Digitally-Assisted Analog Front Ends


When sensing signals in the physical world, undesired interfering signals often arise with an amplitude much larger than the targeted analog signal of interest. For example, the signals received by an antenna include many channels with widely varying signal strength, but the receiver may be required to listen to one of the smallest. In sensor interfaces, the sources of interference are manifold including power supply noise, electrical noise from neighboring circuits, and even mechanical sources of interference. Digital signal processing may be used to ultimately extract our target signal, but this requires digitization of both the desired information-bearing signal, and the dominant sources of interference, thus placing considerable burden on the analog front-end and analog-to-digital converter (ADC).


We have developed low-power analog circuitry that identifies sources of spectral interference, as in the prototype analog circuit above. [1]  Future work includes the integration of this circuitry together with analog-to-digital converters. Much of the same circuitry can be reused for both spectral estimation and digitization.

Techniques like this may be combined with mixed analog/digital cancellation of the interference, as illustrated below.  Doing so can eliminate the dominant interference from the analog signal path prior to digitization, relaxing the requirements on analog amplifiers, filters, and ADCs.


Applications of interest include RF receivers for blocker cancellation, and sensor interfaces such as touch screen controllers where the analog front end must typically accommodate large common-mode and differential interference from supply noise, ground bounce, and the neighboring display.

[1] K. Banovic, A. Chan Carusone, “A 1.55mW Mixed-Signal Integrating Mixer for Direct Spectrum Estimation in 0.13um CMOS,” Asian Solid-State Circuits Conference, Kobe, Japan, November, 2012.