Design of the Goulding & Wood slider chest action
-Article by Tom Wood-
On various occasions we have been asked to describe our unique, electro-pneumatic slider and pallet windchest. Particular emphasis is given to solving the age-old problem of devising a remote key-action system for this chest that is musically responsive over a wide range of wind pressures. We are often asked why would we want to consider such an action with the reliability of the modern day pitman chest system for remote action organs and with the technical advancements that have been made in mechanical key-action instruments. The answer to this question actually goes back beyond the formation of Goulding and Wood Inc., twelve years ago, to the late 1950's when John Goulding and I were both associated with the E.H. Holloway Corporation.
In 1960 John Goulding built the first version of this electro-pneumatic windchest system while with the Holloway Corporation. It was used in a small, one-manual chapel organ and this original chest is still in operation today. The first major organ incorporating these chests was built in 1961. Twenty years later Goulding & Wood Inc. upgraded this organ with solid state switching and extensive tonal modifications. The only changes made to the twenty-year-old slider chests were the new toe boards needed for the new stops. Over the years Goulding and Wood Inc. has made a series of modifications and improvements to the original design, but the basic principles remain intact.
To understand the hows and wherefores by which we developed this reliable and responsive remote-action system for the slider/pallet windchest, one needs to know both the process of development as well as a short history of the evolution of remote-action windchests. It is not our intention to imply that the use of this action is the only valid way to build a fine pipe organ. In fact our adopted procedures exclude us from being competitive in certain situations. We recognize that quality organs have been and are being built that utilize time-proven, electro-pneumatic windchest systems as well as slider-chest instruments utilizing mechanical control actions.
The one aspect of organ construction that is the most unclear in many people's minds is the ability to differentiate between control actions (the energy transfer system from console/key-desk to the windchest), and windchest actions (the mechanism within a windchest that actually controls the admission of air into the organ pipes). There is no physical reason why the key action channels of a pitman chest could not be controlled by mechanical action. In George Ashdown Audsley's famous treatise, The Art of Organ-Building, there is a wonderful drawing of Roosevelt's, pneumatic style, ventil system windchest. This was a highly acclaimed action at the turn of the century, and continued to be the accepted action system in England long after most American firms had adopted the pitman system. When looking at this drawing, I still remember the initial shock I had when I realized that the primary valves were connected to trackers! Early advertisements by Roosevelt offered the buyer the choice of either tracker action or tubular- pneumatic action with his "pneumatic" ventil windchest.
The pitman chest is an outgrowth of the ventil system. It is labeled an "universal" windchest since the admission of air to the pipes is not dependent upon switching on and off the actual air supply that blows the pipes. This air is constantly under the pipes with the "work" of opening the pipe valves relegated to a set of low-energy control channels; namely the key action channels, and the stop action channels. Each pipe valve is connected to both its proper note action and proper stop action channel. Both channels have to be exhausted at the same time for the pipe to play, i.e., the right stop has to be drawn and the right note has to be played for the right pipe to sound.
Both the slider chest and the ventil chest require that the chest be barred or segmented into channels that are air-tight since the operation of the pipes depends upon controlling the actual air that blows the pipes. The slider chest is segmented into note channels with the key action controlling the admission of air to a pipe. The ventil chest is segmented into stop channels with the stop action controlling the air flow to any one stop (with the key action then controlling the proper pipe valve).
The inherent disadvantage of the slider chest is that the action that needs to be the most responsive (the key action) must also do the actual work of admitting the wind to the pipes. These opposing conditions have unnerved many an organ builder over the centuries. Large tonal dispositions, high wind pressures, and remote consoles - all are detrimental to maintaining a responsive key action system on slider chests. For better or worse, the trend in organ building that took place near the beginning of this century emphasized all of these detriments. The ventil windchest was developed to overcome these very problems by assigning the energy control to the stop action system which does not need to be as responsive as the key action. And this, in turn, paved the way for the modern pitman systems which do not have to control any of the actual air flow that blows the pipes.
In spite of these seemingly bad points about the slider chest, there are three positive attributes to the slider chest that are indisputable - 1) Simplicity, 2) Superior air delivery to the pipework, 3) Longevity. The use of a simple, moving slide to control the air into the pipes along with the major advantage of needing only one valve system per note rather than a valve for every pipe makes the slider chest the least complex by far of any windchest system. With only one valve per note, all pipes of the same note speak together and with the same attack characteristics. The "common key channel" allows the flow of air to smooth out after it passes the valve which, in turn, creates a steady and undistorted wind flow into the pipes.
The chest does need to be built to exacting tolerances as the stop slide must move freely while still blocking off the air flow to the pipe, and the builder is constrained to placing pipes according to the matrix (or grid) of the chest, except for those large pipes that are to be offset into cases or elsewhere away from the main chest. But because of the simplicity of the action, pipework can be placed more compactly on a slider chest.
Convinced that the good attributes of the slider-chest out-weigh the inherent action paradox, and likewise convinced that not all the tonal developments of the twentieth century were evil - even if not suited for mechanical action control - we set out to fool Mother Nature and her laws of physics.
Most of us know what it is like to play on a late nineteenth century organ that has had its mechanical control action replaced by either external electro-pneumatic or sometimes electro- magnetic "pull-down" actions. You can forget all about that workshop you attended where Marie-Claire Alain taught you all the nuances of French ornamentation. You will be lucky to get through Come to Jesus in whole notes. External, remote controlled pull down systems simply have too much mass in them to be responsive. And, I think most of us have encountered mechanical action systems that, for one reason or another, cannot be defined as sensitive and responsive.
Whether we as organists or builders want to "fess up" or not, the same laws of physics apply to all energy transfer systems whether they be electro-magnetic, electro-pneumatic, or mechanical. Energy is defined as force over distance. Distance is the product of speed and time, and force is the product of mass and acceleration. At some point, builders have to decide what it's going to be - a high force action for quick response, or a light force action for sluggish response - or else devise a power assist system to solve the dilemma. (Most of us know that Cavaille-Coll adopted the barker lever assist to aid his mechanical action control.)
Reducing the effective mass as much as possible is always a good starting place. As others have discovered, placing the pull-down action within the windchest removes the mass of the linkages. There are really only two types of power assists available, an electro-magnetic system, or an electro-pneumatic system. We, along with a few other firms, experimented with electric magnets. There is merit in the simplicity of the electromagnetic system, but there is an inherent property of the magnetic valve that we find troublesome. The greatest force is applied just as the circuit is engaged. As the valve opens, the magnet armature is pulled across the pole piece. As this magnetic circuit closes, the force diminishes. This is to say that there is not a constant force applied over the range of travel.
This fact also implies that the speed of the pallet is not constant throughout its range of travel. If the circuit is not powerful enough, the pallet will not become fully opened (particularly when many stops are on) and a condition known as "pallet fade" is encountered. When the circuit is powerful enough to prevent fading, then the pallet will open very quickly. The tonal result is akin to playing mechanical action very hard and fast all the time with the resulting maximum emphasis on pipe articulation (chiff). To be candid, we find it interesting that some present day builders of mechanical action make use of electromagnetic pull-downs for the coupled manuals. It appears to us that the tonal result obtained from the electric action is just the opposite of what the merits of mechanical action are deigned to be—unless, of course, it is assumed that when the manuals are coupled, the organist will be pounding the dickens out of the keyboards anyway! The other disadvantage to the electromagnetic approach is that as the number of stops or the wind pressure is increased, the power of the magnets can be overridden. This places a constraint on both size and wind pressure of any particular windchest.
Thus, we concluded that the electromagnetic procedure has two major constraints: 1) limited power, 2) bad pipe speech characteristics. The alternative was an electro-pneumatic system where 1) there is always a constant force over travel (and, therefore, constant speed) and 2) the power is not affected by wind pressure since the same wind that blows the pipes would also operate the action. The two major obstacles that needed to be overcome were 1) reducing the effective or dynamic mass of the system to insure acceptable response time (repetition rate) and 2) provide a wind regulation system that would allow the pneumatics to be installed within the pallet box without the operation of the action affecting the wind to the pipes.
The diagram shows our solution for obtaining a responsive key-note action for the slider chest. In many ways, this action can be defined as an internal, electro-pneumatic barker lever.
To reduce the effective mass of the system, the main motor pneumatic that opens the pallet valve is made half as short as the pallet, but twice as wide. This retains the same square area (thus the same power) but reduces the inertia of the system. In addition, the arm that opens the valve is one-third the length of the pneumatic. This provides a three to one power advantage. The reduction in inertia allows the pneumatic to travel twice the distance as the pallet without hindering response time. There is a pallet rail that prevents the pallets from opening farther than necessary. Since the main pneumatics are wider than the pallets, it is necessary to alternate the pallets from side to side to create enough room for these pneumatics. Thus, the pallet box runs the entire width of the chest. The main pneumatics are controlled by a conventional, electro-pneumatic primary action like those that have been around for some seventy years.
Satisfactory winding has been a topic of concern for organbuilders throughout the ages. Flexible winding, swimmer regulators, and the manner of wind flow within a wind chest, particularly the slider chest, have been the subjects of several lectures given by members of the American Institute of Organbuilders (AIO). While our tonal preference does not lend itself well to flexible winding, we, like several other builders, have been concerned with the tendency of swimmer regulators to over respond to pressure changes which create "bumps" in the wind supply to the pipes. Yet, with action pneumatics inside the pallet box, the need for a regulator that responds quickly to wind demands is paramount. To smooth out the aberrations caused by a normal swimmer plate, we developed a unique, passive concussion bellows that is built into the actual plate. These plates with pantograph springs are mounted between the opposing primary boards. The result is most satisfying.
Another variation in our construction of the slider chest that varies from standard, modern day practice is the omission of slider seals. We did not like the idea of adding components to the chest that in number has to equal the number of pipes on the chest. We choose to build the table of the chest and the toe boards from special, high-density medite. This material has no grain which makes it dimensionally stable. Likewise, it has a very hard and slick surface. This allows us to maintain so close a tolerance for the stop action slide that slide seals are unnecessary. The small amount of air that does "escape" around the slides helps to smooth and stabilize air flow within the key channel. Thus, this procedure provides two benefits. Conventional, key-channel relief openings, opposite the pallet valves, are also incorporated into this windchest.
With this construction, we are able to operate the stop action slides pneumatically as well. This allows for a more efficient, less expensive, and above all - a more quiet stop action. Another unique feature is that each windchest actually consists of two, completely independent half chests. The stop action is placed between the two halves with a passage way over them. This provides two distinct advantages: 1) Each stop slide is only one-half the length of the chest - this holds friction on the slide to a minimum; 2) Since each half has its own isolated winding system, wind regulation is optimized. This is a particular advantage with a chromatic chest layout where the bass pipes are separated from the treble pipes.
It is our tonal conviction that manual 16' stops can be unified to serve as light, pedal 16' stops without noticeable detriment to pipe speech and with a high degree of economic sense. We always use an electro-pneumatic pouch action in unit chests in order to maintain similar pipe speech characteristics.
Thus it is that Goulding & Wood, Inc. came to design a slider and pallet windchest system specifically for remote operation. Solid-state switching and combination action systems have revolutionized the nature and reliability of electric action consoles. By incorporating these developments in electrical signal control, we feel that our system is one of the very few that can match the response time of a pitman chest while maintaining all the inherent advantages of the slider chest and its common note channel. This system is not affected by chest pressures. The normal range has been from 2-1/2" to 5" (60mm to 130mm) water gauge. Much of this article is devoted to the historical evolution of wind chest systems, but it is from these facts that we came to make the alterations and adaptations to the conventional slider/pallet windchest which in turn has produced this unique and highly successful action system capable of supporting a variety of tonal styles.
This article appeared in The Diapason (September 1991).
We wish to thank the editor of The Diapason, Mr. Jerome Butera,
for his permission to reprint the article at this web site.