emails.

But no tip works fine.

I wanted only to learn how is USRP working with GNU radio, I thought that the RX chain in the USRP was different.

But I am not the beginner in electronic and RF circuits.

Anyway, no problem, I don't ask more information about this.

Best Regards,

Simona

Il giorno gio 18 lug 2019 alle ore 12:11 Jonas Manthey <address@hidden> ha scritto:

Hi,

 

You don’t send a sine with  frequency 0, you send a carrier that is modulated by a sine with a frequency of 0. A sine with a frequency of 0 is 0: sin(2 pi f) = sin(2 pi 0) = sin(0) = 0.

 

So your carrier is modulated with a zero, which I *think* ends up in no carrier at all, not sure right now how zero-IF architectures behave in that case. I highly recommend reading some communications or RF book/website/guide or talk with your advisor.

 

UHD-USRP clips anything above 1 if left in the default setting (float I guess) and there will be a lot of garbage in the sent spectrum.

 

Cheers,

Jonas

 

From: Discuss-gnuradio [mailto:discuss-gnuradio-bounces+jonas.manthey=address@hidden] On Behalf Of Simona Sibio
Sent: Donnerstag, 18. Juli 2019 13:01
To: Kyeong Su Shin <address@hidden>
Cc: address@hidden
Subject: Re: [Discuss-gnuradio] USRP GNU radio receiver

 

Hi Kyeong,

 

I attached the screenshots about time-domain and frequency domain.

I sent a sine with frequency=0 and amplitude=0,25.

The output is always zero in the receiver.

The situation changes only if I send a sine with a frequency non-zero.

Thank you for your time.

Best Regards,

 

Simona

 

Il giorno mer 17 lug 2019 alle ore 17:41 Kyeong Su Shin <address@hidden> ha scritto:

To whom it may concern:

 

I think there has been some miscommunications or misunderstandings here: the 'Offset' argument for the 'Signal Source' block is for the 'amplitude offset' of the signal, while what Marcus probably meant in his previous e-mail is the 'frequency offset' (you can use USRP's low-freq digital IF capabilities for that). Also, I do not know how UHD sink behaves when you do not input normalized values, but I am pretty sure that you should not input such large values to UHD sink.

 

If readings in the frequency-domain are non-zero (or, in real-world devices, extremely low values), readings in the time-domains are also non-zero. So, "the output became unstable in frequency and in time is always zero" doesn't make sense. Your time-domain plot is probably zoomed out too much.

 

Finally, as Marcus mentioned, throttle blocks _must_ be removed.

 

Regards,

Kyeong Su Shin

 

---

보낸 사람: Simona Sibio <address@hidden> 대신 Discuss-gnuradio <discuss-gnuradio-bounces+ksshin=address@hidden>
보낸 날짜: 2019년 7월 17일 수요일 오후 11:35:00
받는 사람: Müller, Marcus (CEL)
참조: address@hidden
제목: Re: [Discuss-gnuradio] USRP GNU radio receiver

 

Thank you very much!

I sent the screenshot because in the oscilloscope there is not the offset.

Instead, I would like to detect that.

Thank you in advance.

 

Simona

 

Il giorno mer 17 lug 2019 alle ore 15:22 Müller, Marcus (CEL) <address@hidden> ha scritto:

1. Never use a throttle with a hardware block. GRC's log will actually
shout at you that you shouldn't be doing that!
2. I must admit this is about as good as I'd expect. I of course can't
infer temporal behaviour from screenshots.

Best regards,
Marcus

On Wed, 2019-07-17 at 14:18 +0100, Simona Sibio wrote:
> Hi Marcus,
>
> I tried to change the offset in the transmitter but the output in the receiver is the same, also if I send a sine signal with 10 Hz and 0,1 amplitude: the output became unstable in frequency and in time is always zero (see attached, 130K).
> The flow graph is attached.
> Do you have any suggestions?
> Thank you in advance.
>
> Simona
>
> Il giorno mar 16 lug 2019 alle ore 18:59 Müller, Marcus (CEL) <address@hidden> ha scritto:
> > That phase is not any more or less random than that of your local
> > oscillator, so there might be some misunderstanding here.
> > On Tue, 2019-07-16 at 18:52 +0100, Simona Sibio wrote:
> > > Thank you very much!
> > > But, with the offset tuning I cannot edit the phase.
> > > Maybe, I have to change method.
> > > Best Regards,
> > >
> > > Simona
> > >
> > > Il giorno mar 16 lug 2019 alle ore 18:25 Müller, Marcus (CEL) <address@hidden> ha scritto:
> > > > Well, there's leakage / DC blocking filters in most mixers, so yes.
> > > >
> > > > But the solution is simple (and hardware-supported transparently on
> > > > USRPs): Just tune at an offset, and mix down to 0 Hz (==constant!) in
> > > > digital. Your USRP can do that for you, google for "USRP offset
> > > > tuning".
> > > >
> > > > Best regards,
> > > > Marcus
> > > >
> > > > On Tue, 2019-07-16 at 18:15 +0100, Simona Sibio wrote:
> > > > > Thank you very much!!
> > > > > Then, I cannot measure a constant because, in the RX chain, maybe, some filter deletes the information, doesn't it?
> > > > > Thank you for your time and tour tips!
> > > > >
> > > > > Best Regards,
> > > > >
> > > > > Simona
> > > > >
> > > > > Il giorno mar 16 lug 2019 alle ore 18:01 Müller, Marcus (CEL) <address@hidden> ha scritto:
> > > > > > Look up what a quadrature modulator does. It's just a multiplication
> > > > > > with e^{j2\pi f_{carrier}t}, in the end.
> > > > > >
> > > > > > Your constant signal s(t) = c const. would hence be transmitted as
> > > > > >
> > > > > > c·e^{j2\pi f_{carrier}t}
> > > > > >
> > > > > > whereas a complex sinusoid of frequency f and amplitude a,
> > > > > > s(t) = a·e^{j2\pi f t) would be transmitted as
> > > > > >
> > > > > > a·e^{j2\pi f t)·e^{j2\pi f_{carrier}t} = a·e^{j2\pi (f+f_{carrier})t} 
> > > > > >
> > > > > > so, identical to the constant signal, just at a different frequency.
> > > > > >
> > > > > >
> > > > > > Best regards,
> > > > > > Marcus
> > > > > >
> > > > > > On Tue, 2019-07-16 at 17:53 +0100, Simona Sibio wrote:
> > > > > > > Hi Marcus,
> > > > > > >
> > > > > > > Thank you for tip!
> > > > > > > I know that calibration is not so easy but I hope that.
> > > > > > > Could you help me to understand what happened with a constant complex baseband signal?
> > > > > > > I am beginner in this field.
> > > > > > > Thank you for your time.
> > > > > > >
> > > > > > > Best Regards,
> > > > > > >
> > > > > > > Simona
> > > > > > >
> > > > > > > Il giorno mar 16 lug 2019 alle ore 17:39 Müller, Marcus (CEL) <address@hidden> ha scritto:
> > > > > > > > Watch out! What do you get at the *output* of your SDR device when you
> > > > > > > > transmit a constant complex baseband signal, mixed to a carrier
> > > > > > > > frequency? How does that differ from transmitting a complex baseband
> > > > > > > > sinusoid mixed to a carrier frequency?
> > > > > > > >
> > > > > > > > On Tue, 2019-07-16 at 18:32 +0200, sumit kumar wrote:
> > > > > > > > > Ok, so for a constant source, in the time domain you will see a flat signal only! Try connecting a spectrum analyzer and you shud see a peak at DC becz Fourier transform of constant signal is impulse.
> > > > > > > > > And yes, as Marcus said, increase the sampling rate!
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > On Tue, 16 Jul 2019 at 18:06, Simona Sibio <address@hidden> wrote:
> > > > > > > > > > Thank you for the assistance.
> > > > > > > > > >
> > > > > > > > > > I put the module "constant source" in my flow grap.
> > > > > > > > > > With this module, I can choose witch amplitude I want to send.
> > > > > > > > > > I attached the flow graph.
> > > > > > > > > > And, I would want to read these values in the receiver.
> > > > > > > > > >
> > > > > > > > > > Simona
> > > > > > > > > >
> > > > > > > > > > Il giorno mar 16 lug 2019 alle ore 16:58 sumit kumar <address@hidden> ha scritto:
> > > > > > > > > > > Hi Simona, what is the "constant signal" here ?
> > > > > > > > > > >
> > > > > > > > > > > On Tue, 16 Jul 2019 at 17:54, Simona Sibio <address@hidden> wrote:
> > > > > > > > > > > > Hi all,
> > > > > > > > > > > >
> > > > > > > > > > > > I want to use GNU radio to measure the amplitude and the phase of a signal.
> > > > > > > > > > > > I send a constant signal with the transmitter USRP but in the receiver there is a flat signal with amplitude 0 and offset 0.
> > > > > > > > > > > > I tried to send a sine signal and the receveir works fine.
> > > > > > > > > > > > How can I do to send only a constant and measure the amplitude?
> > > > > > > > > > > >
> > > > > > > > > > > > Thank you for your time.
> > > > > > > > > > > >
> > > > > > > > > > > > Simona
> > > > > > > > > > > >
> > > > > > > > > > > > _______________________________________________
> > > > > > > > > > > > Discuss-gnuradio mailing list
> > > > > > > > > > > > address@hidden
> > > > > > > > > > > > https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > > > > --
> > > > > > > > > > > Sumit Kumar
> > > > > > > > > > >
> > > > > > > > > > >
> > > > > > > > >
> > > > > > > > >
> > > > > > > > > _______________________________________________
> > > > > > > > > Discuss-gnuradio mailing list
> > > > > > > > > address@hidden
> > > > > > > > > https://lists.gnu.org/mailman/listinfo/discuss-gnuradio

_______________________________________________
Discuss-gnuradio mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/discuss-gnuradio