Re: [Discuss-gnuradio] Duplex daughter

[Marcus Leech]

If your processing delay is 1 second, that has nothing to do with the daughtercard.  It has to do with the way you've implemented your code.

 

The delay through the USRP is a few sample-times at the input sample rate, and then the latency of to-host

  link buffering, which is negligible, and then the link latency to host, which is again negligible--none of that adds up

  to 1 second of processing delay.

 

Now, on the duplex issue:

 

Imagine for a moment that there's someone shining a bright light in your eyes, and there's another person, several hundred meters away, trying to signal you in morse-code, using a candle and their hand.   It's unlikely you'll be able to see the remote person signalling you with a candle, unless you do some clever things.  Like move the candle spectrum outside the spectrum of the flashlight...

 

Now, there are things you can do, like subtracting-out your transmitted signal from your RX signal by phase-shifting it by 180 degrees and mixing it with the RX signal.   This is essentially how circulators work.  The results aren't that great, and when you have a wide-band, noise-like signal, it's rather hard to shift it by exactly 180 degrees, because it has more than one frequency component.

 

Which is why for same-frequency operation, many systems use half-duplex architectures  like TDD. 

 

 

on Jul 26, 2013, Dong Wang <address@hidden> wrote:

Hi Marcus,

 

Thank you for your help.

I tried to implement  the TDD on my USRP with a half-duplex daughterboard XCVR2450 using same frequency.

But because of the processing time delay, the receiver can not receive the packet until maybe 1 second after the transmitter send it.

that means I can hardly do real-time communication between two USRPs.

So I want to try some duplex daughterboard. It seems that these duplex daughterboard can only  implement  dul-band.

 

Best regards!

 

Dong Wang

2013/7/26 Marcus Leech <address@hidden>

You can use two widely-separated antennae, or use a single antenna and a circulator--circulators are necessarily quite frequency-specific, and only offer about 20dB isolation from one port to the next.

 

But, if it really is same-frequency at exactly the same time (as opposed to TDD), how do you propose to keep your

  very-strong TX signal from completely swamping the incoming RX signal?

 

 

on Jul 26, 2013, Dong Wang <address@hidden> wrote:

Hi all,

 

I plan to implement duplex with USRP n210.

I have a question whether the SBX daughterboard support duplex on the same band?

Like transmit at 2.4GHz and at the same time receive at 2.4GHz.

Thanks a lot.

 

Best regards!

 

Dong Wang

---

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