So nice of you to say so, but it's more than reasonably predictable.
It's very predictable, unlike wetting and clamping, which doesn't appear
to be predictable by any description. For those who don't wish to drive
across town or the county three times sneaking up on it when the last
attempt moved again, this is a one shot fix, and I've done a number of
them as field repairs with a tuning. No, before you ask, I don't carry a
band saw, but a small hand saw is doable for one key. I have never had a
failure or a repeat. I'd rather there was a method that didn't involve
chopping up the keys, yet still worked controllably the first time. If I
knew of one, I'd adopt it and defend it every time the subject came up.
But until I hear of one or get lucky and discover one myself, call backs
for an independent tech are a bit more than a walk down the hall for a
re-do. Kerfing is practical, predictable, and dependable, but ugly.
Practical is my choice until further notice.
Ron N
Original Message------
Ron's method is certainly a successful and reasonably predictable one (one variable being the width of the saw kerf, and how much angle change that thickness of kerf will produce, another being how much the hinge creates a rotational force during clamping). My experience has been different from his in that I have been successful in using heat and moisture to correct warpage and twist in keys.
I was quite frustrated in my first attempts, as there was no real control, and in trying to apply enough twist or bend I ended up breaking a couple keys at the balance mortise. I decided there had to be a better way, and so I set about to design one, with several iterations along the way. The first principle was to immobilize the middle of the key and protect that weakest part against breakage, so I came up with a clamping arrangement spanning the mortise by quite a bit (see my previous post for photos).
The second was to find a way to do a predictable bend (or unbend if you like). In many of my first attempts I came up short or overshot, so I decided to be organized about it and measure precisely where the point of the key I wanted to move was at rest (in the clamped position), and then to measure how far I was pulling it. An obvious by-product of this is that after applying the bending procedure and removing that clamp, but leaving the middle clamped, you can measure the final result and know what you have done without putting the key back in the piano.
From experimentation, I found that heat alone wasn't enough, probably partly because of my dry climate (I find I often need to wipe hammershanks with a wet cloth before "burning" them for alignment, as doing it dry takes much longer during the driest periods). So I started by applying steam and wetting. I found that applying water to the part of the key I wanted to bend gave the best results, and that it paid to take my time and let it become good and wet, absorb water into the interior. So I ended up applying water with a paintbrush, on all four sides, then wrapping in a wet cloth and allowing at least a few minutes to pass (if doing a number of keys, they can all be wet and wrapped before doing anything else). A heat gun provided a means of directing the heat where it needed to go, and a deflector made it possible to do so on all sides of the wood.
The first step is to mark the key while it is in place, noting what needs to go in what direction and how much. For a twist, I will make an arrow on top and a complementary arrow on the bottom (going the other way). For a bend, just one arrow will do. In each case, a number in mm is estimated and written beside the arrow. If it is a twist, sometimes the top needs to move more than the bottom or vice versa, so that is noted.
From experience, I have found that I always want to overshoot the target, and the typical amount is to pull the key approximately three times the target distance (IOW, to the target and then past the target by 2 times the distance). I often end up with the key overshooting a bit, and then settling to the target after a few weeks - not a problem as long as this doesn't cause rubbing in the other direction when I first re-install the key.
I've done quite a bit of this, and followed it for years. Since I got all the ducks in a row in terms of establishing all the steps of procedure, I have been very successful, with both twists and bends, and have found my results stable, with the exception of a couple keys that warped to a pretty extreme degree against the vertical plane (key front up, key back up relative to balance hole area - so that the key is much higher than its neighbors, more than can be corrected by removing wood from the balance hole area). In the one piano with that problem, I glued some veneer to the bottom of the back of the key after as much bending as I was able to accomplish, and that solved the problem (it was a grand, and fortunately high notes without dampers, so I didn't need to compensate by reducing the key end felt thickness).
Is it completely predictable? No, it isn't. Results will vary with wood grain, age of wood, etc. Young, relatively soft wood is much easier to bend than tighter grained old wood. But it is predictable enough to be a workable, reliable procedure.
------------------------------
Fred Sturm
University of New Mexico
fssturm@unm.edu
http://fredsturm.net
http://www.artoftuning.com
"We either make ourselves happy or miserable. The amount of work is the same." - Carlos Casteneda
------------------------------