Restoration of a Knabe Ampico Reproducing upright for Carol Drummond
37) 
38) 
39) 
37) The old key bushings are being soaked out of the keys.
This is a delicate operation because the felt and wood have to be wetted enough to
loosen the glue holding the bushings but not so much that other glue joints suffer from
it.
38) This piano has a special kind of center pin which requires an exceptionally
narrow bushed opening at the centers. My suppliers do not sell bushing tools called
cauls which are designed for this type of key bushings. Therefore, I made my own out
of quarter sawn maple which I carefully thicknessed to the nearest thousandth of an inch
to ensure that the keys will fit the pins will.
39) The keybed on this piano had a rats nest under the keys and much dirt and grime
worked into the wood. Here, I am washing the wood and metal very vigorously with
soapy water to remove as much of this dirt as possible.
40) 
41) 
42) 
40) The keybed has now been cleaned and new felt punchings
installed. The back rail of felt is not done yet in this photo because the felt
needed was on order. I had run out.
41) The new pin block inserts are now epoxied well into position. This final
assembly will be as strong as the original but the hold that the wood has on the tuning
pins will far exceed that which was built into the original design.
42) The customer requested a medium red mahogany color on this piano.
Unfortunately, this piano was originally a very dark brown. It was necessary to
white wood sand all the old stain off the piano in order to dye the mahogany with a medium
red dye which would have a chance at meeting the customer's expectations. This is a
photo of all the loose boards and the main body of the piano after all the white wood
sanding was complete. This work was not something which was originally planned for
the job and the extra days of labor were not calculated into the price. However, the
customer really desired a lighter red finish and I wanted to see to it that she got that.
43) 
44) 
45) 
43) As each board is ready to be stained, I inspect every square inch very
carefully. I'm looking for any dents or scratches which the sanding did not erase.
I'm looking for missing veneer and also for loose veneer. In the case of this
board, I found that the veneer was loose along the bottom edge. It is being glued
back into place using spring clamps and glass clamping cauls. Glass works well as a
clamping caul in lower pressure situations because it helps to keep the wood's surface
smooth and flat. Also, any glue squeeze out which gets past me will not cause the
clamping cauls to be glued to the veneer as would be the case if I used wooden cauls.
44) This is part of the key cover. This thin, delicate board carries most of
the weight and the pivot hardware for the fall board. The design of this fall board
(or key cover) is quite unique. Knabe was very fond of making music racks and key
covers which operated in ways which were different from their competition. Sadly,
these special made designs seldom held up because they were dependent upon delicate wooden
parts holding up under excessive force. I am always repairing the key covers and
music racks on Knabes for this reason. In the case of this piano, the veneer was
still well glued into place but the core wood had been totally stripped out by the screws
which held the hardware against the edges of this board. Not only had the screws
stripped but then the fall board was used some more after the initial damage and the loose
screws cracked the core wood badly. To repair this, I've excised all the core wood
at both ends of this board. In its place, I've epoxied in multi-laminated hardwood
pin block material cut into circles that fit the holes I drilled. Then I've leveled
the surface using mahogany circles which seal the new assembly. This set up should
be strong enough to prevent this same failure from happening again. In addition, I
will use extremely long screws which I will predrill for when I put the hardware into
place during reassembly.
45) Here you see all the boards of the piano after they were white wood sanded and
then dyed red with an aniline dye. The beautiful part about anilines are the way
they make the wood dance. The color of the piano will really sparkle. Under
sunlight or other forms of direct light, the wood will be a very light to medium red
mahogany color. In shadowed areas, the wood will take on a much warmer, darker cast.
Also, the grain of the wood will be much more vibrant as it changes color as you
move your head.
46) 
47) 
48) 
46) The main body of the piano after the stained wood was
sealed, filled and then resealed. The color is now much darker. However, under
the right light, it will seem quite a bit lighter in color or darker depending on how you
light it. I'm sorry for the fuzzy focus on some of the pictures shown in today's
updates. I was having some camera troubles which are now sorted out.
47) This photo and the one to come are both of the fall board with its damaged ends.
I couldn't manage to save the board and the veneer because the board has so little
"meat" to it. Here and in the next photo, you see the restored board with
a piece of new veneer cut oversized and positioned above it ready to be glued into place.
The original board had one long boring piece of veneer on it. The replacement
will be book matched veneer done in two pieces which will make for a little more
interesting appearance.
48) Although I did glue mahogany circles over my repairs hoping they would blend in
well enough to be able to reuse this board's original veneer, you can see especially in
this photo that it just didn't work out.
49) 
50) 
51) 
49) The arrow at the bottom left is pointing to the board
which was repaired above. All the other boards are the various cabinet pieces which
have been white wood sanded, stained, sealed, filled and resealed. These are now
ready to receive the new lacquer finish. I don't normally like to do white wood
sanding to pianos. Although this technique does help me create a more perfect
surface on the wood and does make the fresh stain more vibrant, it thins the veneer.
This is bad because there is only so much thickness to the veneer and you don't
want to remove that. The next time this piano is refinished, it will not be possible
to do another white wood sanding. It is highly likely that another such sanding will
start creating areas where the veneer is sanded through completely. During the
process of refinishing I spend a great many days doing tasks which are all the same and
don't really warrant photographing. It is during this part of the job that the work seems
to be moving the slowest as you watch it over the internet. Rest assured, the work is
progressing at a good speed but it is a lot of the same tasks which just don't need to be
listed here. I can show the piano being rubbed out, for instance, with one photo. However,
the job may take a week to two weeks or more to complete...
50) The time has come to start rebuilding the player mechanism. This is a
photo of the upper action of the mechanism before work began. This is a shot from
the back.
51) This photo was taken after the cover board was removed from the front.
52) 
53) 
54) 
52) These are all the pneumatic devices that are not part of
the pneumatic stack. These include the tracker, play/rewind, wind motor, control
valves, expression cut out and some cover boards.
53) The top of the action has been removed. The head is located to the front
of the photo and the pneumatic stack is in the rear. The stack includes the primary
valve chest, all the secondary valves, the trunks (which are split into two halves
internally for expression purposes,) the lifter fingers and connecting rods and the
pneumatics.
54) This is the primary chest. This is a design that Amphion did not use a
great deal. Usually the primary chest valves are all wood and the leather for these
valves are glued to the valves themselves. In this case, the outer part of the
valves which can be seen in this photo are all made of Bakelite. The leather seats
are actually glued to the top of the chest instead. This design will work just as
well as the other design. The only fear is that the Bakelite could have gotten
brittle with time. If that happens, the outside valve heads will have to be
replaced. The only way to find out if they're reusable is to take the chest apart
and see if they come out or fall apart.
55) 
56) 
57) 
55) This shows another rare occurrence for Amphion.
First of all, the pouches inside this chest are of the very highest quality. These
are made from the thinnest pouch leather they had available at the time. We can't
even get stuff this thin anymore... I always find this type of pouch leather inside
the lock and cancel system of the Amphion expressions but thicker stuff was usually used
elsewhere. These pouches are actually still useable and are of very high
quality. However, they are old and so will be replaced with new leather never the
less. The other thing about this chest which is unusual is the fact that it was
glued shut. Usually Amphion (who always used cork in their gaskets instead of
leather) would glue the gasket to one side of the chest, place the halves together, screw
the chest shut and then glue a strip of rubber cloth over the edge of the joint to prevent
leakage. Cork gaskets require considerable pressure to seal completely. The
strip of rubber cloth was an added safety precaution against future leakage from humidity
swings that would cause the chest to expand and contract. In the case of this chest,
all those things were done but, in addition to this, they glued both sides of the gasket
into place. This meant that it was impossible to ever take this chest apart for
servicing. The only way to get it apart was to chisel it apart as I did. This
ruins the gasket in the process... I see this sort of thing often in pipe organs
but never before have I seen an Amphion primary chest that was glued shut like this.
Yet, I doubt that this was done by some later technician trying to get the system to
seal. The look of the outside sealing strip of rubber cloth looked factory original
to me so I think Amphion is responsible for this little piece of insanity. The new
gaskets will all be made from leather. This seals without the need of further steps
and does not require the extra tension in the screws to keep it air tight. This is
easier on the chest, easier on the screws and more long lasting while leaving the assembly
easy to take apart for future servicing.
56) Here, the primary valves are being taken apart. They are held together by
a dot of glue on the top of the valve head. One swift tap on the top of the valve
stem separates the halves. We were lucky, the Bakelite is solid as a rock and should
be fine. I will take the trouble to reflatten the Bakelite since this substance can
slump over time. I want to make sure the Bakelite is flat so that the seals will be
perfect.
57) These are the secondary valves after they were removed from the trunks.
These, too, were unusual in that the pouch leather inside of them was much thicker than is
usually used by Amphion. The new stuff will be thinner and more sensitive.
58) 
59) 
60) 
58) These are the three tiers of the stack after they were
taken apart. The next job is to remove the pneumatics after marking their locations
carefully.
59) This shows one tier of the stack with the pneumatics off. To get them off
with the least amount of damage, it is best to warm the wood with an iron. This
softens the glue so that the parts can be taken off without damaging the wood. The
warmth, however, drives moisture out of the wood it touches causing the boards to warp.
To correct this, the boards are moistened lightly and then placed in a stack and
clamped up as shown. This enables them to be pulled flat and true. After the
moisture is totally gone, the boards will be taken apart and will be totally flat.
Then they can be cleaned up in preparation for recovering.
60) Here you see the next tier being done. The pneumatics are cut open, the
movable boards are torn free and all the cloth is torn off the wood. Then the
movable boards are warmed so that the lifter fingers can be taken off without damage
(after being marked for locations) and then the fixed boards are warmed to remove them
from the rail. Another common way of approaching this task is to simply split the
pieces apart. This does work but is harder on the wood if you aren't exactly on
target with every stroke.
61) 
62) 
63) 
61) The metal outside valve seats are being removed.
The valve bodies are turned upside down onto a hot iron to warm them. This softens
the burnt shellac that holds them in place. Once they are warm, a special tool is
put down inside to lever the valve seats out.
62) Now all the outer finish is removed from the valve blocks. This outer
layer of burnt orange shellac is all that keeps these blocks air tight. Very little
sealant is used inside of them. Since the blocks must have their bottom layer (they
are made in three layers of wood) split off from them to get the old pouches out, the
finish has to come off in order to expose the glue line.
63) A sharp, straight knife is laid on the glue line and given a gentle tap.
Then the block is rotated and the process repeated around all four sides. After
about two passes all the way around the joint is freed up enough to remove the pouch block
layer. After I get the pouch blocks split off, I remove the old pouches while
keeping the old lifter discs. I clean the old glue off the lifter discs and then
seal the inside of the pouch wells with burnt orange shellac as shown in this photo.
This sealing is not needed for the final product. However, with the pouch wells
sealed, I can test the pouches to be sure they are really air tight before I glue the
blocks together. The original way of doing this made it necessary for Amphion to
glue up the valve blocks after they "think" the pouches are air tight.
Then they had to seal the outsides of the valve bodies and let that dry. Once they
were dry, then the pouches could be checked to see if they were air tight. If any
were not air tight, then you either discard the valve or you run more sealant down the
tiny signal hole and let it soak up into the pouch. Then you have to try to run talc
down this same hole to bind the sealant. This is a big, unnecessary hassle.
With the pouch wells sealed in this way, I know if my pouches are air tight long before I
glue the blocks back together. While I'm at it, I seal the layer of wood that will
separate the air that is above the valve body from that which is below. Amphion did
put a little sealant here but not very much and certainly not enough after all these
years. I do the sealing now so that any shellac that I slop onto the glue joint will
be automatically cleaned up when I dress the joints in preparation for reassembly.
64) 
65) 
66) 
64) The next job is to reseal each trunk board. These
boards are coated with shellac on the inside to make them air tight. For the system
to work properly, these boards must be absolutely air tight, especially at the point where
the two halves are separated from each other inside. Each tier has an internal
separator in the middle so that they are divided into bass and treble for the use of the
expression system. To reseal these boards, I first cover all the holes that are
drilled into the trunks with masking tape. Then, the pass through holes that run
from the valves to the pneumatics (holes which are not part of the internal structure of
the trunk) are opened up by trimming the tape away from them. Then the inside of the
trunk is literally filled to the rim with shellac. Then the shellac is immediately
drained back out. The other air passages are coated with shellac using a brush.
This ensures that the trunks will be air tight. This step is sometimes
unneeded because the original sealant is still holding the trunk strongly air tight.
I seal them anyhow... In this case, the trunks were, in fact, leaking.
65) The spacer blocks for this piano have been refinished with new orange shellac
and are drying in this photo. The photo also shows some of the secondary valves in
process. Starting from left to right; old valves fresh out of the blocks, in the
middle are the wooden valve heads after the leather and felt has been cut off them, to the
right, the valve stems with the inside leather facings still attached to them. The
inside leather facings are held to the valve stems with a simple tack. This
connection is made in this way so that the valve head will wobble in relation to the stem.
If the seats are not perfectly parallel, this wobble will let the valve seal
perfectly never the less.
66) Now the wooden heads are being cleaned up. All the residue is scraped and
sanded off. I use sandpaper backed with a piece of glass to ensure that the wood
will be totally flat once it is clean. You can see that these wooden discs have an
indentation in the centers. On the left you can see a green residue which is the old
green felt dots that were in those indentations. On the right the green felt has
been removed. More about that in a moment.
67) 
68) 
69) 
67) Reassembly of the valves has begun. The first step
requires explanation. The tack I mentioned a moment ago could make a noise as it
bounced against the wooden valve head. To prevent this, Amphion originally drilled
out a nice little pocket in the valve head to leave clearance for the tack. This was
an excellent solution as it let the valve wobble as needed without being too
"mushy." In this later design, the drilled out pocket is much smaller and
a dot of felt is glued into it. The felt keeps the tack from making noise as it hits
the valve head. However, the felt also makes the valve "spongy" because it
is so thick. In the process of rebuilding these valves, I altered the original
design. I replaced the thick felt dot with a thin dot of split suede leather.
This will keep the tacks from making noise but does not stick out beyond the surface of
the valve head. By doing this, the valve is much less spongy which means repetition
will be faster and valve action will be more precise. In this photo, the white are
the wooden valve heads with their new leather outside faces glued into place. The
purple is the inside valve facings ready to be attached to the valve stems and then glued
to the valve. The brown is the old leather with the valve stems still attached.
I use different leathers on each side because the dimensions of the valve bodies
require that the valves be about .200" thick. In order to obtain that
thickness, two different thicknesses of leather must be employed. My thinner leather
just happens to be white while the stuff that is thicker was dyed purple.
68) This is the process of gluing the inside valve leather and stems onto the valve
heads. The old leather is held to the "x" shaped valve stem with a tack.
The tack must be carefully pulled to release the old leather. Then the new
leather is marked for dead center and the tack is driven through that spot. Then the
tack is pressed back into its original hole. Usually, these tacks are held so
tightly that the act of taking them out and putting them in again leaves them very snug.
This is the case with this piano. The tacks gave me no trouble. Then
the leather and valve stem assembly is glued to the wooden valve head. Care must be
taken not to get glue onto the leather dots near the tack. The center area near the
tack must have no glue so that the valve will wobble. To ensure that the leather is
flat and will seal against the valve's inside seats, the valve is placed into a valve
block and pressed down onto the seat to ensure the leather is properly affixed.
69) A picture of the finished valves ready to go back into the blocks.
© Copyright 2010 {David Rodgers' Piano Rebuilding}. All Rights Reserved.