Diode Noise Generator


Semiconductor diodes may provide a means for generating noise useful for a variety of applications including cryptography, signal jamming, sound masking, and instrument calibration. Diodes particularly suited for some noise-generating applications have to be carefully selected for spectral uniformity within specific frequency bands. Such diodes are quite expensive, but in many less-demanding applications, an inexpensive diode costing as little as a few cents may provide excellent performance.

In the circuit below, a selected 1N5239B 9.1V "Zener" type is used as a voltage noise source. These diodes are not characterized for noise during manufacture, so the amount of noise a particular diode produces may vary depending upon the manufacturer and batch. In fact, even diodes from the same batch may exhibit significant variations. For example, this circuit worked extremely well with type 1N5239B manufactured by Motorola Semiconductor in the 1980s, but not with presently-available units manufactured by Fairchild Semiconductor.

Although the subject diode is commonly refered to as a "Zener" (named after Clarence Zener who developed it in 1934), its actual mechanism of operation is better characterized by the "avalanche" effect, which dominates for diodes in which the breakdown voltage is greater than 5.6V. Avalanche diodes with higher voltage ratings generally produce proportionally more noise voltage. A common silicon diode such as the 1N4007, operated in the reverse-bias avalanche region, may produce hundreds of volts of noise. However, such an implementation would require an impractically large voltage source (greater than 1000VDC). Therefore, the circuit below uses a 9.1V diode, since its noise voltage is, on the average, more likely to be greater than that of lower-voltage Zener types, but not so high as to require an unusually-high voltage bias. Furthermore, avalanche diodes with voltages close to the 5.6V Zener limit have a lower temperature coefficient of noise, and will therefore behave more consistently over a large temperature range. This doesn't eliminate the possibility that a batch of noisy lower-voltage Zeners may be found; many published circuits have featured Zeners in the 2 to 4V range. However, the chances of finding suitable 9.1V avalanche types are much better.

To further increase the usefulness of this circuit by allowing it to work with a single low-voltage source (such as a 9V battery), an inductive boost converter (U2 and associated components) followed by 12V regulator U1 has been included. This increases the battery voltage to a level high enough to bias the Zener. With the values shown, this circuit will work equally well for a supply range from 7 to 11V, and consume less than 20mA of current. For 5V operation, simply replace L1 (4.7mH) with a 470uH inductor.

While viewing the circuit output with an oscilloscope, set the "DIODE BIAS" potentiometer so that the waveform voltage distribution appears even, i.e., that the density of the noise signal looks symmetrical around its center, top-to-bottom. Set the "AMPLITUDE" potentiometer for the desired output amplitude. Since the TS922IN dual operational amplifier includes ground and Vdd in its output range, the noise amplitude may be very near its Vdd supply of 12V. However, to prevent saturation at noise voltage peaks, a practical limit of 10V is suggested.

Since this circuit employs moderately high AC gain (up to 200), construction should be on a ground-plane PC board that is well shielded from AC mains or other interference sources. The 10uF capacitors should be either low-ESR aluminum, or better still, solid tantalum types. The 0.1uF capacitors should be X7R dielectric ceramic types, and the 1nF capacitor in the U2 oscillator an NPO ceramic. Resistors may be standard 5% carbon film types. The inductor is not critical; an inexpensive Bourns type RLB9012-472KL (4.7mH) or RLB9012-471KL (470uH) is well suited. The potentiometers should be cermet trimmer types such as Bourns 3299W-1-502LF (5K) and 3299W-1-203LF (20K).


November 4, 2012

Text and image ©2012 by Arthur Harrison


Back to the Circuit Library Index
Back to the Opening Page of Art's Theremin Page