Hello Jonathan and MMDers:  I was interested in Jonathan's recent
posting (040804 MMDigest) since I had the same problem: rebuilding
a rebuilt pump in which the original measurements had been lost.
 I spent some time thinking about the problem and will share my
thoughts with you and would appreciate any comments.

It seems that there are two important distances: (1) the width of the
cloth and (s) the length of the connectors to be made from rubberized
canvas belting.

First the width of the cloth.  It must be no wider than the width of
the board (5-3/4" in this case) + twice the thickness of the boards
(1/2").  This is the width which results in the folds just touching
in the closed position of the exhauster pneumatics.  The width of the
cloth must also be wide enough that the cloth is not stretched when the
pneumatic is at the widest opening, governed by the throw of the crank
and the length of the connectors.  The danger is that the fold will
blow out rather than in as the pneumatic closes (as Jonathon points
out).

Second, the length of the connectors must be such that there is no
tendency to buckle when the pneumatic is at minimum opening which would
result in significant loss of movement of the pneumatic.

In my case I decided to cut the cloth 4-1/2" wide, which allows about
1/8" for trimming to give a span of 4-1/4".  The connectors were cut
3" long with the holes for mounting to the pneumatic and the crank
spaced by 2-3/8".

As far as I can tell the pump is operating well.  The pulley on the
motor is double (was this ever standard?) and, using the smaller one,
the pump itself rotates at 120 rpm.  With all nipples blocked my
vacuum gauge is pegged at 80+ inches of water lift.  The piano is still
a little too loud with the spill set at 25" of water lift.  I do not
like setting it much lower since operation of the dampers and hammer
rail becomes somewhat sluggish.

Regards
John Johns
Ottawa, Canada