[CAUT] Key ratio measurement (was Re: moving capstans question)

[Richard Brekne]

Hi Fred

I see you and I have been thinking along similar lines here at some 
point. I agree with your closing line completely, and its a point I try 
to make often.... indeed Keith, you and I just went over that ground.  
Consider the following for a sec. If first you do want to pre-dict your 
BW... and taking Stanwood's basic equation as our medium for the 
example... you have:

  BW = (SW * R) + WRW - FW

Now if you happen to know (as is very easy to do) what your SW's are, 
and your WRW is... and your FWs... then the only variable left that can 
explain why a resultant BW doesn't match up to what you planned  on is 
the Ratio... yes ?

Now no matter which way you measure the ratio... Stanwood's way 
included... your result is going to be a bit iffy at best.  But as Keith 
points out... it doesnt matter... as long as you catch the error before 
you get all carried away installing all those FW's in time.  If you find 
your resultant BW is say 2 grams off for 5 or 6 sample keys you've 
temporarily balanced to SW's... the really all you need to do in a 
spreadsheet is add / subtract appropriately to your ratio figure and 
recalculate the FW's...  The formula gets turned around thus:

 FW = (SW * R) + WRW - BW.

Here BW is specified...  FW's are solved for... and if the resultant 
BW's dont match your specified... then you need more (or less)  FW's   
Basically this just reveals the <<error>> in whatever Ratio figure 
you've calculated on.  Doesnt really matter which ball park Ratio you 
started with... you make a quick change in your spread sheet so that the 
resultant FW's do indeed yeild the BW you specify.

So you are right... IMHO.... a lot of talk about ratios is cross 
purpose... and things can easily get mucked up a bit that way... quite a 
bit if you are unlucky enough.

I like Jon Pages sequence... he first establishes an optimal placement 
for the capstan based on whippen travel and the line of convergence...  
THEN he applies Stanwood.  That way he gets a great regulation action... 
as well as all the benifits of doing the weigh off ala Stanwood.  If you 
are going to change the action ratio... by moving the capstan etc... 
seems best to me to approach it Jons way.. or some similiar based 
thinking.  The weight thing is best used for what it was designed to 
do... balance SW's against FW's for the existing ratio.

Cheers
RicB


         > I believe on
         > Overs site he mentions something about the need to take into
         > compensation the angles at which the parts are designed to be
        in their
         > rest positions to get the correct reading... because it will be
         > different if you measure same in another position.


        It occurs to me that the "elephant in the room" is the degree to
    which
    in non-Overs action pianos the knuckle is away from convergence (way
    below).
    Maybe that is the answer to my question right there (along with possible
    wipp/capstan non-convergence).
        I did find it rather distressing to try out all four (including
    Spurlock's addition of weight to hammer) methods and come up with a
    different number for each, and not just a little different, either.
    I was
    thinking of coming up with a kind of quickie way of predicting an
    optimum
    blow/dip relationship and ball park numbers. Maybe practical, or
    maybe not
    (probably the old experimental "try this, try that" method is just
    as fast
    in the long run), but it seemed like it would be cool if it worked. No
    wonder manufacturers' specs are just starting points a lot of the
    time <G>.
    Unless they have really fine manufacturing tolerances, like our Japanese
    friends, and probably many of the Germans as well.
        But it does make me wonder if a lot of talk about ratio numbers
    is at
    cross purposes, because different people are measuring different ways.
    Regards,
    Fred Sturm
    University of New Mexico