More Bandwidth needed (more than x310 is supporting).Should I wait or should I modify the board?

[booth]

Dear All,

 The x310 carries the UBX 160 board which support 80 Mhz I-Q bandwidth. The ADC inside x310 (ADS62P48) supports sampling up to 210 MSPS.

I have a 85 MHz wide QPSK signal centered at 700 MHz to demodulate. My assumption is that the x310 can demodulate an 80 MHz wide QPSK signal.

Should I wait for a wider bandwidth radio? I don't know if Ettus-NI plan to release a radip with more bandwidth. Any Ideas?

In the mean time I can open up the UBX 160 board, bypass the 80 MHz LPFs and change the ADC with a pin compatible 250 MSPS ADS62P49 and play with the clocking and FPGA code.

Hopefully that would give me the possibility of demodulating 100 MHz wide signals.

Any thoughts? 

[Shalini.M]

Hi Booth, 

I was looking through the Ettus website and found a UBX 160 board that supports up to 160 MHz of Instantaneous bandwidth: https://www.ettus.com/product/details/UBX16

Does that better meet your application needs? 

[booth]

the UBX 160 board has 80 MHz LPF for the I and Q lines. I assume the 160 MHz bandwidth is for TX mot RX.

[muffin.vi]

The UBX board has 160 MHz for both TX and RX. 

[alynch]

The note at the bottom of the UBX-160 product page says that the 84MHz RX bandwidth limitation is only for frequencies from 10MHz to 500MHz. Since you are working at 700MHz, you will have access to the entire 160MHz bandwidth.

[booth]

The UBX 160 board has 80 MHz LPF for the I and Q lines. For each line there is a 210 MSPS ADC.

I argue that we can observe 160 MHz bandwidth but demodulate only a 80 MHz wide signal.

Am I missing something fundamental here??

[muffin.vi]

I'm a bit confused here. In what situation is it possible to observe/digitize 160 MHz of bandwidth but only to demodulate half of it? 

[booth]

---

@muffin.vi wrote:

I'm a bit confused here. In what situation is it possible to observe/digitize 160 MHz of bandwidth but only to demodulate half of it? 

---

In the QPSK (or other higher order modulation) case. If we have a 160 MHz wide QPSK modulated signal, the I&Q channels would be 160 MHz wide. Since there are 80 MHz LPFs for each channel only half of it would go through.

Sampling wise: Since the ADCs are 210 MSPS (I&Q), the sampling theorem says that you can only reconstruct 105 MHz wide signals (If you remove the LPFs). That makes again a 105 MHz wide QPSK signal that you can demodulate.

In the BPSK case you actually can demodulate 160 MHz wide signals.

[Shalini.M]

Hi Booth, 

Can you give us more context on your application and what you are trying to do? 
Also, have you gotten a chance to test this with the hardware and at that point did you see your signal being truncated/attenuated? I would be interested in seeing screenshots of the plots/results if you could attach those. 

[stan211]

Hello,

Because you have 80 MHz bandidth in each of the I and Q streams, the total bandiwdth per channel seems to be 160 MHz, as suggested by the spec sheets. Nyquist is equal to sampling rate/2 for real signals, but for complex signals it is equal to the sampling rate.

Stan