Audiotronics: Ear-Pleasing Electronics for DIY Construction

You can probably safely assume that most professional and serious hobby electronics enthusiasts who have now reached middle age took their first tentative steps on this path by building simple amplifiers and radio receivers (or copies), gradually mastering the theory through trial and error. And they know that feeling of intense satisfaction and pride when that very first project more or less did what was expected — in our case, that was the “jam jar receiver” published around 1960 in Radio Blan [1, 2]. In this series of articles, we want to try to give the beginning electronics hobbyist that sense of satisfaction again — more so (in our humble opinion) than is possible with ready-made modules from the Far East.

Theory 1

The Basics of Sound Quality

Audio signals are, of course, influenced by the storage method (record, cassette, CD), but (especially) by any form of processing. When an audio signal passes through an electronic circuit, the output of that circuit contains not only the desired (modified) signal, but also a range of unwanted and undesirable side effects. Typically, these are noise, hum, crackles, or distortion which, in the worst case, are not only audible but also extremely annoying.

In all circuits intended for hi-fi applications, these potential negative influences on sound quality must be minimized as much as possible by appropriate circuit design measures. The principle here is: any audible influence on the sound is already too much. This principle also applies to the circuits described in this article series. It should be noted that in circuits with very high gain — for example, microphone amplifiers — some noise is inevitable.

To combat noise, hum, or distortion as effectively as possible, let’s first take a closer look at the origin, meaning, and effect of these phenomena.

Noise

Noise as Interference Voltage

In technical documentation, noise is always specified as signal-to-noise ratio. However, this does not refer only to the noise itself, but to everything that appears at the output as interference regardless of the input signal — even when there is no input signal present. For example, hum is also included when determining the signal-to-noise ratio, even though hum usually plays only a secondary role compared to “real” noise.

The signal-to-noise ratio indicates how many times weaker the interference voltages at the output are than the desired signal — so how many times weaker the noise is than the music. This factor is on the order of a few thousand times, but it is usually not specified as such, but in the pseudo-unit decibel.

What Is Noise?

Mathematically, noise is a mixture of all imaginable frequencies. The noise signal has a purely random course in which no pattern or repetition can be recognized. But not all noise is the same.

The Color of Noise

With noise, we distinguish between white and pink noise. With white noise, the noise contribution in each equally sized frequency range is the same. If, for example, we were to filter out the range from 10 Hz to 20 Hz from the entire noise spectrum, the noise signal here would have the same amplitude as in a range from 110 Hz to 120 Hz or from 19990 Hz to 20000 Hz.

With pink noise, we can divide the spectrum in a comparable way, except we use a logarithmic instead of a linear frequency scale — corresponding to the sensitivity of our ears. Then, in a frequency range from 10 Hz to 20 Hz, we find the same noise share as in the range from 100 Hz to 200 Hz or from 10000 Hz to 20000 Hz. As a result, pink noise sounds much “duller” and broader than white noise.

The noise produced by electronic circuits is normally white noise. This is almost exclusively audible in the high-frequency range, since the noise share in the range from 10 kHz to 20 kHz is as large as in the entire remaining audible frequency range.

How Does Noise Arise?

When a conductor is traversed by a current, the resulting movement of the electrons is never perfectly even. Depending on the material of the conductor, minimal, seemingly random irregularities occur, which are especially noticeable with semiconductors (but also with resistors).

To ensure a circuit produces as little noise as possible, we use only low-noise semiconductors and metal film resistors. In addition, the circuit around the semiconductors must not have too high an impedance (we’ll come back to this in a later installment).

Another cause of annoying noise must be sought in the storage of audio signals on sound carriers — for example, the formerly common cassettes or tape reels. Some minuscule magnetic particles on such a carrier occasionally stray out of line when magnetized, and apart from that, the magnetizable layer is never perfectly uniform, even on the very best tape.

With digital storage media (CD, DVD, hard drives, and the like), the originally purely analog audio signal acquires a slightly “stepped” shape due to digitization, and this is the cause of so-called quantization noise. In the case of WAV files with only 8-bit resolution, this is clearly audible. However, a resolution of 16 bits is usual, and in that case quantization noise is no longer significant, because the “steps” in the signal are then 256 times smaller. When a PC sound card is noisy, it is usually due to another cause, such as an incorrect signal level at the microphone input.

Practice 1

The Power Supply

Before we move on to various DIY projects in the next installments, we of course first need a suitable power supply. Since the circuits are mostly built with op-amps, a symmetrical power supply is required. A universally usable mains power supply with a stabilized output voltage of ±15 V is suitable, which can deliver a current of 30 to 300 mA depending on the transformer used.

Quality Requirements

When dealing with circuits that must process audio signals in hi-fi quality, we must impose a very important requirement on the power supply from the outset — namely, that it provides an output voltage as free of hum as possible. The regulator ICs used here, type 7815 and 7915 (or possibly the corresponding M or L versions, such as the 78M15), are excellent in this respect. The remaining ripple on the output voltage is at most a few millivolts. As ripple on a music signal, that would be unacceptably high, but here it is an interference signal superimposed on the supply voltage and does not end up at the output as such.

When you consider that an op-amp suppresses power supply disturbances very well (the SVRR or supply voltage rejection ratio is about 70 dB), you see that not much of that ripple on the supply voltage remains. The voltage regulators from the 78- and 79-family also offer other advantages. They have built-in current limiting and thermal protection, so they can be considered short-circuit proof.

Universal ±15-V Power Supply

The power supply described here delivers a stabilized output voltage of ±15 V and is therefore suitable for all the circuits described in this article series (with the exception of a few power amplifiers, which naturally need more power). Equipped with a 4.5-W transformer, this power supply delivers up to 140 mA. That’s more than enough for “extensive” circuits with several quad op-amps and LEDs. When the power supply only needs to supply a single, small circuit, a lighter transformer can be used. In any case, the PCB is designed so that transformers of different sizes can be installed.

The Circuit

The transformer in Figure 1 is supplied with mains voltage via fuse F1. This transformer delivers a secondary voltage of about 2 × 15 V (slightly more under light load). The rectifier, consisting of D1...D4, together with filter capacitors C5 and C6, produces a DC voltage of about 2 × 20.5 V. With respect to ground, C5 has a positive voltage of +20.5 V and C6 a negative voltage of −20.5 V. Capacitors C1...C4 help suppress any high-frequency interference.

If you are wondering how a voltage of more than 20 V can arise from 15 V, you’ll find the answer in Figure 2.

IC1 stabilizes the positive voltage across C6 at +15 V. C7 and C9 help to minimize the remaining ripple. Similarly, IC2 stabilizes the negative voltage at −15 V, with C8 and C10 again providing extra ripple suppression. R1 limits the current through an (optional) indicator LED.

Figure 3 shows the component layout on the PCB designed for this circuit. The layout can be downloaded from the Elektor labs project page.

Power Consumption

The PCB for this power supply is designed to fit almost all common PCB transformers from 1 to 10 W. For simple circuits with no more than five op-amps, a 1-W transformer is sufficient; for more extensive circuits, a 3-W version is needed. Circuits with LEDs may under some conditions consume even more power — to be on the safe side, allow for about 0.4 W per LED.

Questions or Comments?

Do you have any questions or comments regarding this article? Send an email to the Elektor editorial team at editor@elektor.com.

Editor's note: This article (250845-01) appears in Elektor March/April 2026. The article series “Audiotronics” is based on the book Audio-Elektronica by Robert Sontheimer, which was published in Dutch translation by Elektor in 2006.