Clockworks presents at AES NYC 2018: Resistor 1/f noise

Systems built using Analog Devices' Automotive Audio Bus® (A2B), SHARC® Audio Module, and Clockworks SignalBlox® product line will almost certainly have analog inputs and/or outputs. Those interfaces need to deliver specific real world performance numbers to meet their system requirements. Among those is typically a signal to noise (SNR) or dynamic range (DR) specification.  All electronic components above absolute zero create noise and most engineers are familiar with Johnson-Nyquist (white) noise.  You may be a little less familiar with 1/f (pink) noise that resistors create.

At AES 145 in New York City in October 2018 Clockworks presented a poster session on 1/f noise in resistors, the associated paper (open access) provides a background on 1/f noise as well as an experiment you can perform on your computer to determine if it would be audible.

Clockworks products generally use precision metal film resistors in the signal path to minimize the impact of 1/f noise. If designing circuitry it's not needed to use metal film everywhere (versus thin film) as low values and ones with very small voltages across them will contribute very little to system noise.  Outside of the signal path any resistor type is generally sufficient; metal film can cost 10x the price of more generic thin film parts so that must be factored in to the design decision. 

About the power resistors in the picture:

They really don't have anything directly to do with the 1/f noise topic as wirewound parts have almost non-existent 1/f noise.  Usually; as discussed in the paper testing showed that not all wirewounds behave as they should.  However wirewound parts can have other unexpected behaviors that make them slightly non-linear so if using them in applications like dummy loads one must make sure the load resistors are not a source of error.  If you're interested in learning more about there are some posts on the Clockworks website.