Restoration of a Knabe Ampico Reproducing Upright for Carol Drummond
181) 
182) 
183) 
181) This photo shows one of the four bellows in the pump
getting new leather flap valves. The moveable board gets one leather flap. The
fixed board gets four strips of leather. Two strips are glued to the wood and a
metal bar is attached over it. Then the flap valve itself is stretched over this
leather and metal seat. The metal strip is intended to quiet the action of the flap
valve and make the pump more silent. Experience has taught me that this isn't
needed. You can actually eliminate the silencing bar and the pump will run just as
quiet. However, I do leave it in with one slight change. The original leather
seats had this bar mounted over the leather. I cut a channel in the new leather to
recess the bar part way into the leather. This leaves the bar sunk down more than it
was originally. I believe that this helps the flap valve seal better. With the
original design, the flap valves would be less air tight while the pump was new until they
had a chance to mold themselves around the metal bar. The pump would tighten up over
time. However, the leather available now, doesn't seem to want to mold around this
bar well enough to make the pump as tight as I want it. It is a known fact that
leather valves seating against leather valve seats will leak. Leather does not seal
well against leather. Amphion knew this and installed these leather seats in order
to create a controlled leak that would prevent the pump from bogging down. The
Ampico pump is one of the most time consuming to rebuild because it usually take an entire
day to remove all the old leather, clean up all the mating surfaces and hardware, cut new
leather and reinstall everything. Other pumps are simpler and can be releathered in
an hour or so.
182) This is the air manifold which connects all the pumping bellows together into
one air column. It was originally gasketed with cork which doesn't really seal that
well against the pneumatic cloth covering the bellows. I replace this cork with
leather after filling the manifold with sealant and then draining out the excess.
Many parts of these players were not sealed when the wood was new because the nature of
the wood itself meant that it wouldn't leak air. However, 90 years later, a lot of
things have happened to this wood. It becomes wise to seal the wood just in case any
tiny leak points have developed in the wood over the years.
183) The new leather glued onto the manifold. The manifold also got repainted
and all the rusted screws and washers were cleaned up to look like new again.
184) 
185) 
186) 
184) All of the various parts of the pump have been restored
except for the connecting arms. This photo shows all the restored parts laid out
ready for reassembly.
185) These are the connecting arms before they were restored. I mark each arm
to ensure that it gets put back onto the pump in the same way it came off. I want
the metal feet to seat against the same bellows they started out with and I want them to
connect to the crank shaft at the same point they did originally.
186) All the felt washers and bushing cloth as well as the spring loaded wedges had
been soaked with oil. This would have made the pump knock and the parts gum up.
All the washer felts could have been reused had they not been soaked with oil.
I had to make new ones which were packed with dry graphite powder as were the new
bushings. Dry graphite is the only acceptable lubricant for these parts. In
addition, I had to clean off all the oil from the wooden wedges, true up their mating
surfaces and burnish them with dry graphite powder so that they would not create any
knocking noises. If the bushings on these pumps are not very tight and all the
wedges in perfect condition, the pump will knock. We call it the infamous Ampico
"Pump Thump!" The spring loaded wedges are intended to keep the bushings
snug and quiet even after they have worn. In my experience pumps which have no such
adjusters run just as quietly as this type of pump. The wedge idea was a good
thought but pretty useless in the long run. However, since they are made this way,
they must work perfectly or they will make noises.
187) 
188) 
189) 
187) The pump is now well on its way toward reassembly.
188) The restored connecting arms have been driven onto the crank shaft. The
bushings are made tight enough that the parts have to be lightly driven to get them onto
the guide pins. Any looser than this and they will become noisy in a very short
time. Although the bushings must be glued into place on the wooden connecting arm,
care must be taken to keep any glue squeeze out off the inside of the bushing. If
even a speck of glue gets on the wearing part of the felt, the pump will squeak.
189) The metal plate with the drive wheel, bearings and crank shaft/connecting arms
has been screwed back into place.
190) 
191) 
192) 
190) The connecting arms have been bolted back onto their
respective bellows.
191) This photo shows that the screws which hold the cut out manifold in place have
had their tops sealed with bee's wax. This was an original design feature at the
factory and should always be repeated during restoration. I've never known these
screws to be a source of leakage because I use leather for the gaskets instead of cork but
it is best to be safe here since weather swings could effect this in ways not immediately
apparent. Amphion learned through experience that these screws needed to be sealed
so we should follow their lead in this.
192) The back of the pump is all reassembled as well. The new lines connecting
the equalizer to the pump's air flow have been put into place also.
193) 
194) 
194a) 
193) Now that the player mechanism has been restored, it is
time to turn our attention back to the piano itself. The new ribs have been shaped
so that they fit the notches in the rastin (the ledge on which the sound board is glued)
and match the widths of the original ribs as taken from the old sound board. Any
places where the original work done by the factory left the rastin fitting the ribs
improperly, repairs are made to ensure a tight fit.
194) Now the long bridge is placed over the fitted ribs and set so that it is
exactly in the position it will occupy after reassembly. Then marks are made
directly on the ribs to show where the long bridge will pass over them. I will then
shape the ribs so that they have the proper crown built right into them. I set this
crown in such a way that the highest point on each rib is as close to being directly under
the long bridge as possible. My jig that is used for shaping the crown on the ribs
has an indicator mark that shows me where to align my marks on the ribs so that the jig
will shape them accordingly.
194a) A photo of the process of shaping ribs so that they will be crowned and also
set to be highest directly under the long bridge. A router is used to cut this shape
into the ribs.
195) 
196) 
197) 
198) 
195) After the ribs have been crowned, I clamp them to the
old sound board and trace the shapes of the original ribs onto the new ribs. Then
the shape is rough cut on the band saw and fine tuned at the sanding station. After
that, I set the ribs back into the piano and fine tune their fit by hand until they are
perfectly flush in their fit into the notches.
196) Gluing the first parts of the new bass bridge together. The first stage
was to glue up the core of the bridge. For this, I used a technique known as
vertical lamination. Strips of flat sawn maple are glued up into the shape of the
bridge. By using flat sawn maple for these strips, once glued, the quarter sawn
dimension of the maple is set between the strings and the sound board. Maple
transmits vibrations along its quartered grain better than it does the other direction.
The strips are favored because they transmit the vibrations of the strings along
the length of the bridge better than solid wood can do. In this bridge (which is
straight) solid wood could have been used with success. However, the laminated wood
will also be better because of the mix of West System epoxy used to glue it up and the
laminated effect making the entire assembly harder, a better vibrator and more stable.
In this photo, a cap of quarter sawn maple is being glued to the top of the bridge.
The cap is needed because the details of the bridge notching and pinning cannot be
successfully carried out through the laminated areas. You need solid wood for that.
197) The long, treble bridge to this piano has one place which curves very severely.
The original bridge was cut from solid wood and the shape made that way.
However, the grain of the wood (when cutting from solid stock) is interrupted along the
length of the bridge as the shape is cut out. If you interrupt the fibers of the
wood, they cannot transmit vibration as efficiently. To combat this difficulty,
vertically laminated material is called into use again. However, the bend in this
bridge is so severe that the maple would break before it bent around this obstruction.
Therefore, the wood is moistened and then heated and bent around a form until it
takes on the shape that will be needed as it is glued up into its final dimensions.
198) This shows all the layers of the treble bridge core after they were bent in
preparation for gluing up. Clamps hold the shape in place until the wood cools and
dries completely.
199) 
200) 
201) 
199) A special jig that I use for making vertically
laminated bridges is called into service. The jig is cut out to the exact bridge
curve of the piano I'm currently working on. The form is lined with thin plastic to
keep it from gluing itself together and then the layers of vertical grain maple are coated
with West System epoxy and placed into the jig. Then the entire assembly is
carefully clamped down until it is under tremendous pressure. Once the epoxy has
completely cured, the new bridge core can be removed and will hold its shape exactly as
set up in the jig with no springing back whatsoever. After this, I will glue the jig
back together again so that I can cut a new shape out of it for the next bridge.
Occasionally, I make an entirely new jig when the old one has been cut out and glued back
together too many times.
200) The core of the bass bridge with its cap is finished and ready for indexing and
gluing to the rest of the bridge.
201) This bass bridge has four primary components. The vertically laminated
core, the cap (both of which have already been made,) the apron and the gluing stand.
Each separate piece has to be pain-stakingly made to match the dimensions of the
originals exactly and then the index holes drilled into the bridge earlier are transferred
onto the new pieces. Then the entire set up will be glued together.
202) 
203) 
204) 
202) The quarter sawn hard maple cap has been cut out and is
ready to be glued to the treble bridge core which is vertically laminated. Now that
the core is out of the press and is the same shape as the original bridge, it must be
glued to the cap on another jig which will impart the appropriate amount of crown to the
long bridge. Just as we went to great lengths to pre-crown all the ribs, the long
bridge must also have its crown built into it. This will make the entire sound board
assembly more strong and durable. It will also increase the power of the tone of the
piano. Most rebuilders do not take this step into their own work. They glue up
the long bridge totally straight and bend it to the sound board later. I believe
this adds undue stress to the glue joint between the bridge and the soft spruce of the
sound board. Far better to create the crown by using the maple to maple joint of the
cap to core.
203) The long bridge has its new cap glued into place with the same West System
epoxy. I like the epoxy for its acoustic properties but I also like the fact that it
is totally impervious to moisture. These glue joints that I am making in this bridge
will be forever. No amount of moisture exposure can ever cause them to come apart.
Other joints that must be reversed in the future (such as the joint gluing the
sound board to the piano itself) will be done with acoustic glues that are effected by
moisture. This way the sound board can be replaced again in the future with little
harm to the joint with the piano.
204)
205) 
206) 
207) 
205) The new bass bridge blank is complete. Since the
original bridge was in such horrid shape in the cap area, it will be better to make the
marks for the bridge pins based upon my own measurements based on the original as a
reference.
206) Many rebuilders would take this bridge blank and glue it to the sound board
just as it is and carry out the rest of the work after the sound board was glued into the
piano. And they are quite correct to do this. However, the reason they do this
is because they want to check the bearing levels for the strings with the assembly glued
into the piano and the plate lowered into its final place. They can then mark the
bridge for height and plane it to the correct height for the sound board and also mark the
locations for the bridge pins based upon the locations dictated by the plate. This
is a quite good way to proceed. However, I have certain beliefs about this process
that cause me to feel this work is better done outside the piano; at least by me.
First, this is a Mason and Hamlin which means that the original bridge dimensions were
certainly perfect from the factory. Any minor adjustments to down bearing that I
feel may be needed can easily be done using the original bridge and adjusting the plate
height and plate felt thicknesses. Every factory has to make sure their bridges are
the correct height or the pianos won't work properly. Even if the factory were to
make mistakes, they are always minor enough that they can be corrected by plate height,
sometimes shimming duplex bars and felt thicknesses. I am personally much better
able to plane bridges squarely, mark bridge pin locations accurately and drill and notch
bridges precisely with the bridge clamped to a bench where I can see it and get at it more
easily. Once it is in the piano it becomes very hard for me to work on with
precision so I do this work outside the piano. My first job is to measure the
thickness of the original bridge down to the nearest thousandth of an inch all along its
length. These measurements are precisely transferred to the new bridge and then the
bridge is planed down by hand on the bench until every inch of bridge matches the original
for thickness exactly. That task has just been completed in this photo.
207) The next step is to take a pattern from the original bridge of the locations of
all the bridge pins. I start with index holes which are drilled into the bridge and
help me to align the bridge pin locations and the sound board to get everything aligned
properly with the plate. Then I use a soldering iron to burn a little hole through
clear mylar which has been pinned to the bridge. These holes become my locating
marks onto the new bridge. The pattern, once made, is moved to the new bridge and
the index holes lined up. With some bridges it is necessary to "figure
out" where the index holes belong based upon several different other measurements.
The best way to index a bridge is to drill a hole through the bottom of the sound
board, through the bottom of the bridge right up through the top. However, if this
hole isn't perfectly square to the entire thing (which it seldom is) the marks will not
help much in the transition to the new bridge. For this reason, I prefer to use my
best judgement and numerous measurements all over both bridges to find the best place to
locate the pattern. The bottom index holes that align the bridge to the sound board
are easily transferred with accuracy by cutting a slice off the bottom of the old bridge
and taking it over to the new one and lining it up. With such a thin slice, angles
in the index hole will not have had a chance to telescope out and make anything
inaccurate. The most important thing to get right when doing this is the locations
treble to bass of the pattern and the length of the strings at note 88. Most pianos
have a 50mm length to their highest note. Some will go as high as 55mm but they
seldom go less than this. It is important to keep this measurement correct or the
tone of the treble can be compromised. A few millimeters discrepancy in the bass, on
the other hand, has little to no effect on tone as long as it is just in the length of the
string that the difference occurs.
208) 
209) 
210) 
208) Back to the bass bridge to finish up. These two
photos show the bass bridge after it was completed. The first photo shows the bridge
after the pin holes had been drilled, the top lubricated and the edges chamfered.
The second photo shows all the bridges I am currently working on after they were finished
and the sealant coat put on them.
209) Here, you see the long bridge has had its pin holes predrilled. A special
collar is custom made to fit over the drill bit. The collar is made from wood that
is slightly softer than maple to prevent it marking the wood at the top of the bridges.
Still, it does leave a slight mark that actually goes away after burnishing,
notching and pinning. The reason for this wooden collar is to prevent the drill
itself from making contact with the bridge. Of course, I could just stop the
drilling before the drill hits the bridge. However, I want the holes precisely the
same depth each. I want them a specific depth which permits me to drive the bridge
pins all the way into the bottoms of the holes. The tips of the bridge pins will
seat into the maple and make better sound transmission. Also, with the holes all the
same depth, I can get the bridge pins to the same average height across the bridge for a
nice cosmetic result. The normal way to do this job is to drill the holes too deep,
drive the pins in on the shallow side and then grind them all to the same height with a
file or belt sander. This gives a good cosmetic result but compromises tone a
little. This way of drilling enables me to not have to grind the tops of the pins at
all. This makes for a better tonal result at the same time it produces a very
special cosmetic result.
210) The top of the long bridge has received four coats of McLube. This is a
special lubricant which can be burnished to a nice, soft gloss and makes the top of the
bridge very smooth and even slippery for better tuning stability and a good cosmetic
result.
© Copyright 2010 {David Rodgers' Piano Rebuilding}. All Rights Reserved.