RF Measurement Devices

cancel
Showing results for 
Search instead for 
Did you mean: 

5644R DC Offset on Streaming

Hi:

 

I am using 5644R to do signal recording project. However I find that the DC offset of 5644R changes with time. 

 

I use the example vi -- RFSA Acquire Continuous IQ.vi and terminate the RF input of the 5644R. I plot the mean value of I & Q every read block. After 1000 reads,  I get a sine wave. The frequency is about 0.5Hz and amplitude is about 0.3mV. How this happens ?

 

I think the DC offset of Zero IF system should be a const and I can calibrate the DC offset by removing the average of the record data. But if it is changes with time, I can't do the calibration.

 

Hope for the help. Thank you

0 Kudos
Message 1 of 5
(3,753 Views)

Can you post a screenshot of the actual IQ data you are seeing?

 

Also, have you seen the behavior with an actual test signal (instead of just a terminated input)?  You can test this by putting the VST in loopback (RF Out to RF In) and using the RFSG Soft Front Panel to generate a CW at the center frequency and power level of your choice.

Cody A.
0 Kudos
Message 2 of 5
(3,531 Views)

Also, can you try adding a center frequency offset? (upconverter center frequency offset attribute in RFSG, downconverter center frequency offset in RFSA) Try 10 or 20 MHz and see if the problem still exists.

0 Kudos
Message 3 of 5
(3,527 Views)

Here is the vi and screenshoot.

It is from the RFSA Acquire Continous IQ example. And I only add average to get DC offset.

I need to use the 5644R as a receiver, and I don't think I should connect Rx with Tx.

The 5644r works well on RFSA SFP, without any DC offset. However it is not so good when comes to continous IQ mode.

Download All
0 Kudos
Message 4 of 5
(3,506 Views)

I think what you're seeing is frequency error and possibly phase noise. I changed the center frequency based on the frequency error reported by your code and got a MUCH slower oscillation (see the attached screenshot and note the x-axis length).

frequencyError.PNG

 

I don't think that you'll be able to remove the DC offset completely if you're centered on the LO. However, the VST supports the super-helpful downconverter center frequency offset attribute. As long as you can fit your signal of interest into half of the 5644R's bandwidth (so max signal bandwidth of 40 MHz), you can just move the LO out of the way. See below where I set the DCFO to 10 MHz:

dcfo.PNG

And if 40 MHz isn't enough bandwidth, there's always the 5646R (100 MHz on each side of the LO) or the 5840 (up to 500 MHz on each side of the LO).

0 Kudos
Message 5 of 5
(3,494 Views)