Measuring center pin resistance

[Phillip Ford]

I wanted to make some additional comments about center pin friction. 
Rather than try to respond to various posts I thought I would start a 
new thread.  Comments below:

     With regard to measuring with a gage vs. measuring with your 
finger (or whatever part of your anatomy you choose to measure with) 
- the end result is to get an accurate measurement within an 
appropriate tolerance that is consistent and repeatable.  If you can 
do that with your finger more power to you.  I've been doing this a 
fairly long time and feel that my fingers are fairly sensitive.  But 
I've repeatedly proven to myself that my fingers are nowhere near as 
accurate as a Correx gage.

     Speaking of gages - I'm not talking about one of these $5 leaf 
spring gages sold by the supply houses (I assume they still sell 
those things).  I'm talking about an accurate gram gage of the 
appropriate range, such as a Correx gage.  Such a gage will set you 
back $100 (US) or more.

     I can see a difference between checking to see if a jack is 
functioning on a console in the field and setting up a performance 
piano action on the bench.  In the first case I probably wouldn't 
'whip out a gage', but would check with my finger.  But I certainly 
don't see anything wrong with using a gage.  In the second case I 
would certainly use a gage.

     I think some of the confusion apparent in the recent exchange is 
evidence that we haven't developed or put in place the proper words 
and/or numbers for unambiguously communicating information about 
center pin resistance.  To begin with, I don't believe there is 
consensus about what word is used to describe what we're trying to 
measure.  Friction, resistance, stiffness, other?  Any votes on which 
seems most appropriate?

     The use of a gram number to describe center pin resistance is 
incomplete at best, and misleading or simply wrong at worst.  This 
harks back to discussions of measuring or describing downbearing.  To 
say that a center has 5 grams of resistance is like saying that a 
piano has 5 thousands of downbearing.  What does it mean?  Nothing 
without further information.  In both cases you have to know the 
measuring point for the number to have any meaning.  Describing 
downbearing with an angle is unambiguous because it removes the 
measuring point from the description.  Describing center resistance 
as a torque is unambiguous because it removes the measuring point 
from the description.  Perhaps it's time we started using torque to 
specify center resistance.  Rather than specify 4 grams 25 mm from 
the center pin, why not specify that the resistance should be 100 
gram-mm?  This seems more accurate and less subject to confusion 
between sender and receiver of the information.

     With regard to the traditional tests that have been mentioned - 
the swing test and the 'Steinway test' (I assume that what is meant 
by the Steinway test is to tap the shank and see if the flange will 
rise) - I did a few measurements on a Steinway flange to give some 
perspective to this.  I repinned the flange several times to make it 
match the extreme ends of these two traditional tests.  All numbers 
below were taken on a old Steinway shank and flange in good condition 
(I didn't try this on a new shank and flange, so your mileage may 
vary).  All gram measurements were taken with a 2-15 gram Correx gage 
at a measuring point on the flange to the side of the screw hole, 23 
mm from the center pin.

SWING TEST - depending on your source I've seen anything from 3 to 7 
swings specified:

For 3 swings:  7 grams at the measuring point or 161 gram-mm torque.
For 7 swings:  3 grams at the measuring point or  69 gram-mm torque.

STEINWAY TEST - tap hammer shank and see it flange will rise:

    Loosest pinning at which I could get flange to rise:  below the 
lowest indicated mark on the gage (somewhat less than 2 grams - for 
the sake of argument let's call it 2 grams) at the measuring point or 
46 gram-mm torque.

    Tightest pinning at which I could get flange to rise:  8 grams at 
the measuring point or
184 gram-mm torque.

RESULT:  The range in torques between minimum that meets one of these 
tests and max that meets one of these tests is 46 gram-mm torque and 
184 gram-mm torque.  Or in other words, if you had two adjacent 
flanges that met these traditional tests you could have a potential 
difference of 184 - 46 = 138 gram-mm torque.

To put that number in perspective - for those who do hammer weight or 
strike weight smoothing per Stanwood - the tolerance that I work to 
when matching strike weights to a curve is 0.1 gram (I don't allow 
any individual SW to be more than 0.1 gram different than the number 
specified by the curve).  So, max difference in SW between two 
adjacent hammers is 0.2 grams.  Typical distance from hammer center 
line to hammer center pin is 130 mm.  Variation in torque between two 
adjacent flanges caused by variation in SW = 0.2 x 130 = 26 gram-mm.

As you can see, the possible difference in torque between two 
adjacent flanges from using traditional tests is more than 5 times 
the difference caused by the variation in SW.  In my opinion, if 
we're going to go to all the trouble to smooth SWs, and to do the 
sorts of things that Jon Page describes, such as sorting shanks by 
weight, and knuckles by size, in order to get as smooth a progression 
as possible of touch from one end of the keyboard to the other, then 
saying - don't bother with a gage, just use the traditional tests - 
doesn't make sense.

Phil Ford
San Francisco, CA