The most outstanding differences between acoustical noise and electrical noise in components, or any electronics system for that matter, are easy to state. Acoustical noise is any unwanted sound or vibration that is audible to the human ear and/or tangible via touch. Electrical noise is, by literal definition, not particularly audible or discernible by tactile interaction with the component or equipment. In most cases, perception of electrical noise requires test instruments, such as an oscilloscope or signal analyzer.

  1. Acoustical Noise Variations
  2. A Cacophony Of Electrical Noise
  3. Minimizing Mechanical Noise
  4. Stifling Electrical Noise
  5. Summary
  6. References

 

Acoustical Noise Variations

Sources of acoustical noise are many, but it isn’t always unacceptable. For example, some noise from a motor driving a light conveyor belt on a noisy factory floor may be acceptable if the motor is within spec for the task it performs and the noise level it creates does not exceed safety standards for people in the vicinity.

However, a motor in a high-end piece of audio equipment needs to be as quiet as possible during operation. Silent would be ideal. This is particularly critical if the motor is near the signal chain like a motor diving a compact disc or turntable belt, or even performing a rudimentary function like opening a control-panel door (Fig. 1). In this case, any acoustical noise is unacceptable.


1. In high-end audio apps, silence is golden, but hard to come by, particularly where mechanical movement comes close to audio-signal paths. In this basic example, the designer minimizes acoustical noise from the turntable by using two belts.

Poorly made transformers are another source of acoustical noise. The vibration of core laminations that aren’t quite completely bound together becomes audible during equipment operation (Fig. 2). Coils in switching power supplies can also crank out a whining, high-pitch oscillation if they heat up a bit.


2. An extreme case, this power transformer exhibits a poorly laminated core that will most definitely raise a mild rumpus in the power section.

Working within the audio-frequency range, from about 20 Hz to 20 kHz, certain types of non-polarized capacitors can generate acoustical noise. Sources include the use of inexpensive capacitors, poor board layout, or both, or underrated capacitors whose capacitance and/or voltage ratings need recalculating. Sometimes a higher grade or special low-noise capacitor can solve the noise problem and compensate for limited printed-circuit board (PCB) design. Obviously, proper cap ratings also will go a long way in solving noise issues.

Regarding acoustical noise, a good rule of thumb in any design prototyping is to question any source of mechanical noise regardless of the component. Even if noise levels are acceptable and within spec and regulation, the noise can indicate a faulty component and imminent system failure.

A Cacophony Of Electrical Noise

Electrical noise is more sophisticated in both nature and detection than acoustical noise. A physiological analogy would be the pain one experiences from stubbing one’s toe (acoustical) compared to an unexplainable sensation in one’s foot that requires an x-ray or CT scan to diagnose (electrical).

Electrical noise is not audible to the human ear. Even in high-end audio equipment, actual electrical noise is not audible, but its effect on the analog audio output is audible as distortion at all or just certain frequencies. The noise itself requires specialized test equipment to detect, analyze, and define.

By standard definition, electrical noise is any unwanted variation in an electrical signal. To greater or lesser degrees, all circuits have some inherent electrical noise. Once again, depending on the circuit or system, electrical noise may be acceptable within certain limits.

Occurring on both power and signal lines, there is no shortage of electrical-noise types. Capacitive or electrostatic coupling is one source. This involves the inducement of charges from one conductor in a circuit to another via the capacitance in their insulators—for example, two coated wires running next to each other with one carrying current. The capacitance of their coatings (insulators) stores part of the charge and transfers it to the other wire not carrying the same current. Hence, the unwanted charge is electrical noise.

Another type of electrical noise, inductive coupling, is magnetic-coupled noise. Any conductor carrying current generates a magnetic field that can induce current in nearby conductors and circuits.

Other types of electrical noise include thermal noise, shot noise, and flicker noise. Thermal noise emanates from the friction between charge carriers in a conductor. This friction generates varying levels of heat. Shot noise includes any random fluctuations in current flow. Flicker noise, also known as 1/f noise, is the level roll-off or unwanted attenuation of a signal at higher frequencies.

In an ideal world, eliminating acoustical and electrical noise should be simple. Just create the ideal design with textbook-perfect layouts using the finest components available, right? Simple, but impossible. Ever shrinking product footprints impose size restrictions that require unique thinking and innovative solutions when it comes to circuit layouts, and using the finest components available is far from cost effective, let alone necessary.

Minimizing Mechanical Noise

Even when using premium components in optimal layouts, some designs are just inherently prone to generating acoustical noise. Power supplies are notorious suspects here.

Switched-mode power supplies (SMPS) are usually pretty quiet when operating at constant switching frequencies above 20 kHz. When working with unique loads, though, SMPS operating in burst-switching mode can start to speak audibly. During burst operation, the usual sources of acoustical noise are capacitors and transformers.

Capacitors used in power supplies undergo a good deal of stress from applied voltages as well as voltage surges and sags. Aging and heat also play a part in this assault upon capacitors.

For instance, ceramic caps tend to expand when voltage is applied and contract back to their original size when voltage ebbs. Naturally, these expansions and contractions will vary in depth depending on both the quality of the component and the voltage levels. Either way, with continual size variations over time, unwanted situations may arise. The capacitor may loosen its solder joints and physically move about, causing intermittent open circuits and generating a good deal of noise as a result. This noise resonates on the PCB, making it audible.

There are a few solutions for this condition. One is to replace the potentially noisy single caps with two that are half the capacitance value and wired in parallel. One capacitor should be placed above the board and the other directly under it on the opposite side of the PCB (Fig. 3). When they begin to expand and contract, the motion of the two capacitors will oppose each other and thereby keep the PCB between them in a stable physical state.


3. On the left, the 0.01-mF ceramic capacitor expands when voltage is applied and exerts some pressure on the circuit board, which will cause noise problems over time. The solution on the right places two 0.005-mF capacitors in parallel with one seated on top of the board and one directly below it. The expanding actions of both components counteract each other, keeping the board stable and quiet.

Another solution is possible if space limitations don’t permit the aforementioned piggybacking solution and the design budget is solvent. It involves specifying a precision, low-noise component such as those available from component makers like Murata and Kemet.

Movable parts, coils, bobbins, and other components are the usual source of audible noise in transformers. Also, electromagnetic fields that appear when current flows through the transformer create push or pull attractions between the coils. Attractions and repulsions can cause coils, cores, and insulations to vibrate.

There are several solutions for transformer noise. Mechanical motion can be damped with tape, lacquer, or other material. Or, choose a transformer that reduces the flux swing in burst-mode switching and/or reduces the current peaks in burst mode, which in turn reduces flux swing.

Finally, never rule out a noisy fan unless you can really live with the obnoxious whirring sound of an air churner that barely meets spec. There aren’t many choices or strategies here other than to get the quietest fan that meets both cooling and budget constraints and, of course, fits the chassis.

Stifling Electrical Noise

Notably, acoustical noise often can generate electrical noise and vice versa. A capacitor getting loose in its solder joint that wobbles around can cause unwanted oscillations (electrical noise) outside the audible range. A poorly configured audio circuit can induce excess heat in the power section that causes a transformer to vibrate (acoustical noise). The possibilities are as numerous as they are annoying.

Some forms of electrical noise include switching noise, crosstalk, jitter, oscillation, hum, and burst noise. For the most part, a good initial design will minimize these disturbances at the very least. However, even the best designs can get noisy in the field depending on what equipment is nearby, temperature and weather conditions, and stray magnetic fields as well as radio frequency interference.

Using test equipment, electrical noise is most often traceable to faulty components, open and short circuits, and mechanical noise sources. Rectifying these sources by replacement and/or repair is fairly easy. For most interference problems, there are more than enough filters and materials available to solve them.

Some esoteric noise can be a challenge. Take burst noise, also called popcorn noise (Fig. 4). Emanating from integrated circuits (ICs), burst noise comes from random transitions between two current or voltage levels. A voltage or current shift sounds like a “pop.” Since shifts happen randomly and at irregular intervals, they sound like corn popping.


4. Burst noise, or more affectionately dubbed popcorn noise, is usually the result of flaws in the semiconductor device. On an oscilloscope display (current versus time), three bias levels for a single chip generate distinct levels of burst noise.

Burst noise is the result of variations in the semiconductor manufacturing process. One solution is to screen certain lots of a chip intended for a design to find the lot with the lowest incidences of burst noise. Luckily, due to advanced manufacturing and screening techniques, burst noise is not as common as other forms of electrical noise.

Summary

Noise, electrical or acoustical, in any electronic design poses concerns. First, it’s unwanted and annoying. Second, noise affects performance and overall efficiency. And third, acoustical noise could be the alarm bell warning of impending failures. Electrical noise will certainly induce a failure. In summary: noise bad, quiet good.

References

  1. Vibration control and noise control strategies
  2. On the influence of core laminations upon power transformer noise
  3. Capacitor Noise Countermeasures
  4. Capacitor Noise
  5. Two-phase electric power
  6. A Burst Noise Cancellation Scheme for Single Carrier Block Transmission with Cyclic Prefix