[ Editor's Note:  Below is a message from Robbie Rhodes which
 [ contains, as an attachment, a message he received from our new
 [ subscriber A.B. Bonds.  Within that attachment are comments
 [ which are delineated with {} marks.  Thoese comments are
 [ Robbie's, not mine.
 [
 [ There's a possibility here to get an ongoing thread regarding
 [ the engineering subtleties of reproducing pianos.  I'd like
 [ to see this happen for two reasons.  One, it will help some
 [ of our readers troubleshoot certain kinds of malfunctions.
 [ Two, it will help those here who are interested in writing
 [ software emulation of these systems (for the MIDI interpretation
 [ of roll data).

In response to my note in digest 951214 I received this interesting
reply from new member A. B. Bonds. Can any  piano techs out there
give further advice about repairing B valves?  {I have snipped out
portions of my original letter to save space. -- Robbie Rhodes }

========== attachment follows =====

To: rhodes@foxtail.com¶
From: "A. B. Bonds" <ab@vuse.vanderbilt.edu>¶
Date: 14 Dec 1995 11:38:15 -0600

> Welcome, A. B. Bonds, and thanks to "BamBam" Bill Jelen for telling
> the world about the Automatic-Music Digest.

Thanks.  I have access to fairly substantial (disk) storage here, and
have been contemplating setting up an Ampico web page.

{ Neat-O!  Tell us more! }

> Have you compared the Ampico A versus B control systems?  Quite
> a difference.

Agreed.  I have put a sensitive electronic vacuum gauge on my B and
find that the level (which is pretty wobbly over time) can shift
substantially in about 50-80 msec.  Cool.  The "servo" nature of the
control flaps that Hickman designed is astonishingly clever.

My A is rather slower, but it is in some ways more reliable, and
certainly works better when things are slightly wrong.

{ snip - description of interference filters for drawer motor }

All I have done so far is to hang a 0.1 uf ceramic across the points.
I'll try your suggestions.  I don't get that much interference, but
have had trouble keeping the points clean.  Dirty points are usually
first apparent in the form of a very slow rewind.

{ You need a small resistor in series with that capacitor,
  in order to limit the spark current when the points _close_. }

When I first got my B'eeste, the motor ran very poorly and there was a
lot of arcing in the brushes.  Turns out that it had not been run
since the early 50's (the rewind mechanism had given up and so they
used it as a coffee table).  The motor sat in the same position so
long that the goo that holds the armature together had sagged.  The
armature was very slightly out of round and the black stuff had oozed
out between the commutator segments.  I chucked it in a lathe and very
carefully turned the commutator true, then undercut the insulation
very slightly with an X-acto knife.  Motor very happy now.  This may
contribute to the interference you were experiencing.

{ snip -- regarding the poor valve performance }

Yeah.  Try the Gershwin "Rhapsody in Blue" (recut from D/A) or "Islamey."
Low vacuum + long hose = trouble, especially in the notes from the
middle of the tracker bar.  I have tried to minimize the length of the
hose on the back of the stack, which is about the only place one has
any flexibility in that regard, but I am still not satisfied.

{ snip -- cleaning the ball check valve }

I have cleaned the ball valves in acetone in an ultrasound bath.
(Note:  the acetone container was sealed, and sitting in water in the
bath, else you can get a Big Bang.)  They generally seem okay.  Every once
and a while a unit valve will stop working, which can be traced to
some junk (usually a splinter of wood) in the ball valve.  My greatest
concern is that I did not repouch the unit valves (save one or two
that especially needed it), although I did reface the buttons.  Most
of the pouches seemed pretty good when I tested them "by mouth", and I
dropped neatsfoot oil onto the pouch surfaces via a hypodermic
needle.  The exercise they now get seems to have taken some toll, and
I am now getting a fair amount of leakage in some valves.

Even when the unit was freshly set up, though, I got lousy action in
those middle notes, which did not seem to depend on whether the unit
valve was "good" or "bad".  I swapped several different ones in and
got about the same no matter, which leads me to believe that there is
some problem there.

{ snip -- adding primary valves to B system }

I haven't had a chance to argue the merits of the A vs the B type
valves.  I like the easy adjustment of the B button travel as well as
the bakelite seat, but they do seem awfully twitchy in terms of
reliability.  We'll have to chat more on this topic.  I am impressed
that someone is making a whole set of valves, which would seem to be
best done these days with a numerically-controlled milling machine.
True?  And where do you hide the primary valves?  It's awfully
crowded down there....

{ Primaries are above the action stack, same as in the A.  -- Rob }

Regards   ab