rfsa captured power is different in different iq rate

[wei-smart]

Dear All,

The problem is as title, detail steps is listed below:

HW: 5840 loopback

steps:

1. rfsg sfp send 1GHz -10 dBm

2. use rfmx chp example to capture power

3, result is as below: 100M and 10M span will get different power.

[YouZi]

 Hi Wei-Smart,

Could you offer the RFmx version?

Thanks

[MMargaryan]

Hello wei-smart.

The bigger span is, the more noise it takes, so consequently noises are added and you are getting different power.

Below you can find useful link regarding the RFmx Measurements.

Optimizing NI-RFmx Measurement for Speed

Hope, it will help you in the future!

Regards, 

Meri

[wei-smart]

Hey you zi,

Thanks for your reply.

it's common behavior in NI instruent in Any RFmx version. you can reproduce it easily. Thanks.

[wei-smart]

Dear MMArgaryan,

Thanks for your reply.

My set up is VST loop back, there's very limited external noise.

Higher sample rate will get same PSD.

For CHP, integration bandwidth is same.Just change span

Even with higher sample rate we get more noise, then why higher sample rate get less power?

That's my confuse part. thanks

[YouZi]

Hi Smart,

These variations in different SPAN on Spectrum Mode do happen.

And on IQ Mode with different IQ rete, the power is stable. 

[YouZi]

Hi Smart,

RFSA spectrum mode, delivers to the API a spectrum assembled of many acquired smaller spectrum, that are not centered at the LO frequency of the instrument. This is done automatically and without your control, but what that means is that your tone will lie at different frequencies in the instantaneous BW of the HW under different conditions. 

About instantaneous BW, please refer to Help.

Here is an example which will simulate illustrating what RFSA does and will measure the actual Ripple of the passband which will show the same variation, but it will be easier to visualize as there won't be any discontinuities like your seeing in this graph.

[wei-smart · Accepted solution]

Dear All, Thanks for your support This problem is caused by LO offset, NOT sample rate, thanks.