Secondary Pouch Board

As previously discussed, the valve chests of many models of player pianos are a sort of “box” made up of two or more long boards. The secondary chest in my piano is typical in that it is comprised of a valve board and a pouch board. The two components are necessary as a valve cannot operate without its pouch.

pouch board, shown in foreground
detail of secondary pouch board with lifter discs

The pouches are made of very thin leather (appropriately called pouch leather), which is cut into circles of a given diameter and glued in a “dished” position into their respective wells, which are simply precision holes drilled into the pouch board.

As with other components of the action, the decision when rebuilding is to replace or refurbish.
And as with other operations in player piano work, opinions vary.

Some say that replacing with new is the only way to guarantee long term reliable and uniform performance. Others say that new leather is of inferior quality compared to the material of yesteryear, and that if the old leather is still in usable condition, then it can be resealed for continued use.

If you decide to replace, pouch leather can still be sourced from an organ supply company.
Even new leather should be sealed, according to some advocates. More on that shortly.

To replace the pouches, clean off all traces of old leather and adhesive from the board, and check the board to ensure it is otherwise undamaged, and still true.

While the board is “de-pouched” is also a great opportunity to renew the sealant in the channels from the edge of the board to the pouch wells. Thick shellac or thinned PVCE glue can work well for this task. Bruce Newman recommends using cheap disposable pipe cleaners, of the appropriate diameter, for this task. It is important for each channel to be independently sealed, so as not to “bleed over” to its neighbour and risk ciphering notes.

The pouches can be glued down using a pouch setter, and hot hide glue is once again the glue of choice. Note that some pouches (like mine, as pictured above) have cardboard or fibre discs glued in the center. If these need to be replaced, use only the minimum amount of glue necessary to securely attach them, and glue the discs before the sealing takes place!

It is important to note that the two important characteristics of pouch leather are airtightness and suppleness. Ideally this membrane would be infinitely strong, airtight and flexible, but in reality there is a balance to be struck in the physical properties of the leather. In other words, in seeking to make the leather leak proof, we must not diminish the suppleness by introducing stiffness with the application of a sealant.

Traditional solutions for this are to use rubber cement diluted to 50% with thinner, or to use neatsfoot oil or a derivative product. More recently some folks like John Tuttle have been experimenting with silicone diluted in a solvent carrier, which you can read about here.

Once the pouches have been dealt with, you are likely looking at replacing the perimeter gaskets as well, or at least you might as well when you are at it. With your newly spiffed up pouch board, you are now ready to test the secondary chest!

Secondary valves: regulation

As with previous topics, I will first state a proviso that the exact details and order of operations for any given task are subject to the type of player action which you find yourself working on.

The following applies to one of the most common type of original player actions: the Standard / Auto Deluxe double valve action.

Regulating the secondary valves consists of just two steps, but they are both as time consuming as they are critical to the proper performance of the action.

If you have read the last couple of posts (which I humbly recommend you do), you will have a basic idea of the functionality of the valves.

To ensure the valves perform accurately and consistently, they must be regulated.

1) Begin with “gapping” or traveling the valves. The amount of travel allowed to each valve will ideally be set to optimize its performance. For the type of action at hand, there is general agreement that this would be about 0.040″. To repeat, this will vary depending on exact valve type, but for my piano it is a good baseline.

Once the rebuilt valves have been set back in their chambers and secured, the gap can be set by adjusting the position of the outer facing on the stem, thereby altering the distance between the two facings, and by extension the distance the valve can travel.

My variant of action has a screw-on outer facing, so this feature (theoretically) makes the gapping adjustment much more efficient. By gripping the stem below the plate and pushing up firmly (but not too much force), and then twisting the stem, the position of the facing is changed.
Twisting the stem clockwise (seen from the bottom of the stem) turns the facing down and increases the gap, twisting counter clockwise has the opposite effect.

For the more common variant of shimming the outer facing, this unfortunately means constantly removing and replacing the valve until the travel is perfect, for each of the 88 valves. The advantage of this type is that the regulation will last a long time, which is a good thing!
Bruce Newman demonstrates here, once again.

A dial indicator gauge is a very handy tool for this task; however I found that I got good results with this simple Jaras tool, whose normal duty is to level sharp keys. I love finding more than one use for a tool! Whatever method you use, the travel should be consistent to within a couple of thousanths of an inch, at most.

Secondary valve regulating tool, alternative approach

Once this very important step has been completed, double and triple checked, secure your valves in place for the foreseeable future (with sealant if applicable), screw the buttons on to the bottom of their stems, and proceed to step two.

2) Set the button to pouch clearance. The buttons must never rest on the pouches, otherwise they risk staying slightly open and causing leaks or ciphers. It is recommended that there be 1/16″ (1.6 mm) distance left between the middle of a dished pouch and the button. This allows for seasonal changes when the pouch may expand or shrink slightly. In Some locations (like Northeast US and Eastern Canada) where there can be significant swings in relative humidity between summer and winter, some technicians advocate 3/32″ (2.4 mm) gap. A wider gap is definitely safer in avoiding the possibility of valves being stuck “on”, but too much will cut into reliable repetition.

Grip the stem with narrow smooth pliers or forceps (to avoid marring the stem), and turn each button until you get the right position.
One helpful way to do this is to turn the valve board upside down, and have it propped up off the work surface by several inches, to have just enough room to actuate the valves with your hand. Have your gauge (e.g. ruler) set up to simulate the imaginary line of clearance you are targeting.
If you slide your gauge along the edge of the valve board, the buttons should all just kiss the bottom of the ruler, without catching it or impeding it.

Here is an image of the process featuring a custom tool made by technician Bill Maguire.
It allows the straight edge to be quickly and securely raised to the correct height to regulate the buttons with accuracy.

Once you have done a couple off passes of regulation and feel confident in your work, John Tuttle has a method to check your results, here

If you have taken the time to do both these adjustments well, then the next phase of testing and troubleshooting should go smoothly.
If not, better to find out sooner than later!

Secondary Valve Chest: Rebuilding Valves

Having taken apart all the valves, done an evaluation, ordered and received new materials, it’s time to start rebuilding.

In all likelihood your valves are pretty old and used up, therefore a wholesale replacement of all ‘soft” parts is necessary.

The construction style and materials information of my valves were discussed in previous posts, so on to the subject at hand.

In my case, I ordered the following new materials: fibre discs (2 per valve), fibre valve stem guides (1 per valve) new leather facings (2 per), new retaining leather washers for the outside facing (1 per) and new sealing leather punchings for the inside facing (1 per).
I also needed to order custom blotter paper gaskets for the valve plates (1 each).
It is good to have extra amounts to spare, of all the above listed parts.
I was fortunate to be able to reuse all the metal parts including the stem, retaining collars, metal seat plates and plate screws, as well as the wood stem buttons.

The rebuilding process will vary with valve type; here is a list of operations which had to be carried out to renew my own secondary valves:

• Clean and polish stems
• Clean plates (soak in denatured alcohol to soften old shellac sealant)
• Lap plate seats (sandpaper on plate glass, successive finer grits 320, 400, 600)
• Polish + wipe plates
• Seal plates (spray lacquer specially formulated for brass)
• Shellac gaskets onto plates (if applicable)
• Lube plates (spray lubricant)
• Glue / assemble valve facings to replicate originals
• Press or screw valves back together, roughing in gap distance
• Add stem guide and plate, place in valve chambers for regulation

These 10 or so operations seem basic on the face of it, however you will find that they are quite time consuming once everything is said done, it will keep you busy for a while!

plates drying after finish spray
putting blotter gaskets on valve plates
glued outer facings
set up for mounting inside facings
press fitting collars and facings onto valve stem

Not only that, but this is very exacting work. All valves must be rebuilt to a consistently good standard, so that they perform identically and reliably. There are many little details to look out for; I will mention more at length later, but a critical one is that both facings on the valve stem have enough “wobble” to seat properly on their respective seats.
If the valves are too rigid, they will leak badly.
Too loose, they will not operate crisply.

Once again I will rely on previously recorded material from others; Bruce Newman has a good video on assembling and regulating valves.

Following my checklist above, once you have all the individual parts ready and the facings are on the stems, add the finished plates and stem guides, then put them all back in their chambers.

installing newly releathered valves in chamber

The observant reader will notice that I did not mention the buttons: we are leaving those off for now, for reasons to be made clear shortly!

On to regulation!

Secondary Valve Chest: Teardown

There are any number of ways I could proceed in further adventures of rebuilding this player piano action, but since we’ve been discussing valves let’s continue with the secondary valve chest.

When you are looking at the front of the stack, the secondary valve chest is the bottom compartment. More specifically what you are seeing is the back of the pouch board.

secondary valve chest overview
Secondary valve chest as seen from outside of stack

Assuming you have already removed the head, the shelf, the primary valve chest and the “L” board connector, you are now in a position to open up the secondary valve chest.

It is fairly self evident as there are many screws to undo, to peel off the pouch board.
Depending on the gasket material and potential use of adhesives, it should come off without too much resistance.

Now what you will see is something like the following:

valve board and pouch board
opening secondary valve chest: pouch board and valve board

Set the pouch board aside for the time being, I will discuss in a future post. Next, it is necessary to remove the valve board from the pneumatic decks assembly, by removing all the screws of the type that are marked with a blue asterisk. Note that there is a raised screw marked in this photo with a red asterisk, which you probably want to not mess with at this time. There may be 2 or 3 of this type.
I will take this opportunity to remind you about the value of screw maps to ensure the screws go back where they came from, when the time comes.

Once you’ve got the valve board off, it’s time to remove each individual valve – 88 of them!
Depending on how they are secured to the valve board, the technique may need variation.
In my case, they were simply screwed and had a blotter paper gasket, so just removing the screws was all it took.
In other variations of action, it is very prudent to exercise caution here: if the valves have plates as pictured, and these plates are secured or sealed with shellac or other adhesive, even if you remove the screws, the plates will resist removal and if you force them without dealing with the adhesive, you will deform the plates and ruin them.
The common approach is to use a soldering iron with the tip touching the plate, until the heat transfers to the adhesive and softens it.

Another thing to note is that there may be two sizes of screws holding down the plates, the ones holding the stem guide may be slightly longer. It will be helpful later to separate them by length as they are removed.

Once that’s all done, now what you have is this:

valves removed
removing secondary valves from valve board

Finally, you must disassemble each individual valve into its constituent parts.

I kept track of everything by putting valve parts in ice cube trays: all components of a given valve in a compartment. The tray compartments were labelled with a strip of tape on each side which was numbered accordingly. You need about six trays for the whole collection, but hey they’re cheap!

The first step of disassembly is usually to remove the wooden buttons from the bottom of each stem.
If they are on there firmly, you will need to grip the stem with forceps or another small but strong and smooth tool to avoid marring the stem.
Once the button is off, the stem guide and metal seat plate will fall off as well.

For these stem type valves, the valve facings are usually held onto the stem by small press-fit collars. These collars, if removed carefully, may be reused, so let’s try for that. There are a couple of ways to do this; I used the following set up:

removing collar
removing collar from valve stem

One hand gripping the stem with a pair of gripping pliers and a non-marring but firm substance (a slotted piece of nylon cutting board); the other hand with a pair of parallel pliers gripping the collar. Parallel pliers are much preferred to grip the collar firmly and evenly at the tip of the pliers. Twist and pull the collar off, straight out to avoid bending the stem.
You can also press the stem off the collar, as done by Bruce Newman.

Once you’ve done all that, it’s time to decide what needs to be replaced, and what you can salvage, then start the rebuilding of the valves!

valves dismantled
secondary valves during disassembly

Valves: Materials

Having taken a peek at the theory behind valves, it’s important to get practical with the materials.

In a Standard / Auto DeLuxe etc double valve system, the two principal valve types found in the stack are the primary valves – which are an outside valve – and the secondary valves, which are an inside valve. The terminology surrounding the player piano components has been the subject of debate; there is a mostly accepted convention about nomenclature, but better to not get bogged down too much for the purposes of this discussion.

The secondary valves are a stem-type valve, which varying methods of how the facings attach to the stem. For the purposes of this post, I am confining my remarks to the facings.

In this system both styles of valves have two facings; in an outside valve both facings seal against a wood seat, in an inside valve the upper facing seals against wood but the lower seals against a metal plate.

Leather against smooth wood is generally a reliable seal, all other factors considered.

Leather against a metal ring is far less forgiving, particularly if the leather is poor quality (uneven, too fuzzy, or stiff, or porous), or if the valve seat area of the metal plate is corroded or not perfectly flat. There are other reasons (beyond choice of leather) as to why the secondary valves require extensive exacting work to get them to perform well. More on that later, perhaps.

At the time of this writing, high quality leather, the kind required for inside lower valves seats (meeting metal plates) is difficult to source as it no longer widely produced and therefore not readily available, even from trade supply houses. Conventional wisdom dictates that this material be a split calf leather with a smooth nap, to prevent lateral leakage.

Old valve leather vs. new

Some rebuilders have resorted to using thin silicone or neoprene material for lower secondary facings. Like other non-traditional materials it is the subject of controversy, however if it can be made to work in a reliable and durable way – the results are what matter.

For the other leather in primary and upper secondary facings, I have received numerous recommendations for the Columbia Organ Leathers company. I am using it in my stack and find it fit for purpose. I know that others use OSI for their needs. These companies are also able to furnish good leather for other types of valves (not yet discussed) such as flap valves and pallet valves.

Wherever you choose to obtain your materials, make sure to inform yourself about the supplier and their product. Using inferior materials is a fool’s errand as it will waste time and money trying in vain to get good results.

As the prospect of restoring valves strikes terror into the hearts of some inexperienced restorers, they may be tempted to skip over doing it, if the older leather facings “look good”. Older restorers sometimes talk about having gotten away with this omission, when doing a restoration job on a player action when they were starting out.

The difference is that when senior rebuilders worked on players in their youth (e.g. the 1950s or 60s), the pianos in question may have been 30-40 years old. The leather in the valves and gaskets may still have had some life left in it at that time.
After 90 or 100 years, it is a safe bet that all original leather and cloth in the player action need replacing, to function as intended.

The bottom line: you can’t have well performing valves without good materials!

Valves: Theory

Although I don’t necessarily want to get too far into the weeds on this topic, I do believe it behooves me to discuss valves, given their importance in a player piano mechanical action.

The key to understanding player pianos is fundamentally a question of comprehending valves: air pressure and flow. Conversely if you don’t understand how valves work – in theory and practice – you probably won’t successfully figure out how to rebuild and troubleshoot your project. A helpful analogy may be to think of the airflow as an electrical signal, and the valves as switches.

Let’s focus on some theoretical basics for this post.
There are a number of different types of valves; the most common are inside and outside valves. Outside valves normally function as a “primary” role, while the inside valves are generally “secondary”. This is how they are assigned in a Standard style double valve player action.

Both styles of valves operate in a valve chamber, which needs a constant supply of suction from the pump to operate. Each valve has a corresponding pouch underneath it, which will activate the valve when the pressure builds enough to inflate the pouch. Every valve will have an input and an output as well. Note the following image, and how the output of the primary becomes the input of the secondary:

valve 1 annotated complete
Typical double valve action, taken from “Rebuilding the Player Piano” (Larry Givens, Vestal Press, 1963)

It is important to bear in mind too that the input signal doesn’t directly come into contact with the valve. The signal determines whether the pouch (below the valve) is up or down, which in turn determines if the valve is opened or closed.

Primary valves receive a small signal (usually via tracker bar), and augment it. They are signal amplifiers.

Secondary valves also multiply the signal, but the signal is reversed. This is an important note. In other words, if the input to the secondary (or any inside) valve is atmosphere, the output is suction. If the valve receives suction (i.e. the valve is closed) its output will be atmosphere.

Here is another line drawing of the signal path in a Standard double valve style action.

double valve system annotated v3.png

There are three discrete stages which occur almost simultaneously.
When a perforation in the music roll passes over a hole in the tracker bar, this admits a small amount of atmosphere and initiates stage 1. This small amount of atmosphere allows the pouch to rise and activates the primary valve. The now open primary valve allows a larger signal of air to enter the connecting channel to the secondary pouch, which is then raised and opens the secondary valve in the same manner (stage 2). The activation of the secondary connects the pneumatic to the valve chamber with its reduced level of atmosphere; in other words the valve outputs a signal of suction. Stage 3 occurs when the atmosphere has been reduced in the pneumatic the outside pressure of the atmosphere forces the pneumatic to collapse. The pneumatic (in this design) collapses upward, actuating the piano wippen and playing a note.

The cycle ends when the paper again covers the tracker bar hole; with no more atmosphere signal the suction in the primary valve chamber quickly equalizes the pressure on both sides of the pouch (aided through a small “bleed” passage next to the pouch). This in turn shuts off the signal to the secondary valve and then to the pneumatic, which equalizes once more with the surrounding atmosphere, and reopens.
This whole cycle may take a fraction of a second.

The valve chambers for the primary and secondary valves are under constant suction (as generated by the pump); in the diagram just above the blue arrows indicate the path of the suction back to the pump. This is to differentiate from the “signal” paths which are colored gold and orange.

I hope you have enjoyed valve theory 101; if you have interest I invite and encourage you to read more on the subject in one of the resources previously mentioned. Don’t feel bad if you don’t understand everything at a glance, it took me hours of study to really get what is going on in these diagrams!

On to materials!

Pump part iv: evaluation

This post finds me in the unenviable position of trying to play catch up after a long absence in posting.

In going back over my notes, I will endeavor to get everything up to speed.

When last I posted the pump materials had all been individually finished, then reassembled.

The important thing to conclude the whole affair, is to bench test the pump to predict how it will fare in field use.

I did my evaluation in stages; before mounting the various valve boxes on the trunk I tested with just the exhausters and reservoirs in place, sealing off the open ports with acetate pieces secured with duct tape around them.

pump trunk first test stage

We are attempting to make this collection of wood and cloth pieces as airtight as reasonably possible. Things to look and listen for when doing this kind of testing:

“Lock up”: a well sealed pump will exhaust all of the air from the trunk and reservoirs after just a couple of vigorous pumps. Once that atmosphere is (temporarily) exhausted, the lower pressure inside the trunk will not let you pump open the exhausters, until some atmosphere leaks back into the system. If you can get the pedals to lock up after a couple of pumps, you are off to a good start.

Reservoir suction retention: a visual indicator of how airtight the pump is. After lock up the reservoir springs will begin to force the large pneumatics open, in opposition to the reduced pressure trying to keep them closed.
The duration of time it takes for the reservoirs to completely reset and relax is another key indicator of system tightness.
One minute is the ideal, but can be difficult for an amateur rebuilder to achieve.
Thirty seconds is good; fifteen to twenty seconds fair to minimally acceptable.
Anything less than ten seconds will be not so satisfactory, and will find you pumping constantly to try and maintain suction in the system.

If the retention time is too low, try listening closely over all areas of the pump for audible leaks.
This can be awkward to do while pumping to maintain suction; you may need a helper.
I was satisfied enough with my 30 second time, then I added the valve boxes (cutoff and theme, etc).

Soft function valve box

I did some more comparative testing, as seen in this sort video. I have a pressure gauge set up to measure the differential when I actuate the soft function. It is important to note that both pressure AND flow play a part here, so a pressure gauge alone may not tell the whole story.

checking pressure differential with soft function

For a good idea of how a restored pump is supposed to test on the bench, check out this video from Bruce Newman restorations. 

If you are having trouble chasing down leaks or getting things to work, consult the Reblitz book or the Mechanical Music Digest archives online.
You’ll get there!

With the pump finally done, it is on to the stack next — yippee!

Pump part III: Pedals of Metal

The major functional components of the pump were discussed last time, but we still need to talk hardware. There are small bits of hardware (metal arm and springs) that get reattached to the exhausters, after the hardware has been cleaned.

There is an assembly of metal pieces that form the treadles and linkage arms which allow the considerable force of the player’s legs and feet to be transferred smoothly and efficiently into kinetic energy which drives the exhausters.

This assembly is naturally very well secured to the pump and so there will be many screws to remove before it can be detached. As always make a map of where things go and take photos.

pump disassembly with first two screw maps
treadle linkage screw map

Once that is all apart evaluate the condition. Hopefully the fact that the assembly is made from sturdy metal (steel or iron) will mean that it is in serviceable condition with no breaks or deformities.
In my case there was a crack in one of the treadle plates so I had to take it to a local machinist for welding. It was a relatively minor repair.

cracked left treadle

Cosmetics on the other hand is a different matter.
There is a very high chance that there are decades of grime covering the metal and possibly corrosion as well.
Clean all the surfaces with solvent or your cleaning agent of choice; you may need to take a firm bristle or soft wire brush, or very fine steel wool to get it to clean up.
If the plating is heavily damaged or corroded you will have to decide if it is desirable or possible to have the pieces replated. In my case I am not so concerned about that aspect, I like the patina as I have previously stated.

Here is a photo of before and after cleaning; the angle of light is not optimized but it’s a real difference that you can both see and feel.

Before and after simple cleaning

One issue I ran into is that the old bearings for the link pins were worn through. These were originally made from vulcanized fibre, like the valve backing discs in secondary valves.

bearings/bushings worn out completely

It’s not so easy to get exact replacements of these, but the good news is you can make your own from brass hobby tubing (much easier to source) and it will do the same job.
The pins are 1/4″ O.D. so with 9/32″ and 5/16″ tubing, you can make several sets of nested-pair bearings that will fit fairly snugly but with enough play to move freely.
I used a small pipe cutter to cut the pieces cleanly.

cutting new nested bearings for center pins

Other miscellaneous things that need doing are replacing the treadle mats. Take out the backing boards, rip off the old rubber (if needed), clean the old adhesive, cut new mats to fit, and glue them back on the boards. It will make a noticeable difference in the look.

new mats for the treadles

After that just give the bearings a little lubrication (e.g. 3 in 1 oil) and reattach the treadle assembly to the pump. Put any self-contained tubing in place, and it’s ready to go back in the bottom of the piano!

Pump II: Exhausters

I described in a recent post how the exhausters function as a kind of “lungs” of the player mechanism, pumping out atmospheric air to keep the pressure low inside the system.

I also gave an overview of tearing down the pump for rebuilding, with a bit more detail about the reservoirs. The process for the exhausters begins essentially the same way.

The first thing you have to determine is if removing the whole exhauster bellow from the trunk is advisable or possible. Unless you want to make new boards from scratch consider this one carefully. If you think you can remove them prior to disassembly, do make sure you’ve located and removed all retaining screws which may be fastening the stationary board to the trunk.
If you try to pry off the exhauster with screws still holding it, you will cause needless damage and make a lot of extra work for yourself.

There are times when the stationary boards of the exhausters should stay in place, as removing them would ultimately cause more work than it would save.

Take measuremnts of the opening span (if you haven’t already) and also note where mounted hardware on the movable boards goes and how it is positioned. This will help you reconstruct, later – even if “later” is only next week! As an example, I recorded the following photo of the tension arm and spring:

Outer side view of one exhauster, with tension arm disconnected

What I ended up doing was putting a piece of painter’s tape on the outside (lacquered side) of the moveable board, which indexed the holes for future reference. This will save a heap of trouble later.

will need to mark both of these screw holes (including the one with screw eye still in place)

Getting the cloth off usually involves a combination of heat and/or moisture and elbow grease.
It can be a bear of a job, especially if the cloth had previously been glued with synthetic adhesive.
I found it helpful to get the cloth off as intact as possible, to aid in planning the layout of the replacement cloth.

Normally the process of removing cloth on the smaller pneumatics involves slitting the cloth along the crease and separating the halves, not worrying about damaging the hinge (as they will later be replaced).

However the exhauster hinges were generally made quite robustly, so examine them and if they appear to be in good shape, you may choose to reuse them. To check this, hold the “open” ends of each board next to each other, and gently try to move them in opposing directions, i.e. along the length of the hinge. There should be virtually no play or wobble. If the hinge is still well attached and not frayed or worn, you could reuse it.

If you do need to redo the hinge, be careful taking it apart so you can see exactly how it was installed in the first place, and copy it as faithfully as possible, using the same materials if possible (usually thick cotton/twill ticking).

Once the hinge is done you can turn your attention to the flap valves. Each exhauster generally has two flap valves, an inner and an outer. Again, check to determine the condition of what is currently in place. If the leather flap is still supple and whole, it may be fine. If, when you rub it with your fingertip or a fine brush, it starts to crumble or flake apart, it is afflicted with dry rot and should be replaced. Copy the type, size and shape of the leather in fabricating a new one.

Older flap valve leather: somewhat oxidized, but not rotten. I will reuse these.

To recover the bellows you will need a “heavy” fuzzy cloth. Mine was about .060″ thick.
The best kind to use is pure cotton backed rubber cloth, but at the time of this writing it is getting harder to source.

Cut the cloth with the right directionality {insert John Tuttle link} and then do a careful layout and measurements of how you will cut in preparation for recovering. You should cut with only a very slight overage margin, to make less work for the fitment. This is especially true if the bellows is still attached to the trunk and you must recover in situ. Make notes of where the center and corners of the bellows are, and mark them on the inside of the cloth with a soft pencil or marker.

An additional consideration here is the placement of stiffeners, which are used in many kinds of exhausters. If you were able to remove the original cloth largely intact and also remove the original stiffeners with minimal loss of material, the original cloth will help with layout and the original stiffeners may be reused.

A pair of new exhauster bellows cloths, with stiffeners already glued, ready to glue to bellows

Devise a method to clamp the exhauster open (will depend on style of bellow) at the correct span, and ready your cloth and glue (hot hide glue works best, as always). Check your setup carefully as mistakes during glueup will waste a fair amount of time and expensive cloth too. You may consider having a warm iron on standby and close at hand, as a gluing aid, if needed.

Using your marks and the stiffeners as a guide carefully but efficiently glue the cloth in place on the bellows. The glue sets up fairly quickly but you will have time to make one or two quick checks to verify that you have enough glue, where you need it, and not too much, where you don’t. An example of the latter would be a large glue bead forming inside between the front edge and the outer edge of the stiffener, which will cause the stiffener to bind and the bellows will not work as efficiently. Pay attention to the corners as the cloth can bunch a bit here, and may need a bit of finessing to sit properly.

If there were retaining tacks or staples used at the hinge end of the bellows, copy the original and be careful when tapping in the tacks to support the moveable board on the bench when you hammer that side. Otherwise you could damage the hinge, which you really don’t want to do at this point!

If the scenario was difficult and slowed down the gluing, you can use the iron (not too hot) to go around the perimeter to “reflow” the glue somewhat and really get it to soak into the cloth and wood. Don’t linger too long or you will flow out too much glue and rob the joint. When you start to see just a hint of glue coming out at the seam that is sufficient. Expert restorer Craig Brougher demonstrates his process in this video

At this juncture you’ve basically done the job, unless you had previously removed the bellows and now need to remount them to the trunk. A plus of having had to remove them for restoration is that you can check them for airtightness independently, before remounting them.

The two successful hallmarks of bellows restoration are that the bellows is sufficiently airtight in both the flap valves and the cloth around the perimeter, and also that the bellows opens and closes with minimal resistance in a straight and true fashion.

Pump (part I): Trunk, Reservoirs, etc

The trunk is the main body of the pump system, in a pedal-operated player piano.

It is basically a long box of wood with holes (ports) cut in it, over which various pneumatics and bellows are placed. The trunk will contain:

  •  Exhausters: The exhaust bellows are the lungs of the player system; they are what constantly evacuate the atmosphere air from the system, to create the suction which in turn powers the system. Despite the fact that one pushes the treadles with one’s feet, the air is not pushed in to the system (a misconception), it is in fact pushed “out”. This is done by two sets of large flap valves. On the push motion of the pedal, the atmosphere from the system is drawn into the bellows, through the first one-way flap valve. The second step is that the bellows (which is under constant spring tension) closes, forcing the air which just entered to now evacuate via the second flap valve. More here.
  •  Reservoirs: as the name suggests these large pneumatics hold a reserve of suction to help regulate the airflow in the system. Otherwise the direct action of the exhausters would be too jarring and irregular. Unlike the exhausters (which are spring loaded to close), the reservoirs are spring loaded to constantly try to open. This makes them want to tend to constantly draw in suction, which is after all their job! They only have one duty, but it is an important one.
  •  Expression boxes and other peripherals: these have already been discussed in previous posts. On my pump unit, the soft and theme expression boxes are mounted directly on the pump, while the governor/accelerator is mounted separately on the underside of the keybed, connected by hosing.
  •  Treadle linkage: I will discuss this separately in an upcoming post.

Restoring: the first and most important task before reassembling everything is to clean and seal the trunk body. If the trunk envelope is not airtight, the piano will not play as it should. A leaky pump will be noisy and ineffective. Using your sealant of choice (traditional is shellac but I chose to use Phenoseal), go around the entire trunk box and check every nook and cranny where a joint is discernable. The surface of the wood itself is likely already sealed with black lacquer or shellac, but over time the joints may open up just enough to leak, so those have to be sealed up. If there are any splits or holes in the trunk they should be repaired (wood glue is suitable here).

Then it’s time to recover the exhausters and reservoirs (which you will have previously removed). If the boards were not damaged during removal from trunk or removal of old cloth, then they can be reused.

Let’s start with the reservoirs. You have removed the old cloth with heat, or water, or a power sander (let’s take it easy with that last method!). In the process of removing the cloth it’s best to leave the open end until last and take out the powerful V-shaped springs once the sides are first cut open. Do this carefully! Once cloth is gone clean off the old glue, which hopefully was hide glue. If it was a modern synthetic glue, you have my sincere condolences. There is probably leather gasket material left on the stationary board which needs to be cleaned off. Scrape off what you can, mechanically, without damaging the wood.
If there is residue left it could either be hide glue or thick shellac; glue comes off easily with water, but shellac will need denatured alcohol or methanol.

Check the condition of the hinge; replace as necessary.
Verify the condition of the springs, they are probably fine but if not replace them as well.
Before recovery I chose to apply a bolstering strap (about 2″ wide) across the open end, to help resist the strong tension from the springs, using the same ticking material as the hinge.
Reinstall the springs.
Prepare the pneumatic cloth, cut to rough shape then glue it back on.
Because of the large amount of surface area to cover with hot hide glue it is handy to have an iron to help set the glue, and smooth out any inconsistencies in the glue seam.  Trim the excess cloth with a new razor blade, after the glue sets up but before it hardens completely.

Copy and make a new mounting gasket for the stationary board, and if the reservoir has a panel cover (to allow access to mounting screws), make a new edge gasket for it, and set all aside for now.

Here are a few photos from the process:

pump mechanism (unrestored) removed from the piano

pump mechanism with the various valve boxes and reservoirs removed.
This process was made much easier because they had been attached using blotter paper gaskets.

Removing old dried shellac from the joining surface with methanol. The cling wrap helps to keep the methanol from evaporating too quickly.

Removing pneumatic cloth from the reservoir. Once the sides are open, remove the springs at this point — carefully!

All cloth removed from the reservoir pneumatic.
Repair or replace the hinge, and the spring felt.

Recovering the reservoir pneumatic. The last steps are to add reinforcing tacks, then a final overlapping strip of cloth on the closed end. This is done with the pneumatic closed to keep the pneumatic cloth from binding.

panel covers for exhauster pneumatics. Removing and replacing the edge gaskets.

A 1919 Willis player piano