[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Discuss-gnuradio] Question about sampling rate and center freq

From: abhinav narain
Subject: Re: [Discuss-gnuradio] Question about sampling rate and center freq
Date: Sat, 30 Apr 2016 21:23:23 -0700

Hi Marcus,
Not at all!
Sorry, I was taking pictures of hardware and hence just took one for
the GRC too.  As I showed in previous picture, I am only using RX-B. I
am not sure how to connections will be for dual mode, but currently I
am just using one. I get complex values in the GRC and I can plot 
complex FFT of them
This is my setup where i am extracting signals from Powerline using RX-B pin of the daugterboard 
There are two *different* projects :
a) I want to understand the powerline channel 
b) I want to transmit and receive signals on the channel
I have a better picture[2] now : http://postimg.org/image/hptdrl2vl/
Interesting! Why is there a time axis? I must admit those graphs look like they took some effort to create; really, a clean drawing on paper would have been as good for me, but thanks.

I am showing the spectrogram when I set center freq at 125 kHz  0 kHz.
I am trying to show what I am getting from the channel 
And, as I mentioned in IRC when I asked for them, it is important to define wheter you're doing I&Q or real sampling.

I am dong I&Q sampling as shown in [1]. By real sampling, I suppose
you meant that I am taking  magnitude before feeding to the Spectrogram.
No, I didn't mean that.
I mean: Is the signal you *want* to observe a single voltage signal varying over time, or is it an I & Q signal pair?

The signal in (a) is the EMI on the channel by various devices on the channel.
I want to observe the I, Q samples on the channel. 
The GRC pic showed the complex datatype of USRP source and I assume that it is giving me IQ samples from channel on RX-B pin.
Please let me know if I am incorrect ?
If I had set real data type of USRP source block or complext to mag in front of complex datatype of USRP source block, I would have got real signals ... according to my current understanding.
It's no use discussing any of your question until we know whether we're looking at a complex equivalent baseband or at a real signal.

I am doing complex baseband as in the picture.
I'm not convinced you actually are in both cases! That's why I'm asking :)
Okay, I assumed I am doing since I am using complex datatype of USRP source and just changing the center frequency and sample rate.
Either both case i am doing complex baseband or I am doing it in neither, because changing center frequency shouldn't make a difference.
My question was related
to the sampling. I thought doing sampling (complex baseband as shown
in GRC)
OK, we need to talk about sampling:
Sampling takes place where your analog signal becomes a series of digital numbers, in the ADC of the N210. That is before you digitally tune to e.g. 125 kHz. The LF and Basic daughterboards don't have tuners!
in 250kHz at center frequency 125kHz is equavelent to doing
500kHz, with center frequency 0Hz  using LFRX.
That really really depends! It depends on whether you want to observe a single real signal or an I&Q signal pair. So: what do you want to observe?

In a) I want to observe the channel EMI
In b) I am transmitting real signal on one side and able to decode a signal on the other side. 
The TX setup is exactly the same  as above, just with a different daugtercard which transmits, receives at center freq 1 MHz.
So here's the problem I'm having with your original question:
You said
I want to sample first 250KHz on the channel.
But: What *are* the first 250 kHz?
If you're just sampling a real signal with a single ADC, then obviously, you get a real-valued stream of samples, and to cover the bandwidth of 0 Hz – 250 kHz you will need to sample at  500 kHz.

I don't get " a real signal" ? 
There is one single ADC per LFRX card's RX pin that will give complex values with a Complex USRP source right from the each RX pins  ? 
I am using RX-B, hence I should be receiving complex samples from channel with complex USRP source. 
Please correct me, because that is what i think.

If we're talking about complex baseband, I'd define (and that's purely something that I like, not something that is uniformly defined somewhere) the "first 250 kHz" to be the spectrum from -125 kHz – +125 kHz, because those 250 kHz correspond to the 250 kHz bandpass signal "closest" to the center frequency of your down-mixer. To get those 250 kHz, your ADC will need to run at 250 kHz, but it will need to produce one I and one Q sample simultaneously.

I would definitely want to understand what am I getting with the current GRC pic.
I think it is complex baseband. Please correct me.
You can configure the N210 carrying a Basic or LF board into either mode – real sampling of up to two channels, and complex sampling of an I&Q pair that you feed into the two SMA connectors on the daughterboard.

Does this mean, I need to use both SMA pins to get the complex samples and I am currently getting real samples ?
We really need to understand what you *want* to be doing!

As I explained above, I want quadrature samples from the channel, which I think I am doing in my GRC and the setup pic.

It might be best to know how to get both - real or complex setup.
I would like to do complex baseband always  and do processing on the received data file as part of analysis, and not restrict to real sampling.
So, I really don't know what literature you've got access to – if you have access to the scriptum for a communication theory introduction, that might be a good start.

I have read some chapters from Oppenheim but I am still not clear what is practically going on in my system.

I really appreciate your time for helping me,

reply via email to

[Prev in Thread] Current Thread [Next in Thread]