ADC12DJ3200: HSDC Pro complex FFT strange result

Yun Xu

Part Number: ADC12DJ3200

Hi support team,

I make my own data file and import into HSDC Pro. The signal to be sampled is 400MHz(-5dBm), JMODE16，Fs=3200MSPS, NCO is 400MHz, GAIN BOOST is on.

The device parameter settings:

Is there something wrong with my settings or hardware?

Regards

Joseph

over 4 years ago

0 Rob Reeder over 4 years ago

TI__Mastermind 38130 points

Hi Joseph,

We are looking into this, in the meantime, it might be good to try importing a file without the NCO on to make sure the file format is valid and you can see a normal FFT. Is that something that can be done in the meantime?

Regards,

Rob

0 Yun Xu over 4 years ago in reply to Rob Reeder

Prodigy 140 points

Hi Bob,

You mean without NCO on like this:

However, here are some updates:

1) I tried to import capture data, of the signal 400MHz(-5dBm), JMODE16，Fs=3200MSPS, NCO is 400MHz, GAIN BOOST is on and set device parameteres like:

and the complex FFT result look like:

2) I tried to import capture data, of the signal 400MHz(-5dBm), JMODE16，Fs=3200MSPS, NCO is 425MHz, GAIN BOOST is on and set device parameteres like:

and the complex FFT result look like:

Why there exist 50MHz and -50MHz spectrum? Should there be 25MHz or -25MHz?

I also attached these 2 data files, please feel free to test:

Fin400MHz， NCO425MHz:

/cfs-file/__key/communityserver-discussions-components-files/73/jm16_5F00_fs3200M_5F00_fin400M_5F00_fnco425M_5F00_1.csv

Fin400MHz, NCO400MHz:

/cfs-file/__key/communityserver-discussions-components-files/73/jm16_5F00_fs3200M_5F00_fin400M_5F00_fnco400M_5F00_1_5F00_hsdc.csv

Expecting your reply. Thanks.

Regards

Joseph

0 Yun Xu over 4 years ago in reply to Rob Reeder

Prodigy 140 points

Hi Rob,

I made some progress, and new question came up:

1) when Fin=350M, Fnco=400M, result seems to be ok:

2) when Fin=375M, Fnco=400M, there are 25M and -75MHz:

3) when Fin=402M, Fnco=400M, there are -10MHz, -2MHz, 6MHz, 14MHz, 22MHz....:

It looks like 4*(Fin-Fnco) spectrum interval. What could be the reason?

Here are the data files:

1) Fin350M-Fnco400M:

/cfs-file/__key/communityserver-discussions-components-files/73/jm16_5F00_fs3200M_5F00_fin350M_5F00_fnco400M_5F00_1_5F00_hsdc.csv

2) Fin375M-Fnco400M:

/cfs-file/__key/communityserver-discussions-components-files/73/jm16_5F00_fs3200M_5F00_fin375M_5F00_fnco400M_5F00_1_5F00_hsdc.csv

3) Fin402M-Fnco400M:

/cfs-file/__key/communityserver-discussions-components-files/73/jm16_5F00_fs3200M_5F00_fin402M_5F00_fnco400M_5F00_1_5F00_hsdc.csv

Regards

Joseph

+1 Neeraj Gill over 4 years ago in reply to Yun Xu

TI__Mastermind 20541 points

Hi Joseph,

I think it looks you are saturating the ADC. If you look at the time domain view you can see input signal is too close to the full scale value of the ADC.

Try capturing the signal with reduced analog input power and see if you get better plots.

Regards,

Neeraj

0 Yun Xu over 4 years ago in reply to Neeraj Gill

Prodigy 140 points

Hi Neeraj,

Thank you for your tips. The codes are indeed saturated. Originally, I didn't turn SFORMAT on. After your suggestion, I turned SFORMAT on, and I got signed codes which look much smaller, and I got beautiful results like this:

Thank your again!

Regards

Joseph

Data converters

Data converters forum

ADC12DJ3200: HSDC Pro complex FFT strange result