Wrong scaling?

Instrument Building
1

So I am building a Zuckermann / The Paris Workshop French Double VI kit. I purchased it shortly before or shortly after their splitting (1997 maybe?), then left it alone for many years.
The design is ZHI’s, while the actual production was made in France by Marc Ducornet / TPW.

The “kit” is a bit less than a kit, now, because I re-did or I am re-doing:

  • the soundboard. I discarded the kit soundboard and re-did from scratch using my valued Abete rosso della Val di Fiemme spruce (or is it fir?) I purchased from Rivolta in 1996.
  • The 8’ bridge, as it was completely warped (not a fault of the factory, the bridge was a victim of a small flood in my house due to a major fail of my hydraulic system).
  • The jacks and the tongues. I didn’t buy the wooden ones as I planned to make them. Of course I don’t want to use the plastic jacks.
  • I’ll quill it in bird’s (swan) feathers.

As you may know, this kit is derived from the Edinburgh 1769 Taskin. I have a drawing of the Taskin and there aren’t macroscopic differences excepted the zhi-tpw has 63 notes instead of 61, reaching g’‘’.
Everything else is as “historical” as can be: everything is solid good quality wood, no plastic, no plywood, and everything has the shape it should have. Well, registers and lower guides have regular rectangular mortises of the correct size, but they are routed and not sawn/chiselled. I can live with that so I’ll let them alone, if I don’t change my mind before stringing.

As I made - ehm, I’m making - a new 8’ bridge, I have checked the scaling, just to be sure.
I found it more or less “historical”, the length of c’’ is 354 mm (13.94 in) (long string, of course).
However I’ve found some deviations from the ideal that make me worry.
I am not suprised, as on page 10 of the manual DJ Way writes:

======

Space for an extra row of jacks on this instrument would seriously distort the treble scaling-which is already distorted a bit to allow the instrument to extend to g"‘. French double harpsìchords ought not to have a higher note than f"’, and some day I will take my courage in both hands and eliminate the top two notes of this instrument - but this is a subtlety that affects only a few notes in the top treble, and is not to be worried about.

======

The scaling is indeed lenghtened a bit, though the stringing list doesn’t change from iron to high-tensile steel, so probably deviations are not dangerous (however, the stringing list is disputable, I’ll not follow it. More on this in the upcoming weeks or months). Moreover, the scaling seems to be a tiny bit too rapidly foreshortened starting from c’’ and lower. The harpsichord is 237.5 cm (93.5 in) and the curve of the bentside seems a very usual and traditional French curve.
Nonetheless, I am worried and I wonder if any participant would mind reading and commenting the scaling. Remember, it’s not the actual scaling in my harpsichord, measurement are from the drawing.
What do I do? close my eyes and strictly follow the drawing or calculate a more “right” scaling and glue the bridge(s) accordingly? (or, if I find a table with the complete Taskin scaling I could just copy that - it’s difficult to get it from the drawing: anybody has such a table? maybe Claudio who has written an essay on Taskin’s scaling and stringing?) Or maybe I could eliminate the top two notes and relocate the bridges accordingly? and modify the keyboards, not an easy task, and fill the space in the case… uhm, I hope you will not advise so, too much work. I wouldn’t eliminate the strings leaving the keys “muted”.

Scalingfrench_Pagina_1
Scalingfrench_Pagina_11654×2339 104 KB

Scalingfrench_Pagina_2
Scalingfrench_Pagina_21654×2339 96.1 KB

2

Hi Domenico.
The Edinburgh 1769 is one of the few extant Taskins where the stringing list was NOT inscribed in the instrument.
I am sure my article on Taskin Scaling and Stringing will be of help.

All the best for your very interesting project!.

3

HOW TO HAVE A TOP G’‘’ AND ALSO FOLLOW THE ORIGINAL TASKIN SCALING.
Needless to say, I would stay with the original scaling. The bridge displacement needed for the extension to g’‘’ (in order to allow for space for the 4’) is minimal indeed and would only affect notes well above c’‘’.
This said, I recommend a different solution:
given that the extension to g’‘’ is needed for the music by D.Scarlatti and A.Soler, which is NOT meant for a 4’ foot, I would omit the two topmost 4’ notes f#‘’’ and g’‘’. In this way you can use the original 8’ scaling with no issues whatsoever!

1 Like
4

I believe I tuned a Richard Kingston instrument which lacked a top 4’ note or two. It was not noticeable as the 4’ became a less obvious part of the timbral envelope of the top notes.

1 Like
5

Thank you Claudio.
Of course I have read your valuable article, it’s at it I was referring to when I wrote:

anybody has such a table? maybe Claudio who has written an essay on Taskin’s scaling and stringing?

I have downloaded it on my computer as a reference.

This is brilliant. As Ed Sutton says, those two notes would be scarcely audible, so no big loss. Moreover, it’s not unhistoric: some harpsichords lack the treblemost 4’ notes. The triple-manual Hass lacks more than one octave of the 2’.
I’ll check to see if there are any drawbacks other than the space for the 4’.

Do you think the displacement of the bridge as is would audibly impair the sound?

6

Of course I am aware Claudio’s suggestion of eliminating the top two notes of the 4’ will need a lot of work:

  1. If I relocate the bridges, plucking points could be modified to a possibly non-negligible amount. To avoid this undesired consequence, nuts could require to be slightly relocated as well.

  2. The 4’ hitchpin rail would be relocated as well to work as a htichipin rail AND a boudin/contrechevalet. Since it’s ends are housed in the liners and the curve itself will be slightly different, this means I’ll have to make a brand new hitchpin rail. If the ends end in different locations on the liners, I’ll have to:

  3. plug the housings in the liners and chisel new housings.

  4. Of course I will have to calculate the exact position of each and every bridge and nut pin and I can’t use the drawing as suggested in the manual (cut the soundboard part of it and lay in place in the case, stab through the drawing in the bridges and in the nuts).

Now I ask: is all that work worth it, sound-wise or breaking-point-wise?

7

Dear Domenico

Yes, the final years of Zuckermann kit production in the early 1990s prior to separation was made by Atelier Marc Ducornet in Montreuil (Paris). Some clarification now: Please note The Paris Workshop is the trade name we chose for the entirely different kit designs produced by Atelier Marc Ducornet after the split with ZHI Stonington. The TPW kits are quite separate designs to the three levels of production of finished instruments from Atelier Marc Ducornet: The simple logo instruments signed “AMD”; the midrange Atelier Marc Ducornet instruments; and the highest level which are true Marc Ducornet instruments. The old Zuckermann Harpsichords Inc, of course, continues to trade from the same Stonington Connecticut premises under Richard Auber as Zuckermann Harpsichords International.

You are building a Zuckermann French Double VI. Although it is clearly recognizable as French, it is not based on the 1769 Taskin, nor indeed any one particular historic instrument. (As Claudio’s scaling data compilation shows, every surviving Taskin is different, so I’m unsure what following Taskin’s original scaling actually means.) Rather, it evolved by David Way’s gradual modifications of an initial design by William Hyman in the 1970s. The 1769 Taskin does not have a sloping nameboard, for a start. The scaling of the Zuckermann is slightly shorter in the treble, c’’ 354 compared to 357mm off my drawing of the 1769 Taskin. In the extreme treble, even shorter, f’‘’ 143 compared to 148 for the top note of the 1769 Taskin.

Sometimes it can be difficult even for a professional to determine what is important when working on an instrument. While it’s essential to retain curiosity, it’s easy to get lost in the details.

You have the parts—except bridge—and drawing to make an instrument of which the decent design works. I encourage you to keep your project simple!

Regards

Carey

8

Yes, yesterday night I realized I can’t move the 8’ bridge forward because it would end directly above the treble end of the 4’ hitchpin rail, which would deaden the sound. The hitchpin rail is to be housed in the cheek liner and I don’t want to relocate it (maybe housing it in the upper bellyrail, instead).

I am puzzled by this:

I was under the impression that the zhi scaling in the extreme treble is actually longer than the pythagorean. Are you saying the 1769 Taskin scaling is even longer? But then, what exactly DJWay mean when he writes “[…] the treble scaling-which is already distorted a bit to allow the instrument to extend to g"‘“?

I’d like to study a complete Taskin scaling (or Hemsch or other French harpsichords). By “complete” I mean all the notes or at least 2-3 notes per octave. Where could I find such scaling tables?

Thanks a lot Carey for your injection of realism. I was probably taking a dead way or at least a much worked-up way. Now back to my French Double VI.

(by the way: why did DJWay and the people at Zuckermann and the people at TPW put those two treblemost notes? If one wants to play that handful of sonatas by Scarlatti and Soler that request a g’’’ (maybe 5-6 sonatas? 10?), one can usually reshape the melodic phrase to avoid that g’’’. The only spot where I miss a g’’’ is the repeated g chords in the K 427. Oh well…)

9

You’d have to ask them. But I’d say they wanted to increase the general customer appeal and increase sales. Certainly not for historical reasons. Just their design decision, combined with the desire of people to have one instrument for the entire 300 years of repertoire, common at that time.

10

Le 19/01/2024 09:29, Domenico Statuto via The Jackrail écrit :

that handful of sonatas by Scarlatti and Soler that request a g’’’
(maybe 5-6 sonatas? 10?)

23 Scarlatti sonatas, according to my notes:

356
357
383
398
406
422
427
449
454
455
470
471
476
478
479
480
493
494
497
514
538
549
553

1 Like
11

Thanks Dennis. More than I thought but still not many.
I think you should add to your list:

K 380 (bar 55)
K 426 (bar 190, 192, 194)

12

I am surprised nobody as far as i can see has input all the Scarlatti sonatas into Music21 and produced data such as the ambitus for the lot. Once they are input this sort of analysis is trivial. it would be a useful and interesting resource.

13

Thank you Carey Beebe for your most informative details about the kit Domenico is now assembling. I fully agree with all your observations. Not being strictly based on the 1769 Taskin, it is pointless to try and copy its scaling, in many cases the differences will be audible but slightly so and certainly not detrimental, and surely Domenico would better heed your advice and keep things simple. A simple change here can bring lots of trouble there, as Domenico has already observed with the boudin.

14

Dear Andrew let me disagree. I have a kit up to g’‘’. Many of D. Scarlatti’s and A. Soler’s sonatas up to g’‘’ are true masterpieces, and in many of them reworking them to go up to only f’‘’ is, in my opinion, seriously detrimental to their musical value.
Modifying an early design just 1.5 inches in width to accommodate two more notes is not a problem at all .

[Note: needless to repeat, although Hubbard and others when they designed their extension to g’‘’’ enlarged the 8’ treblemost scaling to be able to insert that 4’ treble end, and in the process improving the sound quality of their steel stringing, at this point we all agree that this is just not necessary and the 8’ bridge can be leaved as is, just omitting the useless treblemost 4’ strings. With this proviso, the range extension up to g’‘’ produces no problem whatsoever.]

15

It was me who said that, not Andrew. Well, sure you lose something if you haven’t those notes in your harpsichord, but fact is harpsichords with g’‘’ are very few. 1 to 100, maybe? even less, probably.
As in conservatories and in private houses the “all-around” harpsichord is considered to be the standard French double going up to f’‘’, I think it would be a good thing somebody (you, Claudio?) would come up with viable workarounds to play those 23-now-25 sonatas. Let’s open another thread, however, as here it would be off-topic.
In any event, each harpsichord goes preferibly with its own music. If one has to re-design a French, Claudio is right: let’s stuff in all the notes we need, but without toying with the scaling:

Sure. Thanks again for your insights, even though I will probably leave the zhi kit as is, excepted than certain details such as the jackrail support/lock and others.

16

Forgot to say that I’ll use some spectacular brass pieces made by our David Law, instead of the pieces provided in the kit (which of course would be more than adequate);

  • case hinges;
  • the three stoplevers (they are cranked, David has done them beautifully in brass, I sent him the originals in iron);
  • three stoplever escutcheons (I don’t like the “naked” mortises in the nameboard);
  • two fancy brass screws to fasten the name batten (I hate the utilitarian look of regular screws and would better leave the name batten unfastened, just as in the Ruckers harpsichords and in most of the Frenchs);
  • the jackrail hook.
17

Extended compass.
Two of my late English doubles have had extended compass up to g’‘’ (g sharp if you include transposition) and it worked well. In order to keep the 4 foot bridge on the soundboard - which is desirable - I had to lengthen the scale of both the 8 foot and the 4 foot, but the 8 still works with P wire. The 4 went to steel for the top 4 notes but actually that was conservative. P wire might well have worked. The belly rail needed a bit more hollowing out too.

18

Dear Domenico

Those measurements I gave show that the Zuckermann scaling in the extreme treble is slightly shorter than the 1769 Taskin.

Many makers change the compass of an instrument to suit needs. EdS notes a Kingston he came across, and Huwsaunders notes the changes he made to his iron-scaled English doubles to accommodate additional treble notes. Those famous makers doing brass-scaled Mietke models with extended trebles either have to be content to do away with the top four or so notes and terminate their 4´ bridge short, or else be satisfied that the top few 4´ strings are going to be virtually equal in speaking length and require stringing in iron or even steel.

Regards

Carey

19

I have now both the Taskin 1769 and the Goermans-Taskin 1683-1784 complete chromatic scalings.
Many years ago, the Russell Collection website used to publish very detailed data sheets of their instruments. By “very detailed” I mean even the length of the sharps or the distance of the rose to spine and so on. The Goermans-Taskin data sheet page had a long report on the soundboard restoration. I have not been able to get them again, they seem not to be found again (if I am not grossly wrong).
I managed to reconstruct what the URL was in 2004 and then used the Wayback Machine (glory to them!) to actually get all the relevant pages again. Of course if you need them just ask.

As for the actual data, the current page for the 1769 gives only the C’s and F’s. Some of them are wrong. The right ones are in the old data sheet, which by the way coincide with the numbers Claudio gives in his Taskin Stringing and scaling article.

In the construction manual the Taskin 1769 is many times said to be the model for the ZHI French double, that’s why I keep confronting.
As for the relation between the Taskin 1769 and the ZHI French double, I’ll need to double check my measurements on the drawing. If I have measured everything right, the ZHI French Double is in the range from -6.08% to +2.27% of the corresponding note of the Taskin (8’), and in the range from -3.23% to +8.24% of the corresponding note of the Taskin.

I have made a graph. In the 8’, the Taskin has a smooth line, the ZHI has two “bumps” (very abrupt, so I believe it’s a meaurement error on my part, I’ll check). Here is the graph.

GraphTaskinZHI-8
GraphTaskinZHI-81371×946 134 KB

Overall, the ZHI is shorter than the Taskin along the entire compass excepted the very first 6 lower notes.
The difference is less than 1% only from e to b.

As for the 4’.
Again, the Taskin has a smooth curve, while the ZHI has “bump” (albeit less pronounced than in the 8’).
Here is the graph.

GraphTaskinZHI-4
GraphTaskinZHI-41374×956 106 KB

Here the ZHI is longer than the Taskin 1769 for most of the compass, not only in the bass but in the tenor-treble as well. Range is from ZHI being -3.23% to +8.24% of the corresponding note of the Taskin. The treblemost 23 notes are all but one more than 3% longer, and 13 notes are more than 5% longer.

I am perplexed (and I am starting to think I am giving too much attention to scaling).
There is a potentially dangerous spot at the treblemost 5 notes of the 4’, as the “equivalent scaling” there is 371, 378, 384, 396, 408 mm, well longer than the usual 357 or so. I am aware of the drawing hardening effect, but then the Taskin scaling in the last three notes (remember the ZHI has a f sharp’‘’ and a g’‘’) is 357, 365 368. I now am afraid my strings will break up there.

2 Likes
20

If you have read my previous post by mail, last sentence was:

“Good news are that there is no dangerous spot for historical (or “historical”) wire, so I’ll live with that.”

Please be aware I edited after the 10-minutes time so my edit is not showing in the mail. My last sentence is just the opposite:

“There is a potentially dangerous spot at the treblemost 5 notes of the 4’, as the “equivalent scaling” there is 371, 378, 384, 396, 408 mm, well longer than the usual 357 or so. I am aware of the drawing hardening effect (tensile pickup), but then the Taskin scaling in the last three notes (remember the ZHI has a f sharp’‘’ and a g’‘’) is 357, 365 368. I now am afraid my strings will break up there.”