Rectifying Without Rectifiers
July 1952 Radio-Electronics

July 1952 Radio-Electronics

July 1952 Radio-Electronics Cover - RF Cafe[Table of Contents]

Wax nostalgic about and learn from the history of early electronics. See articles from Radio-Electronics, published 1930-1988. All copyrights hereby acknowledged.

It took a couple times reading through this "Rectifying Without Rectifies" article to get the gist of what author H.B. Conant was talking about. He begins by pointing out the negative aspects of using nonlinear metallic rectifiers in a bridge circuit for an electric meter, then goes on to describe an improved "translator" circuit that uses - wait for it - nonlinear metallic rectifiers (or nonlinear resistors made of Thyrite material). If my interpretation is correct, basically the new and improved circuit incorporates a bias voltage that forces the nonlinear element (be it a metallic rectifier or a nonlinear resistance) to operate in a region which passes a higher current level to the meter movement when low values are being measured. One of the drawbacks mentioned of a traditional (at the time) bridge circuit was the need for separate calibration / marking of the meter's scale on the front panel, but then he says the translator meter also does not have uniform scales on all voltage ranges. I are a bit confused. For now, I'll stick with my traditional meter.

Rectifying Without Rectifiers

Rectifying Without Rectifiers

"Translator" assembly.

New technique frees a.c. voltage measurements from the limitations of metallic rectifiers.

By H. B. Conant*

The most widely used instrument for measuring low-frequency a.c. voltages is the type that combines a small metallic bridge rectifier and a d.c. meter movement. (See Fig. 1.) Its low cost, simplicity, compactness and ruggedness have made it popular for service instruments. It is used almost universally in broadcasting and sound recording to measure volume levels.

In spite of its advantages and wide range of applications it has two serious drawbacks: (1) The current through a metallic rectifier does not vary in direct proportion to the applied voltage. In other words, the resistance of the rectifier does not follow Ohm's law. The relationship is logarithmic:

Standard a.c. voltmeter circuit - RF Cafe

Fig. 1 - Standard a.c. voltmeter circuit.

"Translator" type a.c. meter - RF Cafe

Fig. 2 - "Translator" type a.c. meter.

Current paths in the circuit - RF Cafe

Fig. 3 - Current paths in the circuit.

Two-bridge translator - RF Cafe

Fig. 4 - A two-bridge translator.

Logarithmic scale calibration - RF Cafe

Fig. 5 - Logarithmic scale calibration.

R ∝ I/log E.

At very low voltages the resistance of a metallic rectifier is so high that not enough current reaches the meter to give a readable indication. (2) At high audio frequencies the capacitance between the rectifier surfaces allows current to flow in both directions. This effect increases with frequency and makes this type of instrument useless for measuring r.f. voltages.

A third disadvantage, found in multi-range instruments, is also due to the nonlinear resistance of the rectifier.

For maximum accuracy, a separate scale calibration must be made for each range. This is impractical and confusing on service-type instruments, and compromise calibrations are generally used, with an accuracy of about 5%.

The Translator

A new circuit called a Translator overcomes these basic limitations. It uses metallic rectifier elements (or non-linear resistance materials like Thyrite) in the bridge arrangement shown in Fig. 2.

R1, R2, R3, and R4 are the nonlinear elements. R5 is a current-limiting resistor of ordinary permanent-resistance type. Battery E supplies a bias current, and M is a d.c. milliammeter. R6 is a balancing potentiometer for setting the meter pointer to zero. The four bridge arms are closely matched.

The bias current flowing through R5 and the arms of the bridge develops a series of voltage drops in the circuit. With respect to point D, points A and B would have the same negative voltage, and point C would have twice the voltage of A and B. R6 is adjusted to the same voltage as C so that no current flows through the meter. If a.c. is applied across A and B the following action takes place: (See Fig. 3) During the first half-cycle the voltage at point A is increasing and the voltage at point B is decreasing. The resistance of non-linear arms R1 and R2 decreases, while arms R3 and R4 increase. When point A reaches the same voltage as point D, the voltage at B will equal the voltage at C. At this instant no current can flow through R3 and R4, and these two arms are effectively eliminated from the circuit. The current path under these conditions is shown by the arrows in Fig. 3. The increased current through R1, R2, R5, and R6 changes the voltage at C and unbalances the meter.

As the voltage at point A continues to rise, potential differences again develop across R3 and R4, and their resistance decreases. Only part of the additional current drawn by R3 and R4 flows through the meter circuit. The meter reading increases logarithmically. (This characteristic is highly desirable for many applications. Ordinarily, logarithmic-scale instruments require specially-shaped magnets or complex circuitry.)

On the negative half-cycle the functions of the opposite arms are reversed. The main current path is now through R3, R4, R5, and R6, but still in the original direction through the meter.

Higher bias current extends the low-voltage sensitivity of the circuit. (Increasing the bias current solely by reducing R5 will decrease the meter response by the greater shunting effect.)

Reducing the bias current increases the maximum voltage that can be read on any range, at the expense of low-voltage sensitivity.

Circuit Variations

Both copper oxide and selenium rectifiers were tried in developing the translator circuit. With copper-oxide units, the low-voltage sensitivity was excellent. Slight instability in the selenium rectifiers caused some drift in readings.

The question of whether or not rectifying action was involved in translator operation was settled by constructing a bridge of nonrectifying Thyrite resistors (G.E. No. 8396839G1). Aside from requiring a higher bias voltage, the Thyrite bridge performed very well, although it was not as sensitive to very low voltages as the copper-oxide bridge.

A translator can be made by connecting two full-wave bridge rectifiers as shown in Fig. 4. An assembled unit is shown on the opposite page.

The frequency response of the translator appears to be perfectly flat up to 20 kc. While conventional rectifier-type instruments indicate the average value of the a.c. voltage, there is some indication that the translator shows r.m.s. voltages.

Like rectifier-type instruments, the translator does not have uniform scales on all voltage ranges. The scale uniformity and accuracy are improved by using a different value for R5 as well as a different multiplier on each range.

Logarithmic scale calibration makes it possible to read low and high voltages accurately on a single range. (See Fig. 5.) With the translator circuit an instrument covering 0−10,000 volts in only three ranges can be built without difficulty.

* President, Conant Laboratories.

 

 

Posted July 11, 2022