Subcarrier Frequencies

[RichF88]

What I would like to do is to modulate two signals to subcarrier frequencies, and then modulate them again up to the final carrier frequency. Is this possible using the NI USRP drivers and the Modulation toolkit?

[Anthony_F]

Hello RichF88,

This should be easily accomplished with the help of the MT Upconvert Baseband VI. Just take the baseband waveform and use the "carrier frequency" input for your offset. The USRP will do the final upconversion to the carrier frequency as usual.

[RichF88]

Anthony,

Thank you for your quick response. I tried to implement your suggestion, but came up with a problem. I attached the VI that I used, and made the parameters default values. The following is what I got when viewing the signal on a spectrum analyzer:

This was with a Root Raised Cos Filter applied to the PSK modulation. The intent was to do a BPSK modulation, upconvert it to the 49.125 KHz subcarrier, and then allow the NI USRP to again upconvert it to the final 70 MHz carrier. However, I expected the signal to be symmetric around the 70 MHz carrier. Am I doing something wrong?

Also, as a side question, the Upconvert VI is implemented in software correct? Is there any way to force it into the FPGA in the USRP hardware? The USRP I am using is an Ettus USRP N210 with a WBX daugherboard.

Thank you,

Rich

[Anthony_F · Accepted solution]

The first thing I notice is that you do not use the coerced IQ rate output of the niUSRP Configure Signal VI.  Right now your requested IQ rate is 204.8k and your actual IQ rate is 204.918k.

• You will need this because the USRP cannot support all IQ rates. Use the MT Resample Waveform function to resample the waveform just before the niUSRP Write Tx Data VI.

If you would like a second copy of your data at - 49.125 kHz then you will need to also upconvert to -49.125 kHz and add these waveforms together. 

I would suggest taking an FFT of the data output from the MT Upconvert VI. If you don't have negative frequency components at baseband, I wouldn't expect to see negative frequency components at RF 😉

---

@RichF88 wrote:

  Is there any way to force it into the FPGA in the USRP hardware? The USRP I am using is an Ettus USRP N210 with a WBX daugherboard.

---

The NI-USRP driver does not support modifications to the FGPA.  You can however use GNU Radio and UHD if you wish to modify the FPGA.

[RichF88]

Anthony,

Makes sense to me. Thank you for your help.

[wicak]

so,, MT resample could force the IQ rate that cannot be supported by write tx ?

How is it possible?

And the desired sample rate, is it the desired IQ rate (samples/symbol), or symbol rate (symbol/second), which one is correct?

So if I want set my IQ rate 8M samples/symbol, sampling rate 8 samples/symbol, the coerced IQ rate of my USRP is 961.538k .

I want my symbol rate to be 1M, how could I achieve that?

[Anthony_F]

---

@wicak wrote:

so,, MT resample could force the IQ rate that cannot be supported by write tx ?

 

---

Modulation Toolkit might create waveforms with an incompatible IQ rate.  MT Resample Waveform forces the IQ rate so that niUSRP Write Tx Data can support it. Remember the IQ rate (same as sample rate) is expressed as samples/second, not symbols/second.

If you would like a symbol rate of 1Msym/sec then multiply this by samples per symbol to determine the minimum IQ rate needed. Use the coerced IQ rate to resample the waveform to the abilities of the USRP.