At 3:20 PM 6/29/95 -0600, stanwood wrote: >The data below shows FrictionWt as (DownWt-UpWt)/2 >Before and after moving the capstans and heels >approximately .375" to change strike ratio from an average >level of 7.5 to 5.5 > > >(chart omitted) >The piano played liked a nightmare, now it plays like a dream. > >Small knuckles would not have solved this problem. > _______________ I was out of town for the weekend or I'd have responded sooner to this post. I'm not entirely sure, but this very interesting data documenting capstan relocation on a particular piano and the profound improvement it obviously realized, seems to be presented as a proof which somehow should contradict my earlier statement about knuckle sizes. I do not see a contradiction here at all. There are two different kinds of friction in the action that we must be concerned with, and there are many sources for each kind. I do not know the proper scientific names for each kind of friction, but one is the type found in tight centers or bushings that are unrelated to geometry, the other type is a binding directly caused by less than optimal geometry. This is nothing new. The trick of course, is correctly diagnosing bad geometry where it occurs, and solving that problem without creating others. I appreciate this revealing data about the capstans on that piano, but it does not really say anything about friction as it relates to knuckle size except that on this piano they are most likely correct. If you are using the Hamburg shanks, than the knuckles are the correct size for that leverage. The point I tried to make is that the pre-84 shank from Steinway has an improper-sized knuckle, resulting in bad geometry and excess friction. Changing this knuckle to the correct smaller size would solve at least one geometry problem and reduce friction. I just did this on a 1984 D a couple of weeks ago. Unfortunately, I do not have impressive data collected like what we saw on this D from Harvard, but only changing knuckles on the pre-84 shank (which are about .410") to the smaller one from Renner USA (about .350"), lost an average of 4-5 grams DW. I did not notice much change in upweight, it was a nominal improvement. The smaller knuckle on a lesser leveraged shank also gives the jack a little more room to clear it in the window, which really helps. I will go further and say that if you had installed smaller knuckles on the aforementioned Hamburg shanks, ignoring the capstan problem, that still would have somewhat reduced total friction within the action, but clearly in this case that would not have been the correct solution. It would not have solved the real problem and would not yield a satisfactory result. The interesting sistuation here is that a too small knuckle on the Hamburg shank will reduce overall friction, in spite of that incorrect geometry. This says that total friction is not necessarialy always inversely related to good geometry, and that some friction is good. I think that this apparent contradiction with the normal inverse relationship between friction and good geometry must be observed in the context of its involvement with so many variables. It is difficult if not impossible to completely isolate all of the individual contacts and lever ratios as they relate to total friction. Chris Robinson spent quite some time with this issue a few years ago and did some geometry classes on knuckle sizes. I also believe that Ken Sloane wrote an article some time ago for the CAUT newsletter about his experience with the pre-84 shank when he ended up with some horrendous DW numbers. But to be fair, if you never use the early Steinway dimension shanks, then this is not a problem that you would ever run into. I have pursued this issue primarily with the hope that technicians will not misjudge the early dimension shank from some bad experiences due to an incorrectly- sized knuckle. This shank does work well (with the appropriate weight hammer) as David's strike ratios have confirmed. As long as you concede that smaller knuckles have less friction than larger knuckles, independent of other variables, I do not see any conflict. If I misread the intent of your data, please forgive my confusion. And thanks for sharing that data, it really is an impressive improvement. I don't have my atlas here to look up its date, but that is an old number. What is the oldest Steinway you remember seeing with significantly wrong capstans? Dennis Johnson St. Olaf College
This PTG archive page provided courtesy of Moy Piano Service, LLC