cancel
Showing results for 
Search instead for 
Did you mean: 
Visitor
Visitor
253 Views
Registered: ‎08-05-2019

Strange noise structure in ZCU111

Jump to solution

Hi,

I have been working with ZCU111 RFSoC evaluation board and now facing with strange noise problem.

It happens when I get data with lower sampling rate with using "RF Data Converter Evaluation User Interface".

The data at 2.048Gsps, there is strange structure even thouth I terminate the ADC input with 50 ohm.
However, the data at 4.096Gsps, there is no strange structure.

2048Mspnoise 2020-01-31 144053.png4096Mspnoise 2020-01-31 144053.png

Of course, in FFT view, noise level in 2.048Gsps is higher than in 4.096Gsps (about 15dBFS)

I also checked this strange structure can be seen in other channles on other tiles.
And it can be erased with using decimation. (ex. There is no strange structure in the data at 4.096Gsps with 2x decimation. This effective sampling rate is 2.048Gsps.)

4096Msp_Eff2048Msps_noise 2020-01-31 144053.png

So, I would like to know
1. Is this noise is because of some kind of "sampling principle"?
2. If not, is this normal in RFSoC or just my board failure?

I don't have much knowledge of RF signals......
Many thanks!

Tags (3)
0 Kudos
1 Solution

Accepted Solutions
Highlighted
Moderator
Moderator
221 Views
Registered: ‎08-08-2017

Re: Strange noise structure in ZCU111

Jump to solution

Hi @n_particle 

Looking at the time domain captures you have shared , it looks to me ADC callibration is not converging. What is the frequency specifications of signals you are trying to capture ?

1. To perform accurate calibration of the sub channels for gain and time skew, requires diversity of data points passing through sub channels.

2. When the ADC input frequency has a strong correlation to the sample rate, such as a common denominator, this leads to low diversity of data in interleaved   channels, or an insufficient number of unique data points for the calibration to accurately converge

The ADC callibration is documented in Product guide PG269

https://www.xilinx.com/support/documentation/ip_documentation/usp_rf_data_converter/v2_2/pg269-rf-data-converter.pdf

 

 

-------------------------------------------------------------------------------------------------------------------------------
Reply if you have any queries, give kudos and accept as solution
-------------------------------------------------------------------------------------------------------------------------------

View solution in original post

2 Replies
Highlighted
Moderator
Moderator
222 Views
Registered: ‎08-08-2017

Re: Strange noise structure in ZCU111

Jump to solution

Hi @n_particle 

Looking at the time domain captures you have shared , it looks to me ADC callibration is not converging. What is the frequency specifications of signals you are trying to capture ?

1. To perform accurate calibration of the sub channels for gain and time skew, requires diversity of data points passing through sub channels.

2. When the ADC input frequency has a strong correlation to the sample rate, such as a common denominator, this leads to low diversity of data in interleaved   channels, or an insufficient number of unique data points for the calibration to accurately converge

The ADC callibration is documented in Product guide PG269

https://www.xilinx.com/support/documentation/ip_documentation/usp_rf_data_converter/v2_2/pg269-rf-data-converter.pdf

 

 

-------------------------------------------------------------------------------------------------------------------------------
Reply if you have any queries, give kudos and accept as solution
-------------------------------------------------------------------------------------------------------------------------------

View solution in original post

Highlighted
Visitor
Visitor
175 Views
Registered: ‎08-05-2019

Re: Strange noise structure in ZCU111

Jump to solution

Thank you for replying, @pthakare.

I checked the calibration section in PG269 and noticed there is description about Dithering.

calibration 2020-02-03 174123_LI.jpg

This section says that I need to disable dithering when I use the ADC at 2.048Gsps.(ZU28DR maximum sampling rate is 4.096Gsps) I completely overlooked this description.

After disabling "dither", the noise structure becomes normal.

Thank you for your help!