How to isolate the signals between TX/RX and RX2 on the USRP N210?

[0711]

Hello everyone,

    I want to use one USRP N210 to transmit and receive signals simultaneously. But when I try to transmit signals from TX/RX and receive signals from RX2(the transmitted and received signals are on the same center frequency), I found that the there are signals leaked from TX/RX to RX2.  What can I do if I want to isolate the signals between TX/RX and RX2?

   Thanks very much!

[0711]

PS: my daughterboard is SBX. Thanks again

[cochenob]

I'd be interested in hearing what NI/Ettus has to say about this.  I've noticed similar behavior when using an external source and the USRP in Rx-only mode.  In this scenario, I observed 'spurs' in the magnitude response that were located at +/- n*f_diff...where f_diff is the frequency offset between the input signal and the USRP LO.  If the input and LO happen to be fairly close, then all of those spurs get jammed up close to the tone of interest and it can reduce the useful dynamic range of the device. 

It may be something that's not really too much of an issue if you're doing comms...but if you're trying to do some kind of spectral analysis (especially with narrowband signals), it seems like it could be an issue.  Was sort of disappointed to see this kind of 'leakage'.

---

Brandon

[ErikL]

Hello -

I believe two questions are being asked on this thread:

1. Isolation:  The internal switch provides about 30dB of isolation between TX/RX and RX2. As with most transcievers, if you transmit and recieve on the same frequency at the same time you will recieve your own transmission as well.

2. Spurs: The USRP represents and optimized price vs performance and some images and spurs will occur.  Often they can be predicted as you mention. One thing you can do is use the CORDIC to acquire a signal at IF and then digitally shift I and Q to the ceneter frequency and bandwidth of interest.   This is done by setting the 'lo' via a property node and 'center frequency' to different values.

Other NI RF instrumentation options are available if you're looking for  RF performance.  The NI 5791 was recenty released and may be of interest.

Best Regards,

Erik

[cochenob]

Thanks Eric.  I had a feeling that it was likely a combination of the two.  I think I see that the latest version of the driver allows independent control of the 'analog' LO and the 'CORDIC' LO that would make the frequency shifting you mention possible.

Not to hijack the original post here, but since you mention it...I was just reading about the 5791's earlier this morning in an NI mailer.  Seems as if you created a PXI-based SDR by designing your own front end to interface to an "NI-targetable" FPGA (something you can't do with the USRP...yet.)

Does NI still have plans to provide the ability to target USRP FPGA's with NI's RT tools...or is the 5791 essentially a "work-around" alternative for that?  Being able to poke the FPGA on the SDR's is highly desirable, but being latched onto a PXI chassis is essentially a non-starter for me in terms of system integration and small form-factor packaging for sensors.

[ErikL]

Yes.  An NI USRP that is targetable with LabVIEW FPGA is a top request and is in development.  Send me an email if you need more information (niusrp@ni.com).  There is not plan to spport the current line of USRP devices with LabVIEW FPGA.

Erik

[0711]

If I want to receive and transmit at the same frequency, but do not want to receive my transmission, can you give me some suggestions?

Actually  I want to make an RFID tag based on  USRP  N210, SBX and Labview, working on the 915MHz.  The processing flow is showing in the following figure.

Thanks again!

[ErikL]

Please take a look at the following reference design.  You can always ignore incoming information when you are transmitting.

Attachments:

Bursty demo.zip 1000 KB

(from this post: http://forums.ni.com/t5/USRP-Software-Radio/How-to-transmit-and-receive-using-a-single-attena-with-U... )

[joqoko]

I also have the same problem. Since the discussion is in some months ago, have you found a way to isolate the signal from USRP itself? Thank you very much!

[HustLiliAn]

Maybe we just can do this by using TDMA, FDMA, or CDMA.