Add a new dimension to your music with this superb multiway springline system.
Natural reverberation is caused by reflected sounds, and in a normal room most of these reflections come from the walls and ceilings. It is important to differentiate between reverberation and echo; an echo is heard as a clear repetition of the original sound, whereas reverberation consists of numerous reflected sounds which normally come only marginally behind the original sound. Reverberation can nevertheless last several seconds due to sounds being reflected from a number of surfaces and therefore travelling a considerable distance before they reach the listener. The decay time of reverberation in a small room with carpets and normal furnishings (which tend to be good sound absorbers) is usually only in the region of a few hundred milliseconds or less, but in a large hall a decay time of a few seconds is quite common. The decay time, incidentally, is generally accepted as the time taken for the sound to decrease by 60dB. This should not be confused with the delay time, which is the time delay between the direct sound and the first reflection being received, and is naturally much shorter than the decay time in most cases.
Reverberation tends to enhance sounds by giving depth and colour, and artificial reverberation is understandably a very popular effect. There are several ways of obtaining artificial reverberation, but the most popular method, and one which provides excellent results at low cost, is the use of a springline based system.
A springline normally consists of two long springs mounted side-by-side, and having an input transducer at one end and an output transducer at the opposite end. A signal applied to the input transducer results in sound waves being transmitted down the springs, and these are designed so that there is a significant delay time. Also, in order to produce a more realistic effect it is normal for the springs to give different delay times. The sound waves produce an output signal when they reach the output transducer, but some of the energy is reflected back to the input transducer, and then back to the output transducer where a further output signal is produced. In fact, the energy will be reflected up and down the springs numerous times, gradually diminishing until no significant output is produced from the output transducer.
This is obviously analogous to sounds being reflected around a room, and for such a simple system can give a very realistic effect.
Only a limited amount of circuitry is necessary in addition to the springline, and basically all that is required is an amplifier at the input to match the signal source to the low impedance of the input transducer, and a high gain amplifier at the output. The latter is needed because of the large losses through the springline. Some designs have a built-in mixer so that the reverberation signal can be mixed with the unprocessed input signal in the desired proportions. However, this is not an essential feature, and for studio use it is usually more convenient to use a separate mixer. This design does not therefore have a built-in mixer.
A slight complication arises due to the inductance of the electro magnetic transducer used at the input of a springline. The inductance results in a given input signal voltage producing a greater current flow at low frequencies than at high audio frequencies, and this effectively boosts bass frequencies and cuts treble frequencies.
Some designs do not include any form of equalisation to give an improved frequency response, and this gives reasonable results but there is an undesirable 'boominess' due to the excessive bass and a noticeable lack of treble output.
The design incorporates equalisation to give a reduction in bass response and treble boost so that an improved frequency response is obtained. It must be emphasised that the frequency response is far from flat over the audio frequency range, and it would not be desirable to have a flat response anyway. Natural reverberation tends to have relatively little bass and treble content, and artificial reverberation must emulate this in order to give realistic results. The frequency tailoring of this design can easily be modified to suit individual tastes and requirements, as will be fully explained later on.
In fact, this design can easily be adapted to suit individual requirements in other ways. The basic circuit is designed for use with a Maplin short springline unit, but it can easily be modified to suit the Maplin long springline unit, and all this requires is a change in the value of one resistor. The prototype is actually a versatile stereo unit which has four Springlines (each having its own driver/equalisation/output amplifier module) and a built-in mains power supply. Each channel can be separately switched to use a long springline, a short springline, or both connected in series. When switched to the mono mode it is possible to have two short springlines in series or two long springlines in series. It is even possible to have all four springlines connected in series in order to give a very long decay time.
However, if you require a mono springline having a fairly short decay time, all that is required is a short springline plus a driver/output amplifier module and a mains power supply module. It is just a matter of selecting the springline or springlines that suit your requirements, and then a driver/output amplifier module for each springline must be constructed, together with a power supply module (which will power up to four driver/output amplifier modules).
The Maplin short springline unit has a decay time of 2.5 to 3 seconds and a delay time of 25 to 35 milliseconds. The corresponding figures for the Maplin long springline unit are 7 seconds and 35 to 45 milliseconds respectively.
The circuit diagram of the driver and output amplifiers is shown in Figure 1. TR1 is employed as an emitter follower buffer stage which gives an input impedance of about 50k and a low output impedance to drive the subsequent stage. C1 and R1 decouple the supply to the bias components for TR1 so that these do not couple noise from the supply lines to the input of the circuit.
Springlines are normally driven with a fairly strong input signal so that as little gain as possible needs to be used at the output of the unit. This minimises problems with microphony. In this circuit the springline is driven using a TDA2006 integrated circuit power amplifier, and this device is rather like an operational amplifier having a high power Class B output stage. Pin 1 is the non-inverting input and it is biased to half the supply voltage by R7 and R8. C5 decouples any noise which might otherwise be fed to the non-inverting input of IC1.
R9 and R6 are a negative feedback network which sets the closed loop voltage gain of IC1 equal to the value of R9 divided by the value of R5. This gives a voltage gain of unity, or about -6dB, depending on the value used for R9. A 220k is used if the circuit is to be used with a short springline, and a 470k component is employed if the circuit isto be used in conjunction with a long springline. Greater gain is needed with a long springline simply because it has slightly greater losses.
At low frequencies C4 has a very high impedance, and C4 and R5 therefore have no significant affect on the circuit. At higher frequencies the impedance of C4 decreases, and the shunting effect C4 and R5 have on R6 produces treble boost. R5 limits the maximum boost to about 20dB, but the value of this component can be increased if less boost is required, or decreased if greater treble boost is desired. It could be replaced by a potentiometer or a series of switched resistances if variable boost is required, and there is plenty of room for experimentation here.
IC1 provides little or no voltage gain, but it is mainly needed to provide equalisation and to give a low output impedance to drive the input transducer of the springline. The input impedance of a short springline is nominally 16 ohms, and that of a long springline is 8 ohms. However, as explained earlier the inductance of the input transducer produces a very low input impedance at bass frequencies with a consequent boost in the bass response of the circuit.
This is counteracted in this circuit simply by using R10 in series with the springlines input transducer. At low frequencies where the input transducer has a low input impedance R10 introduces large losses, but at higher frequencies where the input impedance is substantially higher the losses through R10 become comparatively small.
Once again this resistor can be changed in value to tailor the frequency response to suit individual tastes. A lower value gives increased bass response, a higher value gives increased bass attenuation. Variable equalisation can be obtained by using a low value (20 or 25 ohm) wirewound potentiometer or a series of switched resistances in place of R10.
R11 and C7 are needed to aid the stability of the circuit, and these do not have any significant affect on the response of the circuit over the audio frequency range.
The output amplifier uses TR2 as a high gain low noise common emitter amplifier, and TR3 as an emitter follower buffer stage which gives the unit a low output impedance. The full voltage gain available from TR2 is not needed in order to give the circuit an overall voltage gain of about unity, and R14 is therefore used to introduce negative feedback which slightly reduces the voltage gain of TR2.
Figure 2 shows the circuit diagram of the power supply, and this is a straightforward circuit having full wave rectification and the output stabilised at a nominal potential of 15 volts by monolithic voltage regulator IC2. The maximum output current is 1 amp and this is sufficient to supply four springline circuits.
The switching of the Multireverb Unit is shown in Figure 3. S1 is the mode switch for the right hand channel and it enables the signal to be routed through the short springline only, the long springline only, or through both springlines. S2 provides identical switching for the left hand channel.
S3 is the mono/stereo switch, and in the stereo mode it merely connects the two stereo channels through their own separate paths. In the mono mode the output of the left hand channel is isolated from the left hand output socket and is instead routed to the right hand input (which is isolated from the right hand input socket). The two input sockets are connected in parallel, as are the two output sockets.
Thus, in the mono mode S1 and S2 provide the same options as in the stereo mode, but as the two channels are connected in series it is possible to obtain two short springlines in series, two long springlines in series, or two long and two short springlines in series. It is also possible to obtain a short springline plus a long springline in series, which is also available in the stereo mode. However, just a short springline or a long springline cannot be obtained in the mono mode, although this can obviously be achieved by switching to stereo and ignoring one channel of the unit.
The springline driver and output amplifier circuits are assembled on one printed circuit board (one board being required for each springline used) and the power supply is assembled on a separate board. Readymade printed circuit boards are supplied with a driver/output amplifier board and a power supply board as a single panel which must be cut along the broken line in order to separate the two boards or can be left attached as in Figure 4. Of course, if you are using more than one springline there will be spare power supply boards which can be saved for future use.
Details of the driver/output amplifier circuit board are provided in Figures 4 and 5. Note that IC1 is bolted to the printed circuit board using a 6mm M3 bolt and fixing nut. The leadouts of IC1 are preformed and need to be slightly reformed before IC1 can be fitted on to the board properly.
Figures 4 and 5 give details of the power supply printed circuit board and the other power supply wiring. The fuseholder is bolted to the board using a 6mm M3 bolt. IC2 does not have to dissipate a great deal of power, and a ready-made finned heatsink is adequate even if four springline circuits are powered from the supply unit.
Mechanical construction of the unit can obviously be varied to suit individual circumstances and requirements, but the basic arrangement used for the prototype should serve to give some good guidelines. The case has chipboard sides and a chipboard rear panel on which the springlines and printed circuit boards are mounted. The top and base panels are made from 18 swg aluminium, and the three controls and four sockets are mounted on the top panel. The base panel is drilled to take the power supply cable which connects the output from the power supply to the reverberation unit. In our prototype the power supply board and mains transformer were fitted in a separate case which stands on the floor beneath the reverberation unit (which is designed for wall mounting). This keeps the mains transformer well separated from the springlines so that there is no possibility of inductive coupling between the transformer and the springlines. Flowever, if required the PSU board can be left connected to the driver board and mounted in the case as in Figure 4.
Finally, the wiring required to complete the unit is detailed in the wiring table and Figure 4.
Feature by Robert Penfold
Previous article in this issue:
Next article in this issue:
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!