Professionals interested in the quality of sound reproduction typically have a “trained ear” which is sufficiently perceptive to detect the insertion of noise and distortion of the original signal. The resistors in the preamplifier, audio amplifier, or volume control of a sound-reproduction system often cause these undesirable effects. As requirements for sound reproduction become increasingly demanding, the selection of circuit components for these audio systems is critical to minimizing noise and signal distortion that can compromise the fidelity of the end product.
Audio noise generated by resistors results from the motion of electrons within the resistor, which creates an unwanted AC signal which gets superimposed over the primary DC signal. This type of noise falls into two categories: thermal noise and current noise.
In the case of thermal noise, the random motion of electrons within the resistive grid causes the unwanted AC signal. The voltage developed by thermal agitation sets a limit on the smallest voltage that can be amplified without being lost in a background of noise. The equation that determines the amount of thermal noise is E2=4kTR (f2-f1), where k is Boltzmann's constant, T is the absolute temperature (in degrees Kelvin), R is the resistance of the conductor, and (f2-f1) is the bandwidth.
Current noise is the bunching and releasing of electrons associated with current flow. The amount of current noise (or lack thereof) depends largely on the resistor technology employed, and it is measurable and is expressed as a function of the input voltage. The magnitude is microvolts per volt applied. A noise index is expressed in decibels, and the equation converting μV/V to dB is:
dB = 20 x log (noise voltage [in μV]/DC voltage [in V]).
For example, 0 dB equates to 1.0 μV/V, and 15 dB equates to 5.6 μV/V.
To ensure high-fidelity sound reproduction, designers of audio systems, and especially the pre-amp, amp and volume control, can take advantage of the latest resistor technologies, their construction, and their effectiveness in preserving pure signals.
Evolution of Resistor Technologies
Carbon composition resistors
The carbon composition resistor was the mainstay of the radio and television industries prior to World War II. Their construction features carbon particles in a diallyl phtalate resin binder. The conductive path is from particle to particle, each of which touches another along the path.
Unlike the glass binder later used in thick film resistors, the resin binder of the carbon composition resistor is subject to mechanical movement relative to the carbon due to the forces created by voltage strain, moisture ingress, mechanical strain, and thermal strain. These strains cause the conduction sites at the point of contact to change the resistance or even open up. The current thus bounces around from one path to another with audible output, contributing not only to current noise but also to “popcorn noise.” Noise levels for carbon composition resistors ranges from -12 dB to +6 dB.
Carbon composition resistors were eclipsed in the early ‘60s by discrete metal film resistors. It was not noise levels but the rising cost of carbon composition resistors compared to falling prices for metal film devices that was the leading factor in their decline.
Thick film resistors
With a noise level range of -18 dB to -10 dB, thick film networks and discrete resistors made of thick film provide more noise than is acceptable in professional sound recording or reproduction applications. Once again, the construction of the resistor has much to do with the noise it produces. The resistive path in thick film products is an “ink” of glass frit with ruthenium oxide added and screened onto a ceramic substrate. The conduction path is through the oxide particles as they touch one another in the fired glass binder. These “touching sites” are locations for the bunching and releasing of electrons, and this is the reason for the noise in thick film resistors.
Metal film resistors
The introduction of metal film technologies brought significant reductions in both resistor size and noise. Metal film resistors are manufactured through the evaporation or sputtering of a layer of nickel chromium onto a ceramic substrate. The thickness of the layer is value-dependent and ranges from 10 Angstroms to 500 Angstroms thick. The thicker this layer is (the lower the value), the less noise is inserted. Higher values are noisier because the occlusions, surface imperfections, and non-uniform depositions are more significant to the production of noise when the nickel chromium layer is thin.
Grinding or laser adjusting techniques are used to generate the resistance grid. The first of these methods leaves a ragged edge and the second leaves a curled edge with eddy-current paths. Both are a source of noise, which is why metal film resistors have a noise range of -32 dB to -16 dB.
Wirewound resistors are made of alloys similar to that used in foil resistors, described below. As a result, the only noise insertion caused by these devices comes from the tabs used to connect the fine wire to the coarse external leads. The major objection to wirewound resistors, however, is unrelated to noise. A typical noise rating is -38 dB. Of serious concern instead is the inductance that chops the peaks and fails to replicate the higher frequencies of the second and third harmonics.
Resistors composed of metal or metal alloys display the lowest combined-noise level. Foil resistors, such as Vishay Intertechnology’s proprietary Bulk Metal Foil devices, take advantage of this in providing a low-noise solution for audio applications. They are resistive as a result of the inter-granular boundaries between conductive metallic crystals in the alloy. These boundaries are quite long and therefore mask any local site distortions, making these devices’ signal-to-noise ratio the best available.
Foil resistors are formed by a pattern etched in metal. This planar geometry and the devices’ two-axis design permit the current paths to be laid out in parallel, producing a self-canceling of inductance. Path-to-path capacitance is in series, which minimizes the occurrence of lumped internal capacitance. Finally, these low-inductance and low-capacitance resistors cause the least amount of peak-to-peak distortion with no measurable noise insertion. All of these characteristic combine to make foil resistors a clear choice for noise-free operation.
The "sound of silence"
The waveforms shown in Figures 1, 2 and 3, respectively, demonstrate a pure signal, the signal as it is affected by use of commercial components, and finally the signal as it appears when foil-based, "noise-free" resistors are substituted for commercial parts. Note: the scale for all three figures is 20 msec per division (x axis) and 20 mV per division (y axis).
Figure 1: Pure signal used for evaluation; input voltage (y axis) versus time (x axis)
Figure 2: Effect of commercial components on signal; output voltage (y axis) versus time (x axis)
Figure 3: Signal when foil-based resistors are used; output voltage (y axis) versus time (x axis)
The dramatic difference evident to the eye can be equally striking to the ear, even that of an untrained listener.
Noise-free resistors were developed primarily for the high-end sound-reproduction industry, where the purity of the audio signal is of paramount importance. Selection of the appropriate resistors for the signal path is the key in minimizing noise and meeting today’s exacting standards for sound quality.
While the high-end audio market has no single instrument to measure the fidelity of reproduction, the trained and experienced ears of audio engineers recognize the difference when equipment is using foil resistors. In addition to improved sound, the foil-based resistors allow for better resolution and quieter outputs in pre-amp and amplifier applications. A single resistor makes an audible difference.
About the Author
Yuval Hernik has been an Application Engineer for Precision Foil Resistors at Vishay Israel, www.vishay.com since 2004. He is a graduate in Electrical Engineering of Technion (Israel Institute of Technology) and is currently studying for a Biomedical Engineering degree at the Tel-Aviv University.