Hi Stéphane Thanks for airing your thoughts on the matter. I guess this all boils down to whether or not there is any reason to have the shank perpendicular to the key bed or not at impact. You cite having the shank parallel to the string plane at impact, and that makes an outwards rake compatible right away. I will have to re-read Bob Hohf's article to get you a straight answer on that. Perhaps I have misunderstood that but I seem to remember him making a justification for the action delivering its maximum punch with the hammer shank parallel to the key bed regardless of what was being punched as it were. As to errors in measurement you describe below. Strikes me that which ever way you go about it you are bound to get an error of some sorts. I measure typically the string height with a jig I saw described in the PTG journal some years back, which essentially is an adjustable ruler that is attached to a flat block siting on the key bed. You push the thing up to string contact and read the result. Thats dead on accurate in itself. Reading center pin height I usually do taking measurements through the spaces in the keys to either the front or back side of the action rail with an adjustable height square jig. Then I just add the height of the center pin over the action rail measurement. Thats where any error occurs, but its a total error of less then 0.5 mm and how picky can one get ? I have used the pre-hung hammer trick as well, but again have always used it to find the shank at horizontal at impact. Thats been accurate enough as well as one simply attaches a small leveling bubble to the shank. I'm not quite sure how you could use that trick to get an accurate shank parallel to string plane at impact measurement. At least not one that is any more accurate then the measurement above. One way or another some small degree of error seems inevitable I suppose. My main querrie is really this shank at horizontal bit. If I drop that requirement, then I'd be able to get the hammer perpendicular to both the string and shank at impact if the action cavity allows for it I susppose... or at least pretty close to it. And that would account for a backwards rake yes ? Cheers, and thanks again. RicB Hi Ric. When I think of it, I can see only one ideal situation, that is when the hammer is perpendicular to the string at impact time, and the shank is closest to parallel to the string at that time. In this condition, the energy transfer from the hammer to the string seems maximized to me. If the shank is not parallel to the string, there is a portion of the hammer movement that will not be normal to the string plane (but parallel instead), and that movement will spend energy in friction (of the hammer against the string), and the rebound of the hammer will also be slightly impeded because he hammer will have to reverse the direction of that portion of movement too. I fail to understand the importance of the shank being horizontal. What are the ideas of Bob Hohf about that ? The trigonometry is fine, but you could add an analysis of what happens with the measurement errors that you do when measuring - String height over keybed (I don't seem to be able to do this measure with less than 0.5 mm error) - Rounding this value for all notes between two samples you measured (typically at each side of each section), so if for example you measure 231 mm at the right of the treble section and 230 mm at the left of the alto section, all notes in between will have an additional error of 0 to 0.5 mm - Center pin over keybed : here you have two measures and one approximation : the approximation is that when you measure any point on the stack over the keybed, you do this with the action out of the piano (I suppose, or do you have a better way ?) and so the action rests on another surface against which you do the measure, but is the action frame sitting exactly the same way ? I find it difficult to avoid another 0.3 mm error here. Also you don't measure directly the center pin height (or do you ? but how ?) you measure one accessible point at the top of the stack, then measure the vertical distance between that point and the center pin, and here again errors sum up - yet another rounding error for all notes between the extremes, if there happens to be a slight difference in the measurement of the first bass center pin height and the last treble one - angle between string plane and keybed (I can't even figure out how to measure this accurately. My best take up till now was to throw a ping pong ball from 20 cm over the strike line and measure where it falls down after its first rebound, but this assumes that the keybed is horizontal, which I can grosso figure out with a bubble gage) - again approximations across the scale - and finally the error in hammer bore itself This is where the sample hammer with known bore distance trial in situ shortcuts a whole bunch of accumulated errors (in the trigonometry formulas, any measurement you have done will multiply even further all the errors), as there is only one approximation, that is the perpendicular condition of the hammer with the string plane, which is quite easy to measure with a piece of straight wood with a nail put in it at straight angle, which you put on the strings with the nail facing the ground. And if the hammer is perpendicular to the shank, then necessarily, the shank will be parallel to the string plane. But this tells nothing about the horizontality of the shank, and I fail to understand why I should care. So I don't understand either the rake thing, short of trying to float for all the said approximations when it is time to have the actual end user hammer as perpendicular to the actual string as practically possible. What do you think ? Best regards. Stéphane Collin.
This PTG archive page provided courtesy of Moy Piano Service, LLC