ADS42LB49: ADS42LB49 FFT power spectrum problem

Mahmoud Zannerni62

Intellectual 655 points

Part Number: ADS42LB49

Hello,

I am using ADS42LB49 at 100 MHz sampling rate and an FPGA to perform FFT.

The power spectrum does not look uniform when sweeping a signal from 0 to Fs/2 on the input.

It is lower on the low end and the high end of the spectrum as in the attached drawing.

Is this related to the ADC specs? I know if the clock jitter is high it can effect the high frequencies but not sure why the low frequencies are lower as well.

Regards,

over 4 years ago

0 jim s over 4 years ago

TI__Guru* 91947 points

Mahmoud,

The ADS42LB49 has an input buffer which has an auxiliary path which results in a high pass type of response at low frequencies. This can be corrected/compensated either by adding extra circuit on board or by digital processing. See attached app note for info regarding this.

Regards,

Jim

app note.pdf

0 Mahmoud Zannerni62 over 4 years ago in reply to jim s

Intellectual 655 points

Thanks Jim,

Does ADS42LB49EVAL board offer the option to add rc filter as in the doc?

Regards,

Mahmoud

0 jim s over 4 years ago in reply to Mahmoud Zannerni62

TI__Guru* 91947 points

Mahmoud,

Not really. The best you can do is install the appropriate capacitor at C5 and C14 then attempt to solder the appropriate resistor across each of the caps.

Regards,

Jim

0 Mahmoud Zannerni62 over 4 years ago in reply to jim s

Intellectual 655 points

Hi Jim,

Does the Fc changes when the sampling frequency change?

Also, can using IIR filter in digital domain cause any instability issues?

Regards,

Mahmoud

0 jim s over 4 years ago in reply to Mahmoud Zannerni62

TI__Guru* 91947 points

Mahmoud,

What is Fc? For the second question, I would say no per the attached document I sent you earlier.

Regards,

Jim

0 Mahmoud Zannerni62 over 4 years ago in reply to jim s

Intellectual 655 points

Hi Jim,

I was referring to cut frequency of the buffer high pass response.

I have attached a copy of the FFT power output, it looks the FFT power reach its maximum value when the input is higher than approx. 8 MHz

Does the high pass filter effect the input up to 8 MHz?

Also, do I need a better clock source to improve the response on inputs over 25 MHz.

Regards,

Mahmoud

0 jim s over 4 years ago in reply to Mahmoud Zannerni62

TI__Guru* 91947 points

Mahmoud,

This part was designed to have the best performance with IF inputs between 50MHz and 170MHz. What you are seeing at low IF may be as good as it gets. Per Figures 27 and 28 in the data sheet, increasing the clock amplitude should give you better performance.

From data collected shown in Figure 1 of the attached app note, it appears the high pass filter only affects signals lower than 1.6MHz.

Regards,

Jim

5428.app note.pdf

0 Mahmoud Zannerni62 over 4 years ago in reply to jim s

Intellectual 655 points

Thanks Jim,

I was using 1.5 v p-p clock when doing the FFT.

In my application the most important band is 500 KHz to 10 MHz, and I want to be able to achieve 200 MHz sampling frequency.

Is there any better part from TI for this sort of application? with similar cost and an eval with FMC.

Regards,

Mahmoud

0 jim s over 4 years ago in reply to Mahmoud Zannerni62

TI__Guru* 91947 points

Mahmoud,

We have a couple of parts coming out soon that may work. See attached preliminary data sheet.

Regards,

Jim

adc3541.pdf

Data converters

Data converters forum

ADS42LB49: ADS42LB49 FFT power spectrum problem