Not just echo effects
A versatile delay line for deriving 'short echo' and related effects. Design by Peter Rodgers
The echo effect is one that must be familiar to virtually everyone, whether they have an interest in electronic music or not, and the name is really self explanatory. This effect is produced using some form of electronic delaying device, and the delayed signal is fed back to the input of the circuit so that the signal is circulated repeatedly, gradually fading away. A signal fed to the input therefore appears at the output a number of times, giving what is quite a good analogy of a real echo where sounds are reflected 'to and fro' until they decay to an inaudible level.
There are a number of ways in which the delay can be obtained, such as tape methods, digital delay lines, and various forms of analogue delay line. For a home-constructor's echo unit a long 'bucket brigade' analogue delay line probably represents the most practical approach, and this design is based on the MN3011 analogue delay line chip. Even at maximum delay this gives only a fairly short echo effect, but it nevertheless enables some excellent effects to be obtained, and the unit will expand the 'repertoire' of most electronic musical instruments.
Figure 1 shows in block diagram form the various stages of the unit and their general arrangement. Four of these stages, the clock oscillator, the two lowpass filters, and the delay line, are used to provide the delayed signal. The clock oscillator controls the length of the delay, and the output filter is used to prevent the high frequency clock signal from breaking through at the output. The input filter removes any high frequency signals which might happen to be present on the input signal, and which could otherwise produce heterodyne 'whistles' on the output signal.
A low frequency oscillator can be used to vary the clock oscillator's frequency, and thus the delay as well. As the clock frequency is raised, the signal within the delay line is moved out more rapidly than it was fed in, giving a boost in the output frequencies. As the clock oscillator is reduced in frequency the opposite occurs, with a consequent reduction in the output frequencies. In other words vibrato can be introduced using the clock oscillator (which can be switched out if vibrato is not required).
The other two stages are both mixers. The one at the input simply combines the input signal and the delayed signal. The strength of the latter can be varied, and this enables the time taken for echoes to fade away to be adjusted. The second mixer simply combines the output of the delay circuitry with the input signal to give the required output signal.
Both mixers use the same configuration, as shown in Fig. 2(a), and this is a conventional operational amplifier summing circuit. If all three resistors (Ra to Rc) have the same value, the action of the circuit is very simple indeed. The output counteracts the input voltages so that the inverting (-) input of the operational amplifier is maintained at whatever bias voltage is fed to the non-inverting (+) input. For instance, if input 1 is 1 volt positive and input 2 is 4 volts positive, the output will go 5 volts negative to counteract the total input of +5 volts. This gives the required mixing action, and although the signal is inverted through the circuit, in audio applications this is not important (an inverted signal does not sound any different).
If one of the resistors is made variable, like Rb of Fig. 2(a), the gain at this input can be varied by adjusting this component. If it is made higher in value a given input voltage produces less current flow, and a smaller change in output voltage is sufficient to counteract the signal at this input. A lower value gives higher current flow and demands a greater change in output voltage (ie gives higher gain from this input to the output). This feature is used in the circuit to permit the level of delayed signal feedback to be controlled.
The two filters are basically the same, and use the arrangement shown in Fig. 2(b) If we ignore capacitor Cb for the time being, Ra and Ca form a simple R-C lowpass filter, and the series resistance of Rb plus Rc forms a second filter in conjunction with Cc. In both cases the capacitor represents a difficult path for the input signal at low and middle audio frequencies, but as the input frequency is increased each filter capacitor provides an increasingly easy path. At high frequencies most of the input signal tends to be tapped off to earth, and little finds its way through to the output. The buffer amplifier ensures that the circuit has a low output impedance and that loading of the filter with a consequent loss of performance cannot occur.
A drawback of a simple filter of this type is the rather gradual introduction of the filtering, and a level of performance which is inadequate for many applications. This is overcome by the inclusion of Cb. At frequencies where the other filter components provide only moderate losses Cb couples the output of the buffer amplifier back to the input so that the input signal is reinforced and the losses are reduced. The opposite tends to occur at frequencies where the other filter components provide heavy losses since the output signal is too small to provide significant feedback. Instead Cb taps off some of the input signal into the output stage of the amplifier, giving increased losses. This gives a rapid introduction of the roll-off and an attenuation rate of 18dB per octave (ie a doubling of frequency causes the gain of the circuit to fall by a factor of 8).
Delay line chips are very complex devices which normally have the equivalent of several thousand components, but in principle their operation is quite straight forward, as follows.
The input signal is sampled by the first two stages when the switch/buffer stage forces the storage capacitor to take up a charge equal to the input potential. The first switch is then turned off and the second one is activated so that the charge level on the first storage capacitor is fed through to the second one. Then switch 2 is turned off and switch 3 is activated so that the sample is coupled through to storage capacitor 3, but switch 1 is also turned on so that a new input sample is placed in storage capacitor 1. This process is repeated indefinitely with new samples being taken at the input and fed through to the output.
The delay obtained depends on the number of stages, and in practice there would be from a few hundred to a few thousand. It also depends on the rate at which the switches are operated, and this is controlled by the clock oscillator. The output signal is not a true replica of the input signal in that it varies in steps rather than continuously, but provided the clock frequency is at least double the highest input frequency, the lowpass filter at the output removes the steps and restores the correct output waveform.
The clock oscillator uses a chip specifically designed for use with the MN3011 delay line (the MN3101), and this requires few discrete components. Apart from providing the clock signal it also generates a bias voltage for the MN3011.
The low frequency modulation oscillator uses the configuration outline in Fig. 2(c). This is based on an operational amplifier used in a trigger circuit. Ca first charges via Ra until the charge potential reaches a certain threshold level. The output of the trigger circuit then goes low and Ca discharges until another threshold voltage is reached. The output of the trigger then goes high again and Ca starts to charge. This continues indefinitely with the charge on Ca rising and falling to produce a roughly triangular waveform. This waveform is excellent for use in modulation applications such as the present one. A buffer amplifier is used at the output of the oscillator to prevent excessive loading on Ca.
Fig. 3 shows the main circuit of the Short Echo Unit, and Fig. 4 shows the circuit of the mains power supply unit.
IC1a is used as the input mixer stage and VR2 is adjusted to a point which makes it impossible to have excessive feedback (which would result in the unit oscillating). The input filter uses IC1b as the buffer amplifier. IC2 is the delay line chip, and IC3a is utilised in the output filter. IC3b is used in the output mixer, and if S1 is open the delay line is cut off from the output and the echo effect is switched out.
IC4 is the clock chip, and the frequency (echo delay) is controlled using VR4. The modulation oscillator uses IC5, and the modulation frequency can be varied using VR6. The modulation depth is preset using VR5, and with S2 open the modulation is switched out.
The power supply is a simple stabilised 15 volt type which has a very low output ripple content.
Start construction by building up the printed circuit board. This accommodates all the resistors, capacitors, semiconductors and presets. Details of the board are provided in Fig. 6.
Construction is likely to be easiest if you start with the two link wires, then add the resistors, followed by the capacitors, presets, and semiconductors. IC2 and IC4 should be left until the very end, and these should be mounted in IC sockets (a 8 pin DIL type for IC4, and an 18 pin DIL type for IC2). IC2 has a rather unusual pinout configuration with the middle three pins of each row absent so that there are actually just 12 pins and not 18. As the printed circuit board only has holes for the 12 pins that are actually present the socket for IC2 must be modified by either trimming away the unwanted pins using a pair of wire cutters, or simply pulling them out using a pair of pliers.
Both IC2 and IC4 are MOS devices which are vulnerable to damage by static charges. They should be supplied in some form of protective packaging, such as having their pins embedded in conductive foam. Leave them in this until it is time to plug them into circuit, and handle them as little as possible. Unfortunately, ICs tend to have their two rows of pins splayed slightly apart, and it will probably be necessary to squeeze the two rows together slightly before there is any chance of plugging them into the sockets.
It is helpful to fit Veropins at the places on the board where connections to the controls and other off-board components will eventually be made. The board can then be mounted in the case prior to wiring it to the rest of the unit, rather than the more awkward method of wiring it into circuit and then fixing it in place.
The finished board should be checked thoroughly for errors and corrected if necessary. However, it is much better to take great care when fitting the components onto the board and to avoid errors rather than to have to correct them!
A metal instrument case is probably the best type for a project such as this, and the prototype is housed in a case of this kind which has approximate outside dimensions of 250 by 150 by 75mm. This is somewhat larger than is really necessary, and it should be possible to use a smaller case without too much difficulty.
The photographs show the front panel layout of the unit, but it is not essential to exactly copy this. One important point is to keep the mains transformer and mains wiring reasonably well separated from the rest of the circuit so that there is no significant pick-up of mains hum, and a hum-free output is obtained. Note that the neon used for LP1 must be a proper panel mounting type having an integral series resistor for use on the 240 volt mains supply, and must not be a simple neon bulb (which would rapidly burn out).
The printed circuit board is mounted on the left hand section of the chassis or base panel of the case using M3 or 6BA fixings. Stand-offs must be used on the mounting bolts to keep the connections on the underside of the board clear of the case or chassis. T1 is mounted at the opposite end of the chassis or base panel so that it is as far away from the printed circuit board as possible. A soldertag is fixed to the case or chassis to provide an easy connection point for the mains earth lead, and this can be fitted on one mounting bolt of T1. The fuseholder is mounted at any convenient point on the base panel or chassis using a short M3 or 6BA bolt plus a fixing nut.
A hole for the mains lead is made in the rear panel opposite S3, and this hole should be fitted with a grommet to protect the cable.
The wiring is shown in Fig. 6 (in conjunction with Fig. 5), and is all quite straightforward. However, take considerable care not to make any errors in the mains wiring as apart from the damage this might cause, it could also be very dangerous. Connecting the wires to the pins and component tags should not be difficult provided the pins or tag, and the end of the wire, are generously coated with solder prior to attempting to make each connection. It is advisable to use multistrand hook-up wire rather than the single core variety. It is not necessary to use screened cable to connect SK1 and SK2 to the board, but the external input and output leads should be the usual screened jack leads.
Start with VR1 set fully anticlockwise, VR2 at a roughly middle setting, and VR5 set full clockwise. VR2 is set to optimise the large signal handling capability of the circuit, and if a sinewave generator and oscilloscope are available VR2 is set to give symmetrical clipping. In the absence of this equipment VR2 is simply given any setting that gives no significant distortion at peak volume levels.
With VR3 set at minimum resistance (fully clockwise) VR1 is adjusted as far as possible in a clockwise direction without causing the circuit to oscillate. With the vibrato switched in using S2, VR5 is set to give the desired degree of modulation, but if it is adjusted too far in an anticlockwise direction the clock oscillator will only operate intermittently and the unit will fail to operate properly. Results will probably be best with VR5 set just below this point.
The noise level of the unit is very low at the shorter echo times, but the noise level increases significantly when VR4 is adjusted for longer times. The signal to noise ratio should be perfectly adequate though, provided a fairly high input signal level is used. The unit can comfortably take an input of one volt RMS. Incidentally, as is normal practice with this type of equipment, the bandwidth of the delay line has been restricted to about 5kHz in order to minimise noise on the output.
The best way to discover the control settings and types of input signal that give the most effective results is to experiment as much as possible using the unit. However, in general the best effect is obtained using a short percussive input signal, a substantial amount of feedback to give a fairly long decay time, and a medium to long echo time. When using a synthesizer to provide a short burst of sound bear in mind that vibrato cannot be obtained using the built-in modulation facilities of the synthesizer. Due to the brief nature of the sounds this would give an output of (more or less) random pitch! The vibrato effect of the echo unit must be used.
Feature by Peter Rodgers
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!