Variability in design of old clavichords

This article originally appeared as ‘Workshop News’ in December 2007

One of the surprising things about clavichords – especially the ‘historical’ type – is how very much they vary. There seems to be no such thing as a ‘typical’ clavichord; two instruments that to the eye look similar may be utterly different in sound and in the way they respond to the player’s touch.

At present I am overhauling a clavichord for a client in London; it is a diatonically fretted instrument made in the 1970s by a respected maker – the craftsmanship is of a very high order – and (as far as I know) it is a close copy of an antique instrument. The owner complained that the bass was weak, and asked me to do something about it.

At first sight this instrument resembles countless other fretted clavichords made during the seventeenth and eighteenth centuries. I have dealt with – and made – several such instruments and I didn't think it would take me long to sort it out. It was made at a time when it was common practice to string clavichords in much the same way as harpsichords; nowadays we understand that the struck string requires a generally higher tension, since strings that are too slack produce a feeble tone, and the touch is soft, making it hard to control the pitch of the note. In particular, the tension has to rise in that part of the compass where the strings are hitched close to the back left-hand corner and the overlengths are at their greatest. As for the open-wound strings in the bass of the instrument, I suspected that their coverings were loose, hence the almost complete absence of any sound.

Accordingly I calculated what I thought would be a more appropriate set of strings, including new over-wound strings for the bass, and put them on. In much of the compass, there was an immediate improvement, but the bass notes were still disappointing, and I was surprised to find that the touch throughout was still ‘spongy’.

It is from the unexpected result like this that one can learn. On looking more closely at the instrument, I realised that the balance rail (which holds the balance pins, and is where the keylevers pivot) was quite far back, particularly at the left (or bass) end. This is not something that you would notice at a casual glance, but it is a crucial aspect of the instrument’s design, affecting the key-leverage ratios. Now, the leverage ratio for each key is calculated by comparing the distance between the pivot-point and the place where the player’s finger touches the key – call this ‘A’ – and the distance (measured perpendicularly to the case front) between the tangent and the pivot-point – ‘B’. The leverage ratio is B divided by A; if it is greater than 1, the tangent will move faster than the player’s finger, and generally this is the case, since we need a fast-moving tangent to create a decent note. Typically the leverage will lie somewhere between 1.5 and 2 for all but the lowest couple of notes.

On this clavichord, however, because of the position of the balance rail and the length of the keys in front of it, the leverage ratio for the whole of the bottom octave – twelve notes – is actually less than 1, meaning that the tangent moves more slowly than the player’s finger. The only way of getting a decent sound from the bass, and an acceptable balance between bass, middle and treble, is to string the bass even more heavily, and once this was done the improvement was dramatic. Luckily the case was so well made that I have no fear at all for the safety and stability of the instrument!

So, what seems at first a simple and even a rather arbitrary decision by the maker – it may be, to move the balance rail back 10 mm or so – implies a certain kind of stringing and set-up, and has dramatic effects on the way the instrument plays and sounds. There are advantages to this design: the instrument responds very effectively to finger-pressure, so that one can employ all the skills of clavichord technique such as bebung and tragung der töne to good effect. On the other hand, the somewhat heavier stringing means that there is less ‘natural’ sustain, and the player has to work a little harder to draw out the notes (when sostenuto is required).

One could say more about the detailed characteristics of this and other clavichords, but the point to emphasise is how much they vary, even when the designs come originally from the same period and place and look very much the same.

To return to ‘Essays’, click here.

To return to Contents, click here.