An Introduction to Gyrator Theory How inductors can be simulated using resistors, capacitors, and op amps. GYRATOR, believe it or not, is an inductor without any tums of wire. ‘Although the theory behind this interest- ing circuit has been established for some time, only within the past few years have synthesized inductors been used on a wide scale. Before we exam- ine the gyrator in detail, ls review ‘some basic properties of inductors. ‘A pure inductance is a circuit element ‘whose opposition tothe flow of alternat- ing current (inductive reactance) varies directly with frequency. At de or zero hhertz, the ideal inductor has zero ohms of resistance (a perfect conductor) and Zero ohms of reactance. Therefore, we ccan say that it also has zero ohms of lmpedance—the vector sum of re sistance and reactance. However, as we ‘move into the realm of ac, the reactance ‘of an inductor increases according to the formula X, = 2nfL; where X, is mea- sured in ohms; f (requency) in hertz; and L (inductance) in henries. Its re- sistance remains zero ohms. At infinite frequency, the inductor has infinite reactance, and will permit no acto flow. So far we have been talking about an ‘deal inductor. Actually, every inductor thas a certain amount of resistance and capacitance as well as inductance. As shown in Figs. 1A and 18, an iron-core inductor can be modeled as an induc- tance in series with a resistance, RY; and this combination is in parallel with a Capacitance and series resistance, A2. ‘An alr-core inductor (Figs. 2A and 28) behaves as an inductance and series re- sistance AT would. In both cases, L is the inductance of the coll, and A is the resistance of the wire which comprises the coil. The iron-core inductor contains two additional elements, R2 and C, which represent losses within the core. With dc, there are no core losses, and ‘consequently, our mode's C permits no ‘current to flow through F2. At higher and higher frequencies, core losses in- crease. Thus, in our model, increased current flows through 2 as the capaci- tor's reactance decreases. ‘Synthesizing an Inductor. By com- bining resistors. and a capacitor with a ss BY BRYANT, MORRISON gain stage, we can create acrcuit which appears to the “outside world” as a real inductor. To understand how, we wil analyze the inductor models (Figs. 18. and 28) in terms of "port admittance." A Portis a point through which enorgy can fenier or leave. In the case of an electr- cal circuit, it can consist ofa pair of ter- minals to which a circuit element is con- rnocted. The inductors and their models in Figs. 1 and 2 are ports, and when a voltage source is connected across them, an imput voltage (Vin) is applied ‘an an input current (I) Hows. “Admitance, measured in mhos, isthe reciprocal of impedance. in other words, admittance isthe ratio of current to vot ‘age. If an elomont’s admitance is zero hos, no current wil flow through it no matter how high the voltage is across it. Such an element isa perfect insulator oF ‘pen circuit. On the other hand, an ele- ment with infinite admittance will con- duct infinite current, even i alow voltage source is connected across it tis a per- fect conductor or a shor circuit. Combin- ing these two terms, port admitance is, the rato of the current flowing ito the ort (hx) t0 the votage across the port no Referring to Fig. 1B, we can see that resistors At and 2 set the limits of port impedance at both very high and very low frequencies. Atdc, the admittance of the inductor Lis infinite (a short circuit), and only At limits the current through it Capacitor C behaves as an open circuit with zero admittance, so R2is removed from the circuit. At an infinite frequency Lis an open circuit and At is removed from the circuit. However, C is a short Circuit and current through itis limited only by 2. Between these frequency extremes, L wil determine the port's admittance, because it is much larger than. The port admittance of the ai-core Cail at de is simply the reciprocal of re- sistance A, since L has infinite admit tance. At an infinite frequency, the port admittance is zero, because the induc- tance acts as an open circuit, and no in- put current can flow. Analyzing the Gyrator. Now lets apply these concepts to the gyratorcir- cuits (Figs. 1C and 2C). As in the equiv- alent circuits, At represents the ohmic resistance of the coll wire, and Cand R2 are core losses which increase in stop with the applied frequency. However, something new has been added—a gain stage. Any active device can be used, but here we choose an op amp for its simplicity, high gain, almost infinite input Impedance, and very low output imped- lance. The gyrator op amps are strapped for unity-gain, noninverting operation ‘So, within the frequency limits of the de- Vioe (assume infinite bandwidth), the voltage at the output is exactly the same 1a that at the noninverting input. we apply a dc voltage across the in- put terminals of Fig. 1C, capacitor C Fig. 1. Iron-core inductor (A) can be modeled as shown in (B) and simulated using the gyrator eireuit in (C). POPULAR ELECTRONICSPig. 2. An air: 2 coil (A) has an equivalent cireuit shown in(B). Op wmp gyrator (C) simulates the coil’s behavior. does not conduct, and the voltage at the noninverting input is zero. The output is also at ground potential, and because the op amp has very high output admit- tance (low output impedance), we can salely say that AY Is connected across, the port. So, hy will flow oniy through 1, This agrees with the behaviour of ‘the equivalent circuit of Fig, 18, The port admittances are maximized at dc, limit- fed only by the values of both A's (as sumedto be equal) {A infinite frequency, C is @ shod cit- cuit, and therefore the voltage at the op ‘amp's noninverting input (as well as that atthe output is equal to Vig. Since there is no voltage drop across A, itis effec tively removed from the circuit. The only admittance path is through A219 PROPERTIES OF GYRATORS: Advantages 1 munity to ambient magnate 4s no coupling ot eresstak between “indus. oe 2 Very smal size required for large val- asa inductance. ‘3 Inewpensiv, use Foadily avaiabiocom- ‘pone. 4. Accurtely procctable “saturation lov- es. 6, Parameters con be fixed by choice of resistors. Disadvantages 1. Autve davies genetetas noise (oan be held to lw lavas proper devicbs re selected). 2, More. 10 sificult and impractical to simulate, a the circle requ nigh power active 4, Simulated inductors are requeney init ‘ec by thelr acive devices’ usable band- ‘ids and slew tea (nota problem at ‘aut frequencies inmost cases), auuysert ‘ground, which is the same behavior we noted in the equivalent circuit. For frequencies between zero and in- finty, C and R2 act as a high-pass fiter, causing less and loss voltage drop ‘across Rt as frequency increases, and thus less port admittance until 2’ imit- ing effect comes into play. The reactive characterises of the capacitor have successfully been inverted or gyrated so that the port behaves as an inductor. The equivalent inductance in henries is expressed by the formula L = (F1) (R2) (©), with resistances in ohms ané capa- Gitance in farads, With the addition of two resistors, an air-core inductor can be simulated, Air- core coils have essentially no “core loss, and therefore have no parallel re- sistance in their equivalent circuits. Be- cause of this ie gyrator (Fig. 20) uses. the additional resistors to set the gain of the op amp. When the values are prop- rly selected, they provide enough gain to compensate for A2's losses at high frequencies. But the amount of gain myst be carefully chosen—otherwise tne circuit might escilatel I RS equals At and R4 equals R2, the circuit wll be stable and exhibit no parallel resistance. In practice, however, ile is gained over the circuit of Fig. 1C 2s long as the ratio 2/R1 is at loast 90 0 100, because the effects of paralel resistance are negligi- ‘ble in mast audio apelications commonly encountered. actical Design. In synthesizing a useful “inductor,” the same basic rules that govern the’ optimization of wound coils should be followed. For example, series resistance Ff should be kept as small as possible and parallel resistance. R2 as large as possible. This corre- sponds to a coll wound from the hesvi- est wire practicable on the feast lossy core available, For best performance, AY should be no lower than the op amps minimum recommended load re- sistanca, which falls between 100 and 2000 ohms for common op amp types. The largest acceptable value for F1 is desirable, so as not to load the op amp too much, thus preventing high distor- tion and heating effects. To simulate a high-quality toroidally wound coil, A2 should be at loast 100 times greater than Ft, but not so large as to become & major contibutor to the op amp’s input noise. As a rule of thumb, Keep AY around 1000 chms and 2 between 10 kilohms and t megohm, (Once the values of AY and A2 have been chosen, use the formula C = L/ (PINAL) ro find Ye sequicad capaci- tance in farads. At least 100 pF should bbe used to avoid the detuning influences of stay capacitances. It ig important to keep the op amp functioning within acceptable circuit and signal parameters. If for any reason it ‘begins to deviate fear the role ofa volt- age follower, the “inductor” won't work properly. Input signals mustlie within the ‘operating bandwidth of the device, and their amplitudes must not cause the out put stages to cp. In a gyrator, clipping in the gain stage is analogous to core satu- ‘ation, which can cause high distortion levels. However, this is not usually a prob- lem with gyrators. Because they will most often be operated from the same power supplies that other audio stages ‘use, they will not stam to clip until the ‘other amplifiers do. Unlike iron-core coils, whose saturation characteristics ‘re functions of core material, size, umber of turns, and applied current, the gyrator's saturation point is accu: rately predictable, and does not occur before the other active stages of the sys tom also saturate or clip. Using either of tho gyrators we have ‘examined will result in high-quality coils with inductances ranging from milinen rigs to hundreds or thousands of hen ries, Commonly available parts— includ ing relatively small capacitors—can be employed. Added benefits include high magnetic field ierunty and satu characteristics, and (paradoxically) ‘small amounts of required printed circuit board “real estate.” However, there is ‘one imitation, The gyrators we have de- scribed are single ended. That is, one side is grounded. To simulate “floating” Inductors, neither side of which is con- nected to ground, more complex citcuits Using two op amps can be designed. But ‘such gyrators are beyond the scope of this article. ° 39The right gyrator trims the fat off active filters Replacing inductors with gyrators creates almost perfect filters toy Thomas H. Lynch, aunkersamo co, Eecrnic Systems Dison Westke, Cat ©) Analog filters exhibiting nearly ideal performance can be built around a gyrator—if the sight contigueation of this active circuit is used. In effect, the gyrator makes 1 capacitance behave like an inductor, freeing the filter of the problems plaguing conventional inductors, like large size, low Q, winding capacitance, nonlinearity, and magnetic susceptibility Yet most designers look upon the gyrator as an ideal- iste circuit with @ “peculiar” behavior that puts it out of touch with practical applications. This attitude com- pletely ignores its power. Unlike other active-filter Circuits, the gyrator permits the designer to take advan- tage of the large body of dats and techniques already developed for passive LC filters. He can start with a passive prototype circuit and then replace each inductor with a gyrstor, substantially reducing filter size and weight for frequencies up to about 50 kilohertz Fortunately, to0, there is one gyrator realization that works superbly. Not all of them do—in the past, different versions have suffered from drawbacks like instability, poor contro} of loss, sensitivity to component matching, and even excessive complexity. But the preferred version is simple and stable and simulates high-quality inductor, permitting very high-performance filters to be realized. In addition, this gyrator, unlike other active-fiter circuits, preserves the most significant advantage of coupled LC networks— their inherently low sensitivity to changes in component values (see “The surength of LC filters,” p. 116) Understanding the gyrator Basically, the gyrator is x lossless two-port circuit (Fig. La) that inverts a load impedance. When used with 4 high-Q capacitor (Fig. 1b), it simulates the vitual characteristics of a high-Q inductor. The preferred reali- zation for the gyrator requires only two amplifiers and five impedances, as shown in Fig. 2 for both the general impedance representation (a) and the practical RC implementation (b). In the latter case, the circuit simu- fates an inductor having @ value of KC, where K is a constant determined by the resistors KE RRR: Electronics / uly 2, 1977 At first glance, this gyrator’s need for two amplifiers may seem a disadvantage. However, consider the major drawback of most single-amplificr resonators. They generally require an amplifier having a gain in excess of ‘Q% and those that do not usually are extremely sensitive 10 passive-element variations. On the other hand, the gyrator does not require a high-gain amplifier—in fact, stable Qs of better then 1,000 may be obtained with only 40 decibels of gain. Furthermore, unlike other active- filter citcuits, the gyrator is remarkably insensitive to any amplifier parameter, soit may be built with garden- variety devices, even quad chips, as long as they are wnity-gein-stable amplifiers ‘Additionally, with the gyrator, amplifier phase shift ‘enhances Q, rather than diminishing it as in other active- aaron yRaror 44 Meally. Curie’ LC fters simulated with gyators nave character istics approaching the Iaea. Yn ete, tne gyrator is @ sss two~ port ecult (a) that inverts a load impedance. With « capacitive load impedance, the cicut (b) simulates. high-auaity inductor. 18