About ADC (in research and in industry)

[currentmirror2000]

Hi all,

does anybody have any information on multi-channel (8 or 16 or 32 or even 64) high speed (>>1Msps) ADC in on going research or in industry?

from the commercial side, i only found some products from TI like ads5273 which has 8 channels and 70Msps (8 ADCs without MUX). can i have more channels and faster sampling rate in one chip? i'm searching the research area on this topic, but currently got no information. so what's the limitation on the number of channels? is it area or power or commercial reason?

thanks in advance!
py

 

[currentmirror2000]

need help...

 

[brmadhukar]

hi check also the Analog devices website.
But normally you dont get better than 8 -12-15 channels nad very high sampling rate. This is to reduce the aperture effect and subsequent correlation among the samples.
brmadhukar

 

[currentmirror2000]

hi brmadhukar, thanks for your reply.

i've had checked analog devices also but there is no high speed (>10Msps) multi-channel ADCs. in TI, there are some ADCs with 8 differential input, and every channel can operate at 40-70Msps concurrently, which means the overall samples/s is greater than 300-500Msps. the overall power consumption is also high, ranging from 0.5W to 1W. there is no MUX inside this kind of ADCs.

my idea is to put multiple ADCs in paralell without using MUX, just like TI's but with much more channels, so for each channel the sampling rate will only be limited by the respective ADC(S/H, comparators etc...).

there is no correlation among different channels and the method is just simply putting more ADCs into one single chip. i know there are some applications need several channels working at the same time, but what people do normally is to use several single-ADC chips in parallel rather than use one ADC chip with parallel channels.

so does it mean the aperture delay will not be an issue since even by using several single-ADC chips in parallel, the effect also exists.

my prime concern is the power and area constraint. what's your opinion?

py