[pianotech] re design

[Gene Nelson]

John,
Thank you so much for the detailed explaination.
I had no idea that treble strings intentionally slope to the bridge on this 
type of piano.
I am curious if it is reasonable to make the treble string slope angle 
determinations by comparison to the level of the strings in the lower 
sections?
Gene

----- Original Message ----- 
From: "John Delacour" <JD at Pianomaker.co.uk>
To: <pianotech at ptg.org>
Sent: Tuesday, February 03, 2009 1:31 PM
Subject: Re: [pianotech] re design


> At 19:15 -0800 2/2/09, Gene Nelson wrote:
>
>><<The typical hammer for note 88 with agraffe is small compared with the 
>>normal in my experience. I do not necessarily want a larger than normal 
>>hammer - but at least normal - probably an Isaac hammer or equiv. I 
>>believe that the typical agraffe hammers at the top note/s are small 
>>because of strike point issues? Hammer hitting agraffe/plate so they get 
>>reduced in size? No, I do not like thuddy/knocking sound up there.
>
> A treble with studs can sound wonderfully strong and clear if the proper 
> measures are taken and if very careful measurements are taken before 
> boring and gluing on the hammer heads.  I am working on two such pianos at 
> the moment, a 1924 Bechstein C and an 1865 Kirkman Style 2.  Neither piano 
> had any useful sound in the high treble when I took them in.  Both the 
> Bechstein had and the Kirkman hammers were original.  I ordered VFG 
> hammers for the Bechstein and hammers with a special Wurzen felt for the 
> Kirkman.  In both cases the top hammers, from Abel, were considerably 
> fatter than the originals, but I'm used to that.  They can be slightly 
> sanded down as needs be.
>
> Though the Bechstein is well made generally, the metal frame under the 
> agraffes was very rough.  The hammers in the whole top section and a few 
> below had been fouling the frame and must have done so almost from the 
> beginning.
>
> There is one big difference between this Bechstein and this Kirkman. The 
> high strings on the Bechstein are, most unusually, almost horizontal, and 
> the strike height is almost uniform throughout the scale, whereas on the 
> Kirkman they slope up to the bridge at more than 5 degrees, reducing to 
> about 2 degrees at the first break.  The latter configuration is far more 
> common and requires a lot more thought, as I will explain.
>
> There is _one_ optimum strike point for the hammer in the extreme treble 
> and that is roughly 1/16 of the speaking length, say 3mm for note 88. If 
> the string is struck by the hammer head at a perfect right angle at this 
> point, you will get the best sound for that hammer head.  If you plan to 
> remove felt from the top of the hammer, this must be taken account of 
> _before_ the hammers are bored.  Better than removing felt to increase 
> hardness is to inject a little dope round the tip of the moulding.  That 
> said, I have needed to use no dope on either the Kirkman or the Bechstein 
> and the extreme treble on both is excellent.  I hardly ever use dope.
>
> Now, neither the Kirkman nor the Bechstein ever sounded or behaved as well 
> as they do now, and the reason is that the original finishers did not take 
> account of the individual characteristics of the instruments as regards 
> strike height and string angle.
>
> There is a good reason for the gradually increasing upward slope of the 
> strings in the top section of grands that use studs, and that is that it 
> allows more clearance between the near side of the hammer-head and the 
> metal frame, since the hammer, in order to strike the string at a right 
> angle must be cast forward at the same angle as the strings' upward slope. 
> Thus a wider hammer-head can be brought to strike the string at the proper 
> point than if the strings were horizontal. I don't think Bechstein ever 
> quite worked this out; on the older grands the top hammers are 
> exceptionally thin and spindly in order to clear both the belly and the 
> metal frame, and by the time they made this 1924 piano they were still 
> sending out pianos that after a little playing would have problems with 
> treble tone.
>
> The boring length should follow the differences in strike height, so that 
> if the  strike height is, say, 200mm at note 40 (strings horizontal) and 
> the design requires the head to be at 90 degrees to the shank (not by any 
> means the rule), then for a hammer centre height of 150mm, the bore length 
> is 50mm if you make no allowance for wear.  If the strike height is 196 at 
> note 85/88, then the bore length must be 47...
>
> ...But then there is the important question of the _angle_ of the head to 
> the shank, which must be altered following the slope of the strings.  If 
> in the above example I went ahead and bored all the hammers at 90 degrees, 
> my top hammers would understrike considerably and not only would I need to 
> thin the hammer-head to stop it fouling the frame but also I might still 
> not hit the proper strike point, even at the wrong angle, and tone would 
> suffer.
>
> Back to the Kirkman, where the original hammers were bored roughly 
> respecting the strike heights, but all at the same angle, from the point 
> where the strings began to slope upwards to the bridge, I gradually change 
> the bore angle and, when gluing on, gradually increase the length from the 
> hammer centre to the centre line of the hammer moulding.  The result is 
> that, for a string slope of 5 degrees and a bore length of 50mm, the 
> hammer head is glued on 4.36mm further out than the head of note 40.
>
> This is just a worked example, which supposes that the strike line is 
> exactly parallel to the hammer rail.
>
> I am sure that if restorers, and even makers, were to learn these things 
> and put them into practice, we'd hear a lot less about dope and hammer 
> weights.
>
> JD
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>>That said, I have experimented with larger - or should I say heavier 
>>hammers in the high treble with very good results. They just need to be a 
>>bit harder. I have an idea that more hammer mass up there will allow the 
>>strike point to be lengthened slightly but I cannot prove it just yet.
>
>
>