FPGA Central

Hi all!
I have a 400MHz bandwidth analog real waveform & I want to digitis
it using ADC. So i need to sample it at least at 800MHz. But I dont have
suitable ADC to do that (I need to maintain a high dynamic range of aroun
75dB requiring a ADC chip of around 16bits which can sample at 800MHz o
more which i dont have). What i have instead is a 4 channel ADC which ca
sample at upto 200MHz (each channel) . My question is can I sample th
wideband waveform using 4 channels simultaneously sampling at 200MHz (eac
channel), something akin to quadrature sampling , and then combine the
channel streams to get back my original wideband waveform ? Is there an
tangible literature which states how this can be done? ( I have read mos
of the basic sampling theorms papers which give information about highe
order bandpass sampling and interpolants but had a hard time understandin
them)So i am looking around if any of u guys have done this before.

Note: The wideband waveform which i have mentioned is bandpass wavefor
centred around some freq fc>>400MHz

Regards,
Yogesh

"yogesh_gharote" <[email protected]> wrote in message
news:[email protected] ...
> Hi all!
> I have a 400MHz bandwidth analog real waveform & I want to digitise
> it using ADC. So i need to sample it at least at 800MHz. But I dont have a
> suitable ADC to do that (I need to maintain a high dynamic range of around
> 75dB requiring a ADC chip of around 16bits which can sample at 800MHz or
> more which i dont have). What i have instead is a 4 channel ADC which can
> sample at upto 200MHz (each channel) . My question is can I sample the
> wideband waveform using 4 channels simultaneously sampling at 200MHz (each
> channel), something akin to quadrature sampling , and then combine the 4
> channel streams to get back my original wideband waveform ? Is there any
> tangible literature which states how this can be done? ( I have read most
> of the basic sampling theorms papers which give information about higher
> order bandpass sampling and interpolants but had a hard time understanding
> them)So i am looking around if any of u guys have done this before.
>
> Note: The wideband waveform which i have mentioned is bandpass waveform
> centred around some freq fc>>400MHz
>
> Regards,
> Yogesh
>
>
In principle, this can be done. In practice, matching or calibrating the
ADCs and the sampling
instants to the required precision for your dynamic range is a bear.

There was an article in the Agilent Technologies Measurement mag a couple of
years ago about this technique as used in their high speed scopes, it might
be on
the T&M section of their website. There's also a patent by Howard Hilton on
calibration of a pair of ADCs used like this (7049992).

Regards
Ian

yogesh_gharote wrote:
> Hi all!
> I have a 400MHz bandwidth analog real waveform & I want to digitise
> it using ADC. So i need to sample it at least at 800MHz. But I dont have a
> suitable ADC to do that (I need to maintain a high dynamic range of around
> 75dB requiring a ADC chip of around 16bits which can sample at 800MHz or
> more which i dont have). What i have instead is a 4 channel ADC which can
> sample at upto 200MHz (each channel) . My question is can I sample the
> wideband waveform using 4 channels simultaneously sampling at 200MHz (each
> channel), something akin to quadrature sampling , and then combine the 4
> channel streams to get back my original wideband waveform ? Is there any
> tangible literature which states how this can be done? ( I have read most
> of the basic sampling theorms papers which give information about higher
> order bandpass sampling and interpolants but had a hard time understanding
> them)So i am looking around if any of u guys have done this before.
>
> Note: The wideband waveform which i have mentioned is bandpass waveform
> centred around some freq fc>>400MHz

Probably not. First, you need to sample significantly faster 800MHz in a
practical system. (800MHz is a theoretical lower bound under conditions
that are ideal in several respects, none of them realized in practice.)
Second, you would need to synchronize the ADCs' operations more
precisely than their jitter and aperture uncertainty would likely allow.

Jerry
--
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯ ¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯