A new audio amplifier topology - Part 4: Noise in folded cascode stages

Voltage reference elements

There are various types of voltage reference elements available, which offer different tradeoffs for a low noise cascode implementation. Forward-biased diodes (standard small-signal parts or light emitting diodes) usually offer low voltage noise; however, their forward voltage is low, and several parts might need to be put in series to realise a sufficiently high emitter resistor value. Besides increasing the voltage noise generator E_nV, this also augments incremental impedance RV* and its detrimental effect on noise as shown in (4).

Zener (and avalanche) diodes are available with a wide range of voltages. Yet their voltage noise is at least an order of magnitude, for parts with high reverse breakdown voltage even more, higher than that of forward-biased diodes. Inmost cases this will necessitate the use of an RC filter to make this noise contribution negligible at audio frequencies. Similar concerns apply for the various bandgap voltage references available as integrated circuits.

If a differential folded cascode (as e.g. formed by Q3 and Q4 in figure 5 of the main text) is used, the noise contribution of the common voltage reference is theoretically rejected by the common-mode rejection of the following stage (e.g. a current mirror). However, as the common-mode rejection is subject to resistor and transistor tolerances, it is not very dependable without further precautions. Hence sufficient attention to the performance of the voltage reference must be given nonetheless.

From (1) to (4) it is seen that for increasing R* values the total noise contribution of a folded cascode with given quiescent current converges towards I_nQ. Hence the last step in minimising folded cascode noise is the choice of a transistor with low current noise. This corresponds to a device with high h_FE and low excess noise. Note that the choice of a transistor with low voltage noise is not necessary; once the emitter resistor R* is chosen large enough, the contribution from E_nQ is easily reduced well below that from E_nQ.

By now it might have occurred to you that figure A1 also represents a typical current source implementation; a folded cascode is essentially a current source, where the emitter of the pass transistor is used as input node. Hence all the noise reduction strategies outlined above are applicable to the design of current sources as well. This however is often of less practical relevance for power amplifier design. The noise of the current source which biases the input differential pair is mitigated by the common-mode rejection of the amplifier; current sources in later stages are fed to nodes with comparably low impedance, and also a substantial amount of loop gain is available to reduce any remaining contribution.

Even current mirrors are closely related to figure A1 - just consider that the voltage reference is implemented with a series-connection of a resistor and a forward-biased diode. The base of Q* then forms the input node, and the collector of Q* the output node of a standard Widlar current mirror. Noise reduction in current mirrors is hence again done by the means outlined above. As the voltage reference is now replaced with a diode and a resistor with fixed relation to R*, the available design freedom reduces to the choice of large resistor values and a pass transistor with low current noise.

About the author
Samuel Groner was born and currently lives in Zurich, Switzerland. He has been passionate about both art and science as long as he can remember. At present he works for Weiss Engineering Ltd. in the field of analogue hardware design and freelances as classical recording engineer/producer. Besides this, he teaches several courses at a local sound engineering school (ear training, classical music production and audio measurement) and enjoys a manifold activity as pianist, singer and choirmaster. If time permits, he is found on one of the numerous Swiss hiking trails, preferably in company with one of his cameras and a few sheets of black-and-white film. He holds a MSc degree in computer science and a MA degree as Tonmeister (recording engineer/producer).

This article originally appeared in Linear Audio Volume 2, September 2011. Linear Audio, a book-size printed tech audio resource, is published half-yearly by Jan Didden.

For more articles like this and others related to audio design, visit Audio Designline and/or subscribe to the monthly Audio newsletter (free registration).