During the last ten days, on the Pianotech part of http://my.ptg.org/, a discussion has taken place on the subject of string leveling in the agraffe sections of grand pianos. David Skolnik started it off with observations of agraffe imperfections, even in good quality pianos, and an interest in hearing what fellow piano technicians thought. This led into quite a long exchange on agraffes, leveling strings, and the importance of mating hammers with the strings they play. I missed the give and take, but added the following approach, which I have come to in developing my Grandwork™ Regulation Station tools and their linear grand regulation protocol. The variables of plates, agraffes, and strings at the heart of this discussion challenge the piano technician who would successfully regulate and voice grand pianos to limit accumulated error. The Grandwork™ tools have been designed to provide exactly the high quality referencing and precision of execution needed for this task.

"Late to the conversation, again. But I have some useful ideas to add to the good thinking that has happened here.

First of all, I do have the Grandwork™ Underlevel for leveling strings at strike from underneath (which doubles as an excellent string height gauge). It is a worthy tool but not the way I personally level strings...

Allen, and others, have noted planar discrepancy between unison strings as they leave the agraffe or V-bar and as they arrive at the bridge. This twist in the natural plane of these strings along their speaking length presents us with a technical problem. There can only be one place in their length that the strings are parallel to the keybed, and that place needs to be strike.

Parallel to the keybed simplifies. The most simple address of the vertically arrayed mechanics of a grand action is to have each of the linkages in the transfer of power be as vertical as possible. The keytop goes down, the capstan and backcheck go up, the whippen jack tip, tender, and repetition lever go up, and the hammer knuckle, shank, and hammer head go up. The parts then bounce down (up for the keytops) to reset, catch, repeat.

Viewed from the side, travel along each of the arcs to make this happen does its best with motion vertical at half-travel (averaging the motion of multiple places on a part). From above, the closer these arcs can align front-to-back the better. From the front, the more vertical the planes of these arcs can be, the simpler. Accumulated error can undermine our efforts at best tone, power, and repetition (TPR) and good regulation, in general - in the piano, as well as out on bench, lap, or lid.

Addressing this parallel spot vertically shortens the distance between rest and strike and provides a bounce that is straight back to rest: it simplifies the elements between finger exertion and production of sound. We need to make this spot parallel right at strike, and the hammers know exactly where strike is.

If a hammer is well-filed and travels vertically and is vertically balanced on the driving force of its shank at strike, and the unison strings are fit to its horizontal crown at strike, the strings will be level in the right place, well-mated with their hammer, and parallel to the keybed, ready to provide Fred's "focused sound" (and more). This happens naturally, without rework, if we use the real strike in the piano (as opposed to an estimated or as-close-as-we-can-get one) and if we true the hammers before fitting.

If you don't remove the dampers, can a bubble gauge reach strike on any strings of a Steinway? Strike is too close to the capo bar in the high treble, dampers are too close to the plate in the next section down, and strike is too close to the dampers in the tenor, dipping under them entirely at about middle C.

But is leveling the strings in the wrong place that big a deal? It's a slight complication in a situation that has multiple complications, so why go there, if an alternative can leave this complication out?

Without restringing, imperfections of the agraffes are a given (and a challenge, if we do restring). They are an impediment to achieving our goals that must be accepted, respected, and worked around. Overall plate bowing, again, is a given (with a significant benefit, BTW, in consolidating string bearing to soundboard resistance) that complicates our string-to-hammer mating - in the high treble, the treble-side string must be lifted more than the middle string, which then must be lifted more than the bass-side string.

Let me add one more complicating element, referred to in the discussion, before winding up. Underlifted and overlifted strings should be minimized or eliminated. Freshly strung wire is underlifted, initially. The stiffness of the steel wire will not relax under the corner of agraffe or V-bar to extend directly out. There will be curvature. The more curvature out from a straight line at the termination points, the less stable the speaking length will be. Those curvatures are like springs and will diminish power, make tuning stability harder to achieve, and possibly cause tonal complications. A relatively small lift with a stringing hook will reduce the strike-end curvature to a stable place (but will not shift a stable string). And a push or mild tap with a hammer shank will settle the bridge-end curvature.

Overlifting is a more obvious hazard, with similar complications and a potential for string breakage, now or later.

If hammers arrive at their strings varying from vertical, if the strings are not parallel to the keybed at strike, if the hammers are less than ideal in shape, if their travel also varies from vertical, and/or if we've pre-leveled the wire in the wrong place, combinations of these variables may cause us to overlift strings, while trying hard to do the right thing by our needed string-hammer mating.

An alternative is to misfile hammers to achieve the fit not accomplished by moderate leveling. This complicates voicing, in general, but really complicates soft pedal voicing.

The experienced regulator/voicer has filing skills, regulating chops, voicing tricks, and an array of solutions to apply in response to constraints of piano, time, budget, and the vicissitudes of by-eye referencing. Even for them, variables represent challenges and simpler is better. For the rest of us, simpler is way better.

So. My tools or yours, travel and square hammers to vertical, position the action in the piano for best tone, and level strings to hammers in the piano. This needs to be coordinated with settling the wire throughout its lengths to finish with the fit at pitch and the piano in tune.

Finally, Ron, a push back against the idea that kinks don't have an effect. In the early 80's, I restrung a concert instrument for a well-known venue in Boston. Settling wire everywhere was part of how wonderful I was going to make this instrument. And most seemed best. I settled it good and proper on the speaking side of the bridges, only to then have to raise tension back to pitch, a process that began the lifelong process of my kinks being drawn further and further into the speaking lengths I was trying so hard to make best-of-all-possible. Maybe the annoying false beats in this piano were not of my doing - that would make me feel better.

If prep-work is well thought out and executed, moderate practices can produce predictable, best results. Check out my Squaring Platforms, Shank Traveler, and Hammer Squares for ideas on achieving the vertical components of good prep." 

CB  10/18/2014