Understanding Digital Voltmeters (John Frye)

(source: Electronics World, Dec. 1971)

By John Frye

These instruments are coming on strong and technicians should make an effort to understand just how they work.

WHEN Barney entered the service department this frosty morning, he discovered Mac, his employer, sitting tailor-fashion on the workbench reading a thick, heavy book with gold-colored binding. The youth walked over and raised the volume from Mac's lap so he could read the title.

"NonLinear Systems, Inc.: Instrumentation, Systems, MOS /LSI, Pressure," he read aloud. "That doesn't tell me much. What were you reading so intently ?" I'm boning up on digital voltmeters," Mac answered.

"I've wondered how these things work for a long time, but I never ran across a really good discussion in language I could understand until an engineering friend let me borrow this book put out by Non-Linear Systems. It is beautifully written and illustrated; and I've not been so intrigued by a book since, as a teenager, I got hold of a contraband copy of Lady Chatterly's Lover." "Your age is showing," Barney warned. "That old hair curler reads like a Sunday School manual compared to many best-sellers sold openly today. But why all this interest in digital voltmeters? Are you planning to buy one for the shop? I always thought digital voltmeters were like the idiot lamps that replaced meters on the dashboards of cars: they are for the benefit of women and other sub-intellects not smart enough to read a meter." "I heard that, Barney Gallagher!" Matilda, the office girl, called from the other room, "and you're going to get it." Mac grinned at this exchange as he said, "That's not quite the case. The fact that d.v.m.'s are easier to read, especially by untrained personnel, is one of their advantages; but there are many others. I've not made up my mind as to whether or not we need one of these instruments in our work now, but I know sooner or later we're going to encounter them in our industrial service. You find them now in electronics labs, test labs, medical research, chemical processing, power distribution, the petroleum industry, guided missile and aircraft testing, nuclear research and testing, electronics and electrical production lines, quality control departments, and dozens of other places. I want to know how d.v.m's work, what different types there are, what advantages they have over pointer-type instruments, what are their limitations, and if any special precautions are necessary for their use." "Okay," Barney said, perching himself on a stool and leaning back against the wall, "suppose you brief me on what you've learned so far." "I was hoping you'd ask," Mac admitted, closing the book. "I found out long ago that trying to explain something I think I know is the best way in the world to reveal fuzzy areas in my assumed knowledge. A digital voltmeter, or d.v.m., is an instrument that automatically displays a measured voltage in the decimal number system. The first one was originated by Non-Linear Systems in 1952, a 4-digit instrument still in use. NLS and other companies such as Cubic, Electro Instruments, Fairchild, and Hewlett-Packard have made many improvements on the original model, and the d.v.m. has really caught on in science and technology for a number of reasons: (1) operation, including range selection, polarity indication, and decimal placement, is purely automatic, (2) parallax does not degrade the accuracy of the reading which can be seen from up to 30 feet away, (3) accuracy as high as 0.003% and resolution up to 0.00083% can be obtained, (4) thousands of readings can be taken per second for recording purposes, (5) there is extremely small current drain from the source being measured (from 0.0001 to 1 microampere at 10 volts), and (6) the instrument is relatively rugged for this kind of precision." "Sounds good, but how does it really work ?" Barney asked.

"This book compares the action to that of weighing on a balance scale. There the unknown weight is placed in one pan and known units of weight are placed in the other pan until an exact balance is obtained. You then know the weight of the object is equal to the sum of the weights required to balance it. You're too young to remember, but an early v.t.v.m., called a 'slide-back v.t.v.m.,' applied this principle. It consisted of a simple electrical bridge in which the unknown voltage was applied to one leg and a balancing voltage selected by a potentiometer across an internal voltage source was applied to the other. A 'magic-eye' tuning indicator tube was used to indicate a bridge null or balance. The dial setting of the potentiometer required to balance the bridge thus became a function of the voltage appearing across the test leads; the instrument was calibrated by graphing bridge-balancing dial settings vs known applied voltages.

"Suppose we substitute decade banks of resistors for the potentiometer. One decade furnishes 9 voltages in 1-volt steps; another, 9 voltages in 0.1-volt steps; a third, 9 voltages in 0.01-volt steps. Voltage being measured goes to an electronic balance director whose output commands a logic circuit to increase or decrease the internally applied voltage to achieve a balance. Automatic switching selects and applies the internal voltage, and a numeric display follows each step of each decade.

"Let's say we place the test leads across a potential of 7.84 volts. This potential is applied to one side of the electronic balance detector. The internal applied voltage is 0; so a command goes to the logic circuit to increase the internal applied voltage. One-volt increments are added until 8 volts is reached and the balance circuit is unbalanced in the opposite direction. The voltage backs up to 7 volts and this number is displayed while the 0.1-volt bank is activated. It increases the voltage in 0.1-volt steps until 7.9 volts is reached and again tilts the balance, causing the voltage to decrease to 7.8 volts and to activate the 0.01 bank. The internal applied voltage increases in 0.01-volt increments until 7.84 volts is reached. This balances the bridge and stops the selector circuits. The display now reads 7.84 volts.

All this takes place in a fraction of a second. "Admittedly this is a highly simplified and not quite accurate description of the action of a comparison-type d.v.m. but it illustrates the principle. The main difference between comparison-type d.v.m.'s is in the manner and sequence in which feedback voltage increments are produced. The two major classifications are the so-called 'all-electronic' and the electromechanical types. I say 'so-called' because no wide range, high-accuracy, automatic digital voltmeter of the 'all-electronic' type is completely devoid of electromechanical devices; all use electromechanical relays in the range or polarity circuits.

Anyway, in the all-electronic type electronic switches are used to develop the feedback voltage, while this is done with relays, stepping switches, and servomechanisms in the electromechanical type.

"The ramp-type is an interesting version of an all-electronic comparison d.v.m. The input voltage is fed to an error detector where it is compared with a linearly increasing voltage.

When this ramp, or sawtooth voltage is 0--as it is when the test leads are applied-an electronic counter with a built-in time base is turned on; and when the ramp voltage and the unknown voltage are equal, it turned off.

Since the ramp voltage increases at a uniform rate, the uniformly spaced counts that take place between the time this voltage starts to increase and the time the ramp voltage equals the measured voltage is an indication of the amplitude of the latter. If the ramp voltage increases at a rate of 1 volt per second and the timer, counting milliseconds, reads 0.506 second, then measured voltage is 0.506 volt." "Neat!" Barney exclaimed. "But you keep referring to the `comparison type' d.v.m.'s. Are there any other types? "Yes, there is also the conversion type in which we have analog /digital conversion. In this type, electron circuits convert the measured d.c. voltage to a train of pulses whose repetition rate is proportional to that voltage.

Then the number of pulses produced per unit time is counted by a gated electronic counter. Suppose our pulse generator frequency varies continuously and linearly from 0 to 10,000 Hz as applied voltage varies from 0 to 10 volts. Suppose the gate is opened for one second while the measured voltage of 7.84 volts is producing a pulse rate of 7840 Hz. The counter will register 7840 counts, and the decimal will be automatically inserted to indicate 7.84 volts." "Sounds tricky to me," Barney observed. "I'd think the comparison type would be more accurate." "You think right. One problem with the conversion type is erroneous readings produced by a.c. noise whose 'period' is not some exact multiple of the time based employed in the integrating-type d.v.m." "You said that first d.v.m. was a 4digit type. Does the number of digits have anything to do with accuracy of the instrument ?" "Yes. Accuracy of a d.v.m. is usually given as a percent of reading and a percent of full scale. Typical accuracies are ±0.1% of reading plus 0.1% of full scale for a 3-digit instrument, 0.01% of reading plus 0.01% of full scale for a 4-digit instrument, and 0.003% of reading plus 0.001% of full scale for a 5-digit instrument." "What do you mean by `resolution' ?" "That's the total range an input voltage can be varied without initiating a change in reading. The better the resolution, the smaller this range is." "I see some d.v.m. manufacturers advertise a certain percentage of `over range' for their instruments, say 20%. Does this mean you can apply 20% more voltage to the instrument than its maximum range without damage ?" "No. This refers to a usable extension of a basic range which allows high-resolution and high-sensitivity measurements at or near normal range transfer points. We have a little of this same problem in an ordinary voltmeter if it has too few ranges. Suppose we want to measure 1.2 volts on a meter that has only a 1-volt and a 100-volt range. The voltage can't be read on the 1-volt range; yet the pointer barely moves on the 100-volt range. In a d.v.m. with over-range, the lower voltage range is extended to make possible reading voltages somewhat above the normal range transfer point on the lower range where the percentage of full scale accuracy is more favorable." "Don't d.v.m.'s have any disadvantages?" "Sure. One is price. They cost considerably more than ordinary voltmeters. Another is complexity and the need for high quality standards and other equipment for recalibration.

However, a d.v.m. is well worth the money if you need to do high-speed measuring, if you want a meter untrained personnel can read accurately, if you need 0.5% or better accuracy, if you require high input impedance, if you need a precision meter that will be unaffected by rugged operating conditions, or if you want an instrument that can operate in automatic testing or data logging systems.

"But I've barely scratched the surface of what is in this book. It goes into a.c. and resistance ranging of d.v.m.'s.

Every subject I have mentioned is taken up in detail. Several more exotic types of d.v.m's, such as the stroboscopic, the staircase ramp, and the dual slope integrating-type are described.

You can read the book when I am through with it. Since it is prepared by the people who originated the digital voltmeter, I consider you'll be getting your information as nearly from the horse's mouth as possible!"