Linearly Compensated Slope Multiplier

A large part of the electronic multiplier circuits used today is based in principle on the function of a differential amplifier with two transistors, whose amplification is theoretically proportional to the sums of the collector currents (see Tietze and Schenk, Halbleiterschaltungstechnik/Semiconductor Circuit Technology, Springer Verlag Publishers, 6th edition, FIG. 4.41). The most obvious disadvantage of this circuit is the non-linearity of the transmission characteristic (FIG. 4.44 in). This error source can be eliminated by using a differential amplifier introduced into the negative feedback of an amplifier to compensate the transmission characteristic of a second, similar differential amplifier whereby the input of the second differential amplifier is connected in parallel with the input of the first differential amplifier and whereby the signal applied to the collectors of the second differential amplifier is prepared by means of an output amplifier. Such solution was published for example in Linear Databook 1982 by National Semiconductor on Page 3 170, bottom. In a circuit executed in accordance with page 3-170 of the National Semiconductor Linear Databook 1982 referred to above, the impact of non-linearity can indeed be kept very small, but only if either control currents for both differential amplifiers are maintained very small or if the control currents are equally great ("control currents" refers here to the currents which discharge through the common emitter connections and which, as a result, affect amplification). With increasing deviation, non-linearity multiplies with increasing values for the control currents.

The cause of this disturbing effect is the inner resistance in the emitter, which occurs in real transistors, the socalled emitter bulk resistance (referred to as such in the English-language literature). This resistance (R-Eb) acts in principle as an emitter resistance and effects within the two differential amplifiers from page 3-170 of the National Semiconductor Linear Databook 1982 referred to above a negative feedback dependent upon the product of control current times R-Eb as well as transmission characteristics dependent thereon. This means that transmission characteristics compensate only when the product of control current times R-Eb is equally great for both differential amplifiers from page 3-170 of the National Semiconductor Linear Databook 1982 referred to above. Otherwise there occurs a non-linearity which depends on the size of the deviation and which, as input signal, effects the occurrence of cubic distortions in alternating voltages. Such an inner resistance also occurs in simple multiplier circuits of the type illustrated in FIG. 12.37 of Tietze and Schenck, Halbleiterschaltungstechnik/Semiconductor Circuit Technology, 6th Edition, referred to above, producing undesirable distortions.

The compensation of undesirable negative feedback takes place logically through positive feedback of equal size, thereby already addressing the essence of the present invention.

FIG. 1 illustrates an electronic multiplier circuit with compensation for non-linearity according to a first embodiment of the present invention;

FIG. 2 illustrates a modified embodiment of the invention in a multiplier circuit similar to FIG. 1 but providing greater control of the control current or a fixed control current for the first differential amplifier; and

FIG. 3 illustrates a simple multiplier circuit with compensation according to another embodiment of the invention.

FIG. 1 is a further description of one of the possible versions for the elimination of non-linearity in multiplier circuits caused by emitter bulk resistance whereby an input signal is applied to an input amplifier, in whose negative feedback is arranged a first differential amplifier and an input of the first differential amplifier is connected to an input of the second differential amplifier, and whereby further the collectors of the second differential amplifiers are connected to the inputs of an output amplifier and multiplication of the input signal takes place with a factor through control of the current in the emitter branches. Through two voltage dividers, Rb1, Rb2 and Rd1, Rd2, a proportional portion of input signal (uE) is applied to the second input (B) of the first differential amplifier (D1) and a proportional portion of the output signal (uA) is applied to the second input (D) of the second amplifier (D2). In the second differential amplifier, a coupling is made from the output to the input of differential amplifier D2 via the voltage divider Rd1, Rd2. The magnitude of the compensation voltage coupled to input D will be dependent on the magnitude of the output voltage, and therefore upon the control current ID2. The conditions are slightly different for the first differential amplifier (D1): with increasing control current ID1, negative feedback increases via input amplifier (A1) so that output voltage of A1 decreases with steady input signal (uE), so that the signal applied to input B of transistor T2 remains the same, thus causing the signal in D1 at the second input (B) to increase relative to the signal applied to input A, when ID1 is increased. In other words, it is certain for both differential amplifiers that the magnitude of the compensation voltages depends upon the respective control current ID1 or ID2. Since it can also be assumed that the emitter bulk resistances are for the most part independent of the control currents, the harmful negative feedback caused by R-Eb can be virtually compensated when both voltage dividers Rb1, Rb2 and Rd1, Rd2 are set properly.

The following reason is responsible for the fact that compensation of R-Eb in accordance with FIG. 1 is not 100%: the disturbing negative feedback does not only work for the signals applied to the first inputs (A, C) of the two differential amplifiers but also to the same extent for the compensation signals applied to the second inputs (B, D), thus causing a residual amplification error V-R to remain. The effect of V-R depends upon the respective control currents for which both voltage dividers (Rb1, Rb2 and Rd1, Rd2) are set at the lowest non-linearity, and increases with greater maximum values for both control currents. For this reason, the compensation of R-Eb shown in FIG. 2 has an advantage over the compensation shown in FIG. 1 when ID1 has an invariable fixed value. Indeed, V-R for the first differential amplifier is then constant, and when the output amplifier as shown in FIG. 2 has inverted wiring compared to FIG. 1, the proportional portion of uA at input B can be added to the proportional portion of uE. As a result, when ID1 is constant for both compensation voltages, an equally great and constant V-R is in effect which can be taken into account when setting Rb1 and Rb3 so that compensation of R-Eb is now 100% possible.

As a welcome additional effect, the emitter bulk resistance compensations shown in FIGS. 1 and 2 also reduce or eliminate the non-linearity of the control current's overall amplification characteristic, also caused by the negative feedback in both differential amplifiers increasing with rising control currents.

There are applications in which not a reduction but a targeted boost of the cubic distortion is desirable. This is very often used for musician amplifiers in order to achieve a better sound. When a proportional portion of uE is applied to input D, the cubic distortion level can be set in the corresponding portion of uE so that the "soft" overshooting behavior of tube amplifiers can be reproduced. In addition, when using the circuit to affect the dynamics behavior of the most different signal sources, the cubic distortion can also be affected at will with the same component.

Finally, FIG. 3 also shows how the properties of simple multiplier circuits can be improved through compensation of R-Eb: As in the second differential amplifier (D2) of FIG. 1, FIG. 3 shows the compensation of the harmful negative feedback through positive feedback via voltage divider Rd1, Rd2, and the resulting linearization of the control current's overall amplification characteristic. As in the first embodiment, a portion of the output signal uA is connected via voltage divider Rd1, Rd2 to the second input D of differential amplifier D2. Compensation occurs dependent on the magnitude of the output signal and therefore upon the control current Id2 via A2, the voltage divider, and the input D of one of the transistors. Thus, the harmful negative feedback caused by R-Eb can be virtually compensated when the voltage divider Rd1, Rd2 is set properly.

Component values (example) for FIG. 1: