PAiA 4710 Balanced Modulator Design Analysis originally published 1974 see accompanying files: 4710balm.gif for schematic 4710figs.gif or fig1, fig2 Balanced modulators are identified by a number of different names and to the average musician the most familiar is probably ring modulator. A ring modulator is simply a specific means of implementing a balanced modulator using a diode ring as the modulating element. As far as function is concerned there is no difference between the two. For the purposes of this explanation a third name, Four Quadrant Multiplier, not only gives us something to call the module but as we shall shortly see is also an exact statement of what it does. Figure 1 shows three possible ways of combining electronic signals - mixing, amplitude modulation, and balanced modulation - and the results of each process. Notice first that when two signals are mixed in the audio sense nothing changes, (figure la) Two frequencies go in and the same two frequencies come out. Figure lb. shows the result of simple amplitude modulation and is representative of the situation that one normally encounters when working with a Voltage Controlled Amplifier. Two frequencies are applied to the inputs of the VCA and the amplitude of one directly determines the amplitude of the other. By convention, the higher of the two frequencies is commonly designated the Carrier and corresponds to the audio input of a VCA. The lower frequency is designated Modulation and corresponds to a control voltage input. There are a number of things of particular interest in figure lb. For example, notice that this is a multiplier of sorts. When the modulation signal is zero, the output is zero and as the modulation voltage increases the output also increases. From working with VCAs we know that negative control voltages have no effect on the output, the amplifier is already off at zero volts and if anything, negative control voltages only turn it more off. In analog computer terms - and let's face it, a synthesizer is more analog computer than anything else - a VCA is a two quadrant multiplier, indicating that the multiplication process is valid for positive or negative carrier voltages but that the modulation input is constrained to only positive voltages. Notice the output of this two quadrant multiplier. Unlike the mixer, the output does not contain the two input frequencies, instead, it contains the original carrier frequency and two brand new frequencies that are the sum of the original carrier and modulation frequencies and the difference of the two. These Sum and Difference frequencies are commonly called upper and lower side-bands respectively. Of particular note is the fact that the modulation frequency has been "rejected" and does not appear in the output. In a mathematical analysis, this is the reason that a good VCA will not "pop" when subjected to rapid control voltage changes. Figure 1c illustrates a balanced modulator. Once again we have the multiplication process in which increasing the voltage at the modulation input produces increased output level, but notice that in this case we have eliminated the constraint of keeping the modulation voltage always positive. This, then, is a four quadrant multiplier in which the sign (phase) of the output is valid for positive and negative modulation signals as well as positive and negative carrier signals. NOTE: in a balanced modulator the terms "carrier" and "modulation" actually lose their meaning since as a practical matter these two signals are interchangeable. We will continue their use, however, because in the 4710 a distinction is made between these two signals to ease the module's use in auxiliary applications. Notice the frequency content of the output. Not only have we dropped out the modulation signal as we did with the VCA, now we have rejected the carrier as well. Neither of the two original signals is present and only the side-bands remain. If we stick to sine waves as the inputs to the modulator not a lot has been accomplished because the side-bands are sine waves too, but, when non-sinusoidal inputs are used the situation becomes considerably more complicated and musically much more interesting. As has been pointed out innumerable times before, any non- sinusoidal waveform can be broken down into a summation of sine wave components of various phase, amplitude and frequency relationships. What hasn't been strongly stressed is that until you begin working with a balanced modulator you are restricted almost exclusively to sounds that are harmonically structured; they have a fundamental frequency followed by overtones that are all integral multiples of that frequency. As it happens, most natural sounds have these same properties but by no means all. Fig 2 shows two frequency spectra, the first (a) for a balanced modulator with sine waves at both inputs and the second (b) with a square wave substituted for one of the sine waves, In a frequency spectrum the position of some vertical line along the horizontal axis indicates the frequency of some component of the signal being examined while the height of the line indicates the relative amplitude of that component. For 50 Hz, and 800 Hz. sine wave inputs (dashed lines) the output (solid lines) consists of sinusoidal side- bands at 750 Hz. and 850 Hz. Substituting a 50 Hz. square wave for the 50 Hz. sine wave produces the spectrum shown in (b). For simplicity only the first 4 harmonics of the square wave are shown and they appear at 50 Hz., 150 Hz., 250 Hz. and 350 Hz, respectively, These components are dropped out in the modulation process and in their place the side-bands appear. Each of the frequencies that appear in the side-bands is the sum or difference of one of the components in the square wave and the 800 Hz. carrier. Notice in particular that as predicated, these side-band components are non-harmonic; that is, there is no lowest frequency fundamental with additional components appearing at integral multiples. DESIGN ANALYSIS The 4710 Balanced Modulator is built around a type 1496 balanced modulator integrated circuit. The internal workings of this I.C. package are not pertinent to this discussion but certain general points should be made. The 1496 chip has differential inputs for both carrier and modulation ports. Biasing for the carrier inputs (pins 8 and 10) are supplied by the voltage divider consisting of Rll and R12 with this biasing voltage coupled to the IC through R14 and Rl5. Input signals applied to the carrier input jack J4 are capacitively coupled by C2 to input isolating resistor R26 and finally appear across R14. Fixed resistors R13 sad R16 along with trim pot R25 are used to supply adjustable current into the carrier input port to balance out variations that occur in the integrated circuit during manufacture. With no signal applied to either of the inputs there should be 7v +/-20% present at pins 8 and 10 sad at the junction of Rll and Rl2. The modulation inputs (pins 1 and 4) are similar to the carrier inputs except that they are tied to ground through R8 sad R9 with R7, R10 and trimmer R24 supplying tolerance compensating bias currents. Pins 1 and 4 should read 0V +/-0.1V. R19 and R20 serve as load resistors for IC-2 with the in- phase output coupled to output jack J5 by way of C3. quiescent voltages at pins 6 and 12 should be 8v. +/- 20%. The modulation input of the 47I0 module is buffered by the 748 type operational amplifier IC-1. Control voltages are direct coupled to the inverting input of that IC by R2 and R3 while audio signals are capacitively coupled by R1 and C1. With the modulation LEVEL control fully counter-clockwise and no signals present at the inputs, pin 6 of IC-l should be 0V. +/- 0.1V. Rotating the LEVEL control in a Clockwise direction causes less of the output of the operational amplifier to appear as feed-back at the inverting input which in turn causes the voltage gain of this amplifier to go from approximately 4.5 to 90 (13 - 40 db.) for signals applied to the audio input and 0.3 to 6. 3 (-10 to 16 db.) for signals applied to the control inputs. The voltages at pins 2 and 3 of IC-l should be 0V. +/- 0.1V and pins 7 and 4 are positive and negative supply respectively. The overload Light Emitting Diode becomes forward biased and begins to conduct on negative excursions of the output of IC- 1. Because of the forward voltage drops of D1 and D2, as well as the LED itself, conduction begins to occur at about - 2v. The output of the buffer amplifier is coupled to IC-2 through R6. Power supply decoupling is provided by the combinations R2l, C6; R22, C7 and R23, C8. The voltages across C6, C7 and C8 respectively should be 13V., 7V. and -8V. +/- 20%. (c) 1995 PAiA all rights reserved phn 405-340-6300 fair use copy only with this notice