The Matinee Organ (Part 1)
PART 1 OF A SERIES SHOWING YOU HOW TO BUILD A COMPLETE ELECTRONIC ORGAN AND SAVE YOURSELF HUNDREDS OF POUNDS.
In the early part of this century almost every home had a piano, but the advent of recorded music and hi-fi turned most people away from making their own music and thousands of pianos were scrapped. Today we are witnessing a renaissance in home music making, brought about by the arrival at an affordable price of the easy-to-play home entertainment organ.
Home entertainment organs are now available in the shops at prices ranging from a few hundred pounds to several thousand pounds. A careful study, however, shows that ready-made organs selling in the £700 to £1,000 price range are a "best buy" for home entertainment.
Organs priced at under £700 are rather basic and although on the face of it, ideal for the beginner, their limited range can soon restrict the more ambitious player. Organs priced at over £1,000 have extra facilities, but apart from orchestral voicing, string and brass chorales, the extra features have a limited usefulness and unless you are a highly proficient musician they are not worth the extra price. Certainly, if you can afford it, orchestral voicing, string and brass chorales are worth having, but they are still rather expensive.
The organs in our "best buy" price range have four main features: two manuals, a 13-note pedalboard, a rhythm unit and a simplified chord playing facility commonly known as single finger chording. Table 1 shows how the Matinee compares with twelve of the most popular ready-made organs in the shops at the moment. Naturally it was not possible to compare every switch and facility, and all the organs in the survey did have other features which are either shared by the Matinee or restricted to a particular manufacturer. The Matinee too has some features not found on the other organs in the survey as a close examination of the specification will show.
You will see from Table 1 that the Matinee compares very favourably with the commercial organs and the fact that it has drawbar voicing, more than overcomes any apparent deficiency in upper manual voices, because having all the voices on drawbars allows an almost infinite variation of sound, greatly surpassing anything possible with fixed volume stops.
The one stand-out feature on Table 1 is of course the price, but none of the saving is due to a lack of quality or to inferior circuits.
The organ sounds every bit as good as the commercial organs and it looks as good as well. Ready-made punched and printed front panels are available and a complete cabinet of equal quality to those found on the other organs in the survey is included in the price. For those who wish to build their own cabinet, we shall publish full cutting and construction details in the last part of this series and this could save you another £70 or £80 as the price for everything except the cabinet is only £299.95.
Some of the price saving certainly comes from the fact that you have to build the organ yourself, but much of the saving is the result of a major technological advance in integrated circuits designed for electronic organs.
In 1978, a young graduate from the University of Bologna in Italy, Mr Giuseppe Ravaglia joined SGS-Ates, a company that was already actively involved in the design of integrated circuits for the electro-music scene. Mr Ravaglia has the rare distinction of having not only a degree in electronic engineering, but also of having studied music theory at the Municipal School of Music. Since he joined SGS-Ates, he has been working at their headquarters near Milan in Italy on the development of microprocessor controlled organs with programmable voicing and specification.
The first step towards this goal was to integrate on one chip what might be considered the heart of an organ — the tone generation and gating system. By late 1979 he and the team of SGS Digital IC's Applications Group had a working breadboard of such a system. It consisted of over 300 CMOS integrated circuits and yet by mid-1980 every part of this had been condensed onto one integrated circuit chip measuring less than two tenths of a millimetre square. They called their baby, the M108.
Not a very exciting name, but definitely a very exciting chip. To fully appreciate their achievement, it is necessary to look briefly at the way the silicon chip has been advancing on the electronic organ over the past few years.
During the 1960's, an organ with an integrated circuit in it would have been a rarity, but by the late 60's the first IC specifically designed for electronic organs was on the market. It was a top octave synthesiser IC that allowed the twelve musical notes of the top octave of an organ to be generated from one very high frequency. In earlier years electronic organs had to have separate generators for each note in the top octave with the result that one oscillator could go out of tune, making one note sound flat or sharp in every octave on the keyboard and you had to take the whole organ to pieces to cure the fault. The top octave synthesiser IC's eliminated this problem since all the frequencies could be adjusted from one control to pull the whole organ into tune in seconds.
Every note on the organ can then be created by dividing each top octave note by two, several times. For example, the C below top C is exactly half the frequency of top C and so on down to bottom C (see table 2).
The second advance came in the integration of rhythm generator IC's. For some years now, all popular electronic organs have had rhythm generators, mostly coupled with some form of simplified left-hand playing facility such as single finger chording. The early rhythm generators were boards of digital IC's and early chord generators were built from banks of diodes. In the mid-70's, both the rhythm and chord generators were condensed onto one IC or in some cases a two IC set.
The new M108 IC combines the top octave generator, the dividers, the rhythm generator timing, the chord generator and all the gating. It is the inclusion of this latter item that eliminates most of the tedious wiring and makes this a really easy organ to build. Instead of each keyboard needing 200 wires, the Matinee keyboards need just 17! This represents a saving of well over 80% in wiring the whole organ.
The gating is one of the most important features of an organ. What we mean by gating is switching the raw frequency to the main bus bars. For example, if middle C is pressed, we need 261.6Hz to be switched to the 8 foot voices, 523.3Hz to be switched to the 4 foot voices and 130.8Hz to be switched to the 16 foot voices. Then, if for example the 8 foot Flute stop is pressed, any frequency switched through to the 8 foot voices will pass through the Flute filters and the characteristic Flute tone will be heard at the correct frequency.
Various forms of gating are available. The simplest is the mechanical switch fitted under the key, or in our case with three footages we would need three separate contacts under each key. This is mechanically difficult and expensive and a great deal of wiring is required; 3 wires for each key if there are 3 footages. In addition, if envelope shaping is required, a further contact would be needed as a key-down detector, generating even more wiring.
The other most commonly found form of gating is the diode or transistor switch. This has the advantage of needing only one contact under each key regardless of how many footages there are. However, it still needs the same number of wires even though most are no longer connected to the key switches. One major advantage is that the attack and decay can be set individually for each key, though this is expensive to do and is not found on organs in the £700 to £1,000 price range. The disadvantages are that there are lots of components and the output is not symmetrical, so if a fast attack was required there would be an audible thump in the output (caused by the average DC level suddenly changing). Another disadvantage is that some of the signal leaks through a semiconductor switch even when it is fully off because of the inherent capacitance of a diode or transistor. It is a very tiny signal, but in an organ with lots of footages there will be hundreds of switches in parallel all adding their little bit. The result is a just audible broadband noise that is generally referred to as "singing".
The gating in the M108 is performed by MOS switches which unlike diodes and ordinary bipolar transistors do not allow any signal to pass when they are off and thus there is no singing in the Matinee organ. Also the output of the M108 is perfectly symmetrical, however many notes are played simultaneously. And finally the biggest advantage of all, of course, is that massive saving in wiring. Only 17 wires are required for each keyboard and only one simple contact is needed under each key, including the key-down detector.
The key-down detector is required so that we can provide envelope-shaping of some of the voices. In early electronic organs, the raw tones were simply filtered to produce the correct waveshapes to simulate the characteristic sound of an instrument. However, the voices did not sound very realistic, because one of the most fundamental characteristics of the sound of an instrument were missing. This characteristic was the way in which the volume of the sound changed with time, in particular at the beginning and end of a note.
If a key is struck on a piano, for instance, the volume of the sound reaches a maximum almost instantaneously, whereas if an accordion starts to play the sound it takes a whole second or more to reach full volume. We say that a piano has a fast attack; its envelope opens and reaches a maximum very quickly, whereas the accordion has a slow attack. If we have an amplifier whose gain we can control, we can simulate these attack rates by switching a particular preset volume change to the amplifier gain control when a given stop is pressed. All we need to know is when to start the volume change and hence the need for a key-down detector.
Similarly the amplifier could be set to close down the volume quickly or slowly when the note ends. Again different instruments stop sounding at differing speeds. For example an undamped string instrument will carry on sounding for a long time after the string is plucked; the volume gradually becoming less and less, whereas a wind instrument stops sounding almost the instant the musician stops blowing. We say that undamped stringed instruments have a slow decay whereas wind instruments have a fast decay. Again, the key-down detector is used to signal when the last key is released.
For example, on the Matinee the piano voice has a very fast attack and a fairly slow decay if the keys remain held down. However, if the keys are released then a fast decay is set in motion in order to correctly simulate the action of the dampers on a real piano. In addition, a control is provided that works like a "loud pedal" to sustain the notes for a long time or in other words to give a very slow decay. Although the basic tone of the piano voice on the Matinee sounds like a piano, the addition of this envelope shaping transforms the voice and makes it dramatically realistic. The other voices on the Matinee are treated similarly to add realism to the sounds.
The M108 makes all this possible very easily and it is this that has led to the low price and simplicity of construction. In fact the only qualification you need to successfully build the Matinee is the ability to solder correctly. Virtually all the electronics are built on one large printed circuit board with only the main power supply located elsewhere. All the wiring is via simple PCB plugs and sockets so that everything just plugs together for absolute simplicity of construction. All the fixings, metalwork and woodwork are available for you to just bolt together if you wish, guaranteeing you a first class finish that will grace any living room.
Although we shall be going into the circuitry of the Matinee in depth, we wish to make it quite clear at this point that you do not need to read or understand any part of these technical descriptions to successfully build this organ. In fact many of the terms used may be a mystery to you, but anyone reading our beginners course that commences elsewhere in this first issue of E&MM will quickly gain an understanding of these descriptions.
Two M108 IC's are used in the Matinee, one for each keyboard. For the pedalboard a different IC is used: the M147 which will be described later. At this stage it is sufficient to know that the M108's require a master frequency of 1,000,120Hz to generate all the notes of the keyboard and carry out all their other timing functions, and the M147 requires a master frequency of 500,060Hz. In Fig. 2 we show the circuit of the master oscillator used in the Matinee from which all the tones and timing throughout the whole organ are derived.
The main oscillator is formed by IC36, a CMOS device containing four separate NOR gates. When power is applied, the voltage at the input to IC36c is low, at -6V, because C127 is uncharged. Since the input is low, the output of gate 'c' is high. Similarly the output of gate 'b' is low and gate 'a' high also. Gates 'a' and 'b' and R378 form a Schmitt trigger that greatly improves the sharpness of the edges of the waveform. The output of 'a' being high, now charges C127 at a rate set by RV32 which should be adjusted so that middle A on the keyboard produces 440Hz. When this is the case the frequency of the oscillation at this point will be exactly 2.00024 MHz.
When C127 is charged the input to gate 'c' will be high and hence the output now goes low. The outputs of gates 'b' and 'a' are now also forced to change state and the output of 'a' being low discharges C127 until the input at 'c' is low again. This process repeats itself continuously two million times every second. This oscillation is then presented to gate 'd' which simply acts as a buffer for the square wave output.
IC35 is a dual op-amp which produces an oscillation in much the same way as IC36 except that it is a triangular waveform and at a very low frequency. RV28 is the vibrato rate drawbar and sets the frequency between 2Hz and 20Hz. The amount by which the vibrato changes the frequency of oscillation of IC36 is set by the vibrato depth drawbar RV30.
With the vibrato switch SW20 off, the output of IC35 is short circuited to ground, but when switched on the output is coupled through the DC blocking capacitor C123 to the base of TR38. This transistor is supplying the main positive DC voltage that allows IC36 to operate. As the supply rail voltage changes under the control of the vibrato oscillator, so the points at which the switch-over thresholds in IC36 occur, change and the frequency of oscillation also changes slightly.
When the delay vibrato switch, SW21 is on and no key is pressed on the keyboard, C126 is charged and TR36 fully on, short circuiting the output of IC35. When a key is pressed, the positive voltage on D135, D136 is removed and C126 discharges slowly through RV29. This preset adjusts the length of the delay before C126 is sufficiently discharged to allow TR36 to turn off. As it does so the output of the vibrato oscillator is no longer inhibited. At the next instant when no keys are pressed on the keyboard, C126 charges again ready for the next key to be pressed.
A glide switch is fitted on the foot pedal to give Hawaiian Guitar effect. When the switch is operated the current through TR37 changes slightly over a short period set by the position of RV31. The effect of this is much the same as changing the voltage on the base of TR38: the frequency of the main oscillator is changed. The change is set by R375 and R376 to give a one semitone change on the keyboards. When the switch is released the current through TR37 returns to normal at the rate set by C128.
Because the main voltage rail to IC36 is used to produce the organ's vibrato and glide effects the output of gate 'd' can under certain circumstances reach voltage levels that are insufficient to correctly drive IC37. TR39 is therefore provided to apply sufficient amplification to the output of gate 'd' to ensure that IC37 is correctly driven in all conditions.
IC37 is a CMOS IC containing two D-type flip-flops that are connected here to function as frequency dividers. The main oscillator frequency of 2.00024 MHz is applied to pin 3, the input to the first flip-flop. The output at pin 1 is exactly half this frequency: 1.00012 MHz, the frequency required by the M108's. As well as being connected to the clock input of the M108's it is also connected to pin 11 on IC37, the input to the second flip-flop. The output at pin 13 is exactly half the frequency on pin 11: 500,060Hz, the frequency required by the M147.
It may seem strange at first sight that we have set the main oscillator to run at 2MHz when the highest frequency used is 1MHz. There are, however, several advantages in starting with a higher frequency than that required. Firstly the higher frequency gives better accuracy and the Schmitt effect of IC36 gives excellent long term stability and secondly the 4013 generates an excellent square wave shape whose mark/space ratio is exactly 1:1. This circuit, then, provides an ideal waveform to drive the clock inputs on the M108's and M147.
Next month, we shall begin the construction details of the Matinee organ and continue the description of the circuits used.
[Please check the corrections listed in part 6 of this series - where possible, corrections have been applied to the text but in some cases there are circuit diagram amendments.]
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!