Removing Ampico Roll Flanges Nondestructively

If you collect original Ampico piano rolls, you will often find
examples that have portions of one or both flanges broken off, or that
show telltale signs of cracking.  If either or both flanges are broken,
and each corresponding shard is not available for reattachment, it is
not wise to play the music roll on your Ampico piano.  One runs the
substantial risk of tearing (e.g., splits, rips, fold-overs, confetti)
the edges of the paper during the rerolling process.  Even if the
piano's tracking mechanism is in proper adjustment, there is still
the distressing chance of damaging the paper edges during the play
operation.

Replacing a broken left-hand idling flange is a simple matter.  One
need only twist the flange and slide it out of the core of the roll,
discard the damaged piece, and replace it with an equivalent flange
that is still intact.  Such undamaged flanges have a way of accumulating
from the remnants of other Ampico piano rolls whose paper was deemed
unrepairable.  Replacing a cracked or broken right-hand drive flange
is another matter altogether.  Whereas the left-hand flange was just
pressed into place, the right-hand drive flange was originally pressed
and glued into place with an all-too-effective adhesive.

One could discard the core along with its broken flange and then use
a modern replacement core and plastic flanges to remount the original
music roll.  This approach results in the need to find a reliable
source of new high-quality flanges and cores.  Alternatively, there is
a restorative challenge of 'fixing' the original so that as much as
possible of the original construction is preserved.  Is there a way to
separate the broken right-hand drive-flange from the core of the roll
without damaging the core? The answer is yes.  One solution is
described below.

The key to successfully removing the right-hand flange is to know what
adhesive was originally used to bond it to the cardboard core.  Because
shellac was used elsewhere in the player mechanism for a variety of
sealing and bonding purposes, I assumed that its stickiness and fast
drying properties would also be effective for the flange-bonding
application.  One way to evaluate this possibility was to use denatured
alcohol to release (dissolve) the annular grip between the 3/4-inch-long
stem of the circular flange and the interior cylindrical surface of the
cardboard core.

A few (15) simple test trials with broken drive flanges and their
attached cardboard cores showed that denatured alcohol does allow a
drive flange to be separated from its core with no damage to the
cardboard.  Although this does not prove that the original adhesive was
shellac, it certainly supports that view.  Clearly, this result shows
that the adhesive releases its grip (dissolves) in the presence of a
sufficient amount of denatured alcohol for an appropriate time period.
The details characterizing how much alcohol to utilize and for how long
are documented below.

The first step is to unspool the paper roll from the core.  Then
remove the left-hand flange to gain access to the opening into which
the alcohol can be poured.  Without spilling, introduce enough alcohol
(a Goldilocks amount, i.e., not too much, and not too little) into the
vertically-held tube core to cover the 'stem' of the right-hand flange
about 3/4" (within the core interior).  If insufficient alcohol is
supplied here, there will not be enough liquid to soak and spread into
the core-stem interface.  In that case, subsequent twisting of the
flange will likely tear out portions of the interior of the cardboard
core.  Pouring in excessive alcohol will soak into the cardboard too
much, weakening it and/or distorting its shape.

The next step is to wait until the alcohol 'stains' (wets) through
to the exterior surface of the core.  The stain first appears as an
exterior surface wet spot(s) and then spreads, downward, upward, and
circumferentially.  Let the downward stain progress so that it reaches
the inner face of the drive flange, making sure that the stain goes
all the way around the core exterior.

Let the upward stain rise about 3 to 4 inches above the inner face
of the drive flange, also all the way around.  The two stains do not
have to occur simultaneously, and usually will not.  This double stain
circumstance constitutes another Goldilocks criterion.  Too little
stain spread can cause gouges and tears when core-flange twist is
applied; too much stain spread results in mushy cardboard.

Both of these 'spreading-stain' characteristics must be satisfied for a
successful separation.  You may have to add a bit of alcohol to realize
both these requirements.  There is some variability in elapsed time for
this procedure because of variations in the thickness of the cardboard
(core to core, and within a given core) and the porosity of the
cardboard material itself.

Commonly, both of the above conditions are satisfied after about 15
to 30 minutes.  This is the time needed for the alcohol to disperse by
wicking and the adhesive bond to loosen up (dissolve).  Experience shows
that it is better to err on the side of too much soak time rather than
too little soak time.  The last step is to pour out any leftover alcohol
from the core interior and firmly but gently twist (clockwise and/or
counterclockwise) the right-hand flange relative to the cardboard core.
It should readily twist out with no tears or gouges from the cardboard
core.

Leave the core out in the open to dry thoroughly, keeping it away from
open flame or other concentrated heat sources.  When the alcohol aroma
has dissipated, the cardboard core is dry.  Then, one can insert a
closely fitting replacement right-hand flange.  One can use a wrap or
two of 3/4"-wide Scotch tape on the flange stem to ensure a reasonably
tight fit between the flange stem and cardboard core.  Or you can use
that old stand-by, fresh shellac, to preserve originality.

Bill Koenigsberg
Concord, Massachusetts