Specific sub-systems - The Winding System
Here is a cross sectional diagram of a typical pipe organ pressure regulator, one design that has been used in a number of leading pipe organs. The operating principle is as follows. Refer to figure five.
Air from the blower enters through the bottom inlet into the internal valve box section, after which it passes through the valve and enters the remaining space inside the regulator, which is referred to as the regulated pressure region. The top of the regulator is made in the form of a modified bellows as shown, so that the top can rise a certain amount, which gives the pressure regulator a variable internal volume. This is how the device "senses" the instantaneous pressure. Unlike a typical bellows, however, the top remains parallel to the rest of the unit in all positions, that is, it is not hinged at one side.
At the same time as the air pressure pushes the top up, powerful springs pull the top down. Therefore, the top will not rise until the pressure in the regulator is enough to counterbalance the downward force provided by the springs. As the top rises, it pulls up the internal valve disc which closes off the circular opening in the top of the valve box section. In the best designs, the valve disc has an exponential profile. As more air enters the top section, it raises the movable top until the valve will eventually close off the circular opening. When playing makes a demand for air, the top descends slightly because there is a very minute instantaneous lowering of the pressure acting on the movable top, however it is worth mentioning that this pressure drop is extremely small, so that as far as the pipes are concerned, it is negligible. As the top descends, the valve disc, which is attached by means of the rod, drops slightly also and begins to allow a little higher pressure air from the blower to enter the regulated region above the valve disc.
In a well designed pressure regulator, the valve characteristic should be such that when the top of the regulator begins to descend, the valve opens very little, but with just a little more descending of the top, it opens considerably more. This is the reason for the exponential valve disc profile. It is also interesting to note that these regulators will not work if the airflow should stop. That is, if the regulator supplied an entirely airtight or dead-ended system, they would function erratically. In order to work properly, there always has to be a slight amount of airflow. Fortunately, there is always minute leakage in the windchests on which the pipes stand, so that a minimum airflow is maintained at all times. As long as this minute airflow exists, then these bellows type pressure regulators will control the output pressure very accurately. In actual operation, the amount of leakage is just enough to keep the valve from sealing shut entirely, but that is actually a very small amount of air.
If a pipe organ winding system is well designed, the normal motion of the regulator top should be very small, about 1/2" at the most between no playing and full organ. If the regulator tops must descend through several inches of travel, then the output pressure will sag as the springs pull with less force when they are stretched less. If for some reason, the regulator tops should bottom out, they will then be out of control and will not regulate the pressure at all. The usual causes for this problem are major air leaks or the use of inadequate blowers. For proper pressure control, the springs must also be pre-loaded, that is, even when the blower is not running and the regulator top is at its lowest position, the springs still must be under considerable tension. Because springs have a definite tension vs deformation characteristic, the best results are obtained with relatively small changes in spring deformation, and when operating at the higher end of the most linear region of the spring characteristic curve.
In a typical regulator of this type, the total vertical rise of the top is usually not over 6" and sometimes only 3 or 4 inches. The range of top position during normal playing of the instrument should never exceed 1/2 inch. In poorly designed winding systems where the regulator tops must move through many inches of travel between zero and full organ, you will encounter noticeable pressure sagging with a subsequent slight detuning of the pipes and a slight drop in volume when using a lot of stops and/or playing big multi-note chords, especially when the couplers are turned on also. Pressure sagging constitutes very bad design and should never happen. This, by the way, can also happen if the blower is too small for the instrument, so that the distribution pressure at full organ is not any higher than the desired regulated pressure, or if there are many leaks in the wind distribution piping between the blower and the pressure regulators.
Now, if you think about this, you might say to yourself, "Self, if springs change their force with a change in deformation, why not use weights for loading instead of springs on the tops of organ pressure regulators?"
This was indeed the original practice, and it should seemingly make the pressure much more constant and allow a much greater range of regulator top movement. However....(funny how in so many situations there's always a however), weights have inertia and momentum. If you load the top of a pipe organ pressure regulator with weights instead of springs, when you play sudden big chords, there will be an instantaneous significant pressure drop because of the inertia of the weights, and the top will not respond and subsequently open the valve fast enough to keep up with the sudden wind demand. But then, when once moving, the weights will gain momentum and cause the top to overshoot slightly, giving you a sudden subsequent increase in pressure. The result? Severe hiccupping in the pipe speech.
Spring-loaded pressure regulator tops respond virtually instantaneously to wind demand changes and allow the regulator to track the demand much more closely and suppress these instantaneous pressure transients right as they start to form before they can really develop. As long as you can keep the regulator top's total positional excursion to within a half inch, and supply the counter balancing force with pre-loaded springs and not weights, you'll get the best of both worlds; instant pressure transient suppression and really close pressure regulation.