This page has had views

Why do humans endeavour to do anything? The answer to this must be for the challenge. Whether it be to climb a mountain, go to the moon, whatever- ths motivation is purely to see if it can be done. In my case, I had been listening to and appreciating organs for most of my 45 years. The other factor is the Internet. No matter what you are in to, there is any amount of information and experience available on the Web and organ building, both amateur and professional, is no exception. So the die was cast, and the work began..

I decided at the outset that as this was something I had no experience of, I would do it scientifically. When the Wright brothers built their aeroplane, they succeeded where others had failed through the use of a scientific process. This meant building something, evaluating it, if it didn't perform as expected build it again with modifications, evaluate again, and so on until they got it right. This was the pattern I decided to follow.

I decided that the instrument should be fully chromatic so anything could be played on it, cover four octaves with room for expansion, and be of a size to fit the domestic environment. My wife is used to me building odd things by now so this wasn't too much of an issue, although she certainly had concerns at the outset. But, above all, it had to fit within a very restricted budget. So this project is really about how to build an 'economical' pipe organ.

I needed the instrument to be as compact as possible so a vertically stacked layout was chosen. It consists essentially of a 'blower chest' at the bottom, with feeders under and the reservoir on top, and mounted above this is the actual pipe windchest, containing the action.

The footprint on the floor is 12 x 34 inches, and the height of the completed instrument will be about 6 feet.

The quality of the final result depends on a reliable, even wind supply. I decided against the usual type of electric rotary blower firstly to keep costs down, and secondly for reasons of noise level. All other home builders who have used these in the same room as the organ seem to have had to build them into soundproof boxes, which is going to be big, heavy, and a lot of work.

I was fascinated by the blowing arrangements originally used in fair and portable organs and from this practice I derived the present system of feeders operated mechanically by a crank. This system can be made practically noiseless, and the low power requirements seemed to indicate that power from a DC motor was entirely possible.

The difficult decision to make was that of wind pressure. In the event, the answer was provided by the experimental pipes I made, which could be voiced to speak with the volume and tone I wanted on around 1 to 2 inches of wind. I needed a water gauge to measure the pressure, and this, like everything else, was home made. I calculated the wind requirements of the complete organ from information available on the Internet, added 50% for good measure and then calculated the size of feeders required, based on their volume and operating speed. The reservoir will provide a more than adequate volume of extra wind if required.

In organ terminology, a bellow used to feed wind to an organ is called a 'feeder'. The three feeders I have built are of the wedge type and are mounted on the underside of the blower chest. Each feeder consists of a top and bottom board, 10 x 10 inches, leathered with stiff card (mounting board) reinforcement pieces glued inside. The fixed top board is cut out to give access to the interior and inlet valve, and the moving bottom board has a 1.25 inch hole in it, covered on the inside by the inlet flap valve. The valve is a strip of rubber cut from a car tyre inner tube. Originally, the valve strips were fastened at each end; relying on the flexibility of the rubber to allow enough space for air to pass but this was found inadequate and one of the fastenings was removed. Although slightly noisier (the valve 'taps' closed) this works really well. The feeders are screwed to the underside of the blower chest, and the exhaust valves are rubber strip flap valves mounted over holes in the chest bottom board.

The crankshaft runs along the bottom of the organ frame and is fabricated from 15mm copper water pipe and fitting; it uses 90 degree joints to provide the crank throws. At the driven end, it is coupled to the shaft of a Ford Transit wiper motor which is ideal for this application, revolving at about 1 rev. per second with the slow speed connection in use. This has the advantage that the whole instrument can be operated from a 12v DC supply so could be used away from the mains supply if necessary. At the far end, the crank is supported in a bearing made from a 15mm compression fitting. The crank rods are made of M8 studding. At the rod ends, 15mm pipe clamps which screw directly onto the studding are used. M8 nuts lock the clamps in postion on the studding. The feeder bottom board is fitted with a pair of aluminium angle brackets, with an M8 bolt running between them. Where it contacts the 'little end' the bolt is sheathed with rubber fuel pipe, of approximate 15mm outside diameter. All the working surfaces are well greased.

Rather than run the blower at constant speed and use a dump valve or other contrivance to maintain the reservoir lift constant, I decided to electronically sense its postion and use this to control the blower motor. As an electronics engineer, this was relatively trivial for me to do. As stated above, a potentiometer is fitted to the reservoir assembly and provides a feedback voltage proportional to the degree of lift of the reservoir. This is connected to an operational amplifier with a high current output stage, which in turn feeds the motor. The feedback is compared with a preset reference voltage, and as it approaches the reference, the motor slows down, until the two voltages are the same when it stops completely. As wind is used by the organ, the motor will run at a controlled variable speed in order to try and maintain the reservoir full. The gain of the servo amplifier is set so that if it falls below 66% full then the motor will be running at full speed. A mechanical cut-out is provided which will trip the supply off if the reservoir opens beyond its design limit, so that the feeders and crank arrangement will be protected in the event of failure of the motor control servo system.

The pantograph springing idea was suggested by Raphi Giangiulio who uses it in his organ, (and also the anti-roll bar idea- thanks, Raphi!). It is a method by which an ordinary spring is compensated mechanically to provide a load which does not vary with deflection. It provides a remarkably constant wind pressure: its measured performance is that the wind pressure is almost constant until the reservoir is nearly fully closed, when it increases by about 1/4 in. wg. To prove the point, the musical pitch of the middle C pipe was measured using a digital chromatic tuner. The organ was fully pumped up, i.e. with the reservoir fully open, and then the blower was switched off and the reservoir allowed to run down with the pipe speaking, which took about 30 seconds. The variation in pitch throughout the whole range of reservoir travel was just 10 cents. In practice this amount of variation will not be seen because the servo will maintain the reservoir position. I don't know how this performance would compare with professional organ practice, but it seems pretty good to me.

The organ will use direct electric action. I chose this for a number of reasons, but the main one is that it allows very easy interfacing to a MIDI controller. I don't play keyboards to any standard of proficiency, so I wanted to be able to play the instrument from MIDI files. However, playing from a manual is still possible and will be used during subsequent development. A description of how the manual will work is given later, but it does not rely on the MIDI controller being present, because this will be virtually the last thing I will build. A secondary advantage of DE action is that it is relatively easy, using diode matrices (apologies to readers who are not versed in electronics!) to give different pitches and even mutations using a single rank of pipes. So I could play, from my manual, certainly 8ft and 4ft perhaps 2ft if the organ were expanded upwards, 2 2/3 ft, etc. I don't think this functionality will be built (or programmed) into the MIDI controller; it is more in the domain of a software application on the host computer.

The action magnets as they called in organ building parlance are usually some sort of solenoid and whilst the proper thing can be obtained, they're expensive. My experiments have so far have been around autoelectric (car) relays modified so that their armatures operate the pallets through stiff wire linkages. Even when run on only 9v, the relay was found to have adequate power to lift a pallet against wind pressure over a 3/8 in. diameter hole, the sort of size of the toe holes I will use. The pallets are a small wooden block (about 1/2 in. square) with a layer of felt glued on, and over this a layer of soft rubber sheet. These seat well and hardly leak wind at all. Electrically, the magnets are connected to a common return rail on one side, and to individual pins on D-type connectors mounted through the wind chest top board at the other. The D connectors are sealed on the solder side with a good layer of silicone rubber to prevent wind leakage through the connector. In all a minimum of 50 connections is needed but to allow for possible expansion more are available.

The action will be controlled by a PIC microcontroller, which will receive MIDI messages from a standard port. It controls the magnets via a shift register and driver arrangement that in effect is a 48-bit parallel port with some current drive capability.

The first thing I did, in Summer 2003, was to build some wooden pipes and see how they performed. I tried a pipe made like a fair organ violin, and open and stopped flues. A great deal research, experimenting and thought led to my final 'recipe' which produces a beautiful velvety flute tone on quite low wind pressures. It is a stopped pipe, which has the advantage that it is relatively short so saves space, and also it is quieter than an open pipe, so more suited to a domestic environment. I am a bit defeated by the idea of what to call it, but broadly it belongs to the gedeckt/stopped diapason family. With the cut-up used it speaks quickly with virually no 'chiff' and it is economical with wind. Four pipes covering four octaves were made to 'the recipe' to test it across a good range of pitch.

The experimental pipes were constructed from Balsa wood available in different thicknesses from model aircraft shops. The diagram shows the internal layout and approximate dimensions, note the various bevelling and chamfering of the lips and the block; I found this to have a profound effect on the tone, and the efficiency of the pipe. I have not found it necessary so far to fit ears, but I haven't made the biggest ones yet. I found that the stoppers are critically important and they must effect a total seal. I experimented with different ways to make them, but found that the best results (and the cheapest) were obtained using a block of closed-cell foam plastic, cut slightly larger than the inside dimension of the pipe. This fits snugly and seals well even if the pipe isn't made completely 'square'. A wooden stopper handle can be glued to it, and you don't see the foam from the outside so it looks quite good too. The foam slides fairly easily during tuning, but you have to allow for the fact that the foam deforms as you pull the handle in or out.

The current status is:

The Blowing System is essentially good but the crankshaft needs replacing with one with less throw. There is a tendency for the wedge feeders to overtravel and the wind back pressure pushes the tucks in the leather out the wrong way. As the throw given by the copper pipe one cannot easily be reduced, the next one will have to be made of wooden dowel.

The Action needs solenoids. I know I was on about modifying relays, and this is still a possiblity if I can get enough of the right sort at the right price. I thought of trawling car scrapyards for them. But maybe we can dig up some surplus ones somewhere. The search is on!

MIDI Controller Not even thought about it in any detail yet.

Web page background from about.com: desktop publishing

Last Modified 17/2/2008