ADC12DJ5200RF: operation principle for high frequency sampling

Koji Ikeda

Part Number: ADC12DJ5200RF

Hello support team,

My customer inquires ADC12DJ5200RF sampling frequency.

Datasheet shows "The ADC12DJ5200RF device is an RF-sampling, giga-sample, analog-to-digital converter (ADC) that can directly sample input frequencies from DC to above 10 GHz.". However, in customer understanding, more than 20GHz sampling frequency would be required to convert 10GHz analog signal. How ADC12DJ5200 achieve over 10GHz samping, while ADC12DJ5200 support up to 10.4GSPS? They would like to know the operation principle.

Thanks,

Koji Ikeda

over 3 years ago

0 Rob Reeder over 3 years ago

TI__Mastermind 20170 points

Hi Koji,

The 5200RF ADC is more like 7-8GHz if 3dB analog input BW.

The ADC is sampling at 5200MSPS max, the 10.4GSPS sampling is achieved by taking the two ADC channels and interleaving them.

So the sample settling time is within bounds for this BW. Effectively you can take any ADC and interleave them together to open up the Nyquist BW and achieve faster sampling. However, this doesn't mean the analog BW, whatever it maybe, get larger/bigger as well.

Hope this helps.

Regards,

Rob

0 Koji Ikeda over 3 years ago in reply to Rob Reeder

TI__Expert 5335 points

Hello Rob-san,

Thank you for your support.

In general case, acceptable analog conversion frequency is half less than sampling frequency. 10GSPS would support less than 5.2GHz analog signal. However, 5200RF can support 7-8GHz analog input by 10.4GSPS. How can 5200RF achieve such a higher frequency sampling?

Thanks,

Koji Ikeda

0 Rob Reeder over 3 years ago in reply to Koji Ikeda

TI__Mastermind 20170 points

Hi Koji-san,

I believe there is confusion about Nyquist rules. I would review the following papers/articles to better understand this concept. See links below.

Nyquist rule states that the signal of bandwidth (BW) must be sampled at a rate equal to or greater than twice its bandwidth (2BW) in order to preserve all the signal, IF, information.

Under-sampling can exist because there is no mention of the absolute location of the band of sampled signals within the frequency spectrum relative to the sampling frequency.

Hope this helps.

Regards,

Rob

https://training.ti.com/bandwidth-vs-frequency-subsampling-concepts?context=1139747-1140267-1128375-1139110-1134477

https://www.analog.com/media/en/technical-documentation/application-notes/577685821057469457705578853404381574892535191189340957AN_827_0.pdf, page 3

Data converters

Data converters forum

ADC12DJ5200RF: operation principle for high frequency sampling