How to accurate measure IQData?

[billycardiff]

Hi:

We got a 5610 and 5441 together as a signal generator and 5600 and 5142 as signal analyzer. when I generate a two-tone multi-sines with the signal generator, and I want to caompare the difference between the signal out the generator and the signal in the receiver. however, I found the IQ data I received at the signal analyzer side is keep moving and not similiar to the original one. Any reason? I used the built-in example of "uniform space multi-sine generation " and "read IQ into disk"

Thank you very much!

[DanBrowne]

Hi,

I was just wondering what cables you are using to connect the signal generator side to the reciever side. For instance, what are the cable lengths, and termination resistances?

Regards,

Daniel

[billycardiff]

Hi:

Thank you for your reply. the cable I am currently using is 50OHm ad 15cm length. But then I also tried different cables with different length but all with 50ohm resistance, all of them get the same result. So I do not think it was caused by cable.

I generate two-tone signal with 100k space at 100M carrier frequency. at the generator side, the actual sample rate is 1M/s. Then I used 100M carrier frequency and 1M sample rate at the receiever side.

[billycardiff]

Hi:

I trie to snap the screen to make things more understandable.

[DanBrowne]

Hi,

Thanks for posting the screenshots? Would you be able to post your VI(s) so that I can look at your code directly? Maybe i'll then be able to diagnose the issue.

Regards,

Dan

[billycardiff]

Hi:

I was looking through this problem, it looks like unlocked RF local oscillator (LO) introduced frequency offsets and thus make phase drifting. What do you think? The vi is attached below. those vi are directly copy from built-in examples.

Thank you!

[billycardiff]

Hi:

I thought it might due to the LO frequency offset at downconverter part. I found there is a example of demodulating QPSK signal, which seems provide a solution to fix that. However, I don't know whether it could be applied to multi-tone signal?

Thank you!

[DanBrowne]

Hi,

Thanks for the added information - I am still currently looking into why you are seeing this problem,

Regards,

Daniel

[DanBrowne]

Hi again,

I have a suggestion for you to try. After having a look at your configuration in your screenshots, you appear to be using the same IQ Rate (1MS/s) on both the generate and recieve sides. This could be why you are seeing different signal shapes and phase relationships on the reciever side of the system. Generally, you should be oversampling by a factor of between 5 and 10 times the generation frequency on the reciever side in order to get an accurate reproduction of the signals. In your case, therefore, you could either reduce the generation IQ Rate to 1/5 or lower of its current value, OR increase the reciever side IQ rate to 5-10 times it current value. This would achieve a suitable oversampling ratio.

If you could try this, to see if it works, then that would be great.

Regards,

Dan

[billycardiff]

Hi:

Thank for your reply. Just one question: can the 5660 series block set the sample rate? I can not find the revelant block to set up the sample rate. it looks only available for maximum 1.25MHz sample rate, is it right?