Fundamental Limits of ADCs and DACs

Does anyone know what the fundamental limits of DACs and ADCs are?


Here’s an interesting article from 2000 that shows a 10 bit 1GHz ADC as difficult to impossible, and predicts that it will improve by 1 bit for every 2 years. So we should now have a 16 bit 1 GHz adc.

Here’s a 16 bit 250 MSample/second ADC from analog devices. It costs $100 and is the highest performance ADC analog sells.

Here is **broken link removed** module that claims to be the fastest commercially available 8 bit ADC in the world. This is roughly in line with where the guy was saying the technology was 12 years ago, so maybe it is slowing down as it reaches towards lower limits. The abstract for **broken link removed**claims that ADCs were within 1 order of magnitude of their fundamental power dissipation limit in 2006.

I've found much less information about the fundamental limits of DACs. Here is supposedly the **broken link removed**, at 12 GS/s. Here is a 1.25 GHz, 14 bit DAC from analog devices.

The underlying question in this post is, in the future will all radio/microwave signals be directly synthesized and detected, spelling the end of the super-heterodyne receiver as we know it? Or are there fundamental reasons that prevent that from happening?

As long as radio communication uses small to medium relative bandwidth, down- respectively up-conversion before AD/DA processing will most likely offer performance benefits at least starting from VHF bands. Frequency conversion isn't a big thing. IF bandwidth and dynamic of digital receivers marks the progress. As it increases, direct digital "UWB" processing comes in reach. But do you expect that all communication changes to UWB? Rarely.

It 's more likely, that small and medium bandwith transmission still keeps the major share. It is fine with some 100 kHz up to a few 10 MHz data rates. Direct digitizing would multiply their data rates by several orders of magnitude - and the power consumption in the digital domain as well.