felt and striker springs

Some incremental progress since last time...  Here, inserting the guide-nails into the keylevers, which the strikers ride upon.

Same glue-nail-heads-into-place procedure as with the pedalboard, just at a smaller scale!

Making the felt for the lower keylever travel stop.  I ended up using two layers of my fairly-thin felt here.  This felt pad quiets the key on the return stroke; if the finger is snapped off the key rapidly enough, the key accelerates under gravity until hitting this felt, so there can be substantial noise; but still, the upper travel stop is more critical (or at least, it receives more force): the upper felt terminates the "power stroke" which actually launches the striker.  It "feels" like the abruptness of impact with the upper stop somehow plays a part in imparting speed to the striker; but according to my understanding of the laws of physics, the striker will move just as fast even if the keylever is brought to a stop slowly and smoothly (and quietly) by a thick layer of felt.  I will "intuitively" believe this only once I actually feel it under my fingers...

These little felt pads on the keylevers under each striker are a crucial element of my design idea.  This is what is supposed to damp out any tendency of the striker to bounce off of the keylever and return to hit the string more than once.  Piano and piano-forte actions use all kinds of complicated catch mechanisms to prevent a rebound of the hammer, while also keeping the hammer near the string, ready for a quick repetition.  If my trivially-simple felt-and-spring system can accomplish a similar task, I'll be thrilled!  Only by actually building this action, and playing it for a while, will I really be able to tell whether the complex interaction of striker mass, lever lengths, spring strength, key travel distance, striker flight distance ("regulation gap"), and felt absorptive and frictional properties, leads to success (which is defined as that highly-subjective, "I'll know it when I feel it" quality known as "playability").

I was planning to glue the springs into the keylevers, but they jam into the holes so nicely that I think no glue is needed; and I may decide to change to a different wire guage to change the spring strength, so I keep my options open.  Indeed, the best way to remove the strikers, perhaps to sand them more/differently or to add felt, might be to pull out the springs from the keylevers; otherwise, the springs would have to be bent beyond the linear domain to get the strikers off the long guide nails, necessitating a tedious re-bending of each spring to get back to a good action again once the strikers are replaced.  Phew, it's tedious even to write about!  Anyway, so I think removable striker springs is a necessary feature of this action design, which I'll carry over to the big instrument (next project).

The strip of wood in place at the top of the guide nails is not what will really go there; I'm just using the damper spring carriage, since it has the right-sized holes at the right spacing, to stand in for the upper travel stop.  This has allowed me to test the action, ever so crudely, by holding the zither upside down above the strikers.  (Eventually, the zither will rest in a cradle which holds it at exactly the right distance.)  Another test is to hold a finger just above the key-down rest position of one of the strikers, and then hit the associated key.  Although the striker moves so fast that it's hard to see it leave its felt pad, one can feel the momentary touch of the striker, even for very soft strokes on the key, when the fingertip is hovering 1/16" .. 1/8" above that rest position.  These are all good omens...

from the notebook

I don't have nice PostScript drawings for this, but my notebook drawings are probably helpful, so here are the relevant pages so far.

adding a keyboard to my little zither

I've recently become interested in a class of hammer-action string keyboard instruments, which can be called "dulce melos".  The term, which simply means "sweet song", is the ancestor of "dulcimer"; it can apply to both of the types of dulcimer frequently seen today: the small lap-style or "mountain" dulcimer, often with 4 strings or string pairs; and the larger hammered dulcimer, usually played horizontally from a standing-up position.  But also, "dulce melos" has referred in ther past to stringed instruments with keyboards, and in some cases, hammer actions.  So I have adopted the term, to refer to my novel hammer action design, in order to emphasize that it is very different, not only from the modern piano, but even from any of the early forte-piano designs with which I'm familiar.  The key difference in my design, other than its utter simplicity, is the presence of a spring between the "hammer" (or I call it the striker) and the keylever.  In other designs that I've seen, the hammers move ballistically and return either by gravity or perhaps by various spring
mechanisms which are attached to the instrument body, not to the keylever.  You'll see what I mean in the pictures, I hope...  I often make detailed engineering drawings but I don't have a good one for this project,
yet...

Anyway, so in due course I plan to build a large instrument using this new action design.  Maybe 4 octaves or so, double-strung, grand (horizontal) format.  But before I do that, I want to build a prototype to test out and debug the new action, and start to characterize parameters of playability and feel so that on the "big" one, I can hopefully make something that is actually usable, on the first try.  So that's where my little zither comes in...

For years I've wanted to somehow put a keyboard on this little zither.  Most harpsichord-like designs require significant space "above" the strings, which would be below in this case, if I picture a mechanism which faces the zither and mates with it (yes, instrument-building is very erotic).  There's not much space between where I'd want to pluck the strings with a harpsichord action, and the plywood face of the zither.  A very miniaturized version of the action would be needed, and all just to put keys on this "toy" zither...  So it never happened.

But it occurred to me, this dulce-melos action could be designed to only "come from below", even the dampers.  And since I need to build a small prototype anyway, I will finally have, not only a keyboard on the zither, but a damper and "pedal" effect as well.  Can't resist that.

It's also kind of an interesting problem, because the zither is diatonic (7 notes per octave, 15 notes total).  So I've come up with a keyboard layout, similar to the regular 12-tone layout but 4 "naturals" around a group of 3 "sharps".  You'll see when we get there...

Here is my first duck-taped-up prototype of the dulce-melos idea.  It's a bit hard to see...  My "keylever" is balanced on a Sharpie; you push down on the right end and the left end see-saws upwards until the right end bottoms out on the felt; this throws the striker, the little irregularly-shaped scrap of wood on the left end of the lever, upwards a little further than the key itself travels, then the striker quickly flips back down, being on a spring (steel wire, taped to striker and keylever).  Felt in between the striker and the keylever damps out its motion quickly, so that it does not bounce or oscillate enough to hit the string a second time.  By holding the zither upside down just above this contraption, I was able to find the sweet spot where the striker hit one string at a time, and I could verify that even with this very crude version of the mechanism, it was able to repeat quickly without audible second-strike, and it readily gave access to a full range of piano to forte dynamics.  Very exciting!  Look ma, no escapement mechanism...

So, to the University Bookstore to purchase a bunch of their basswood strips, 3/8" x 1/2" x 24".  Works out nicely to make the keylevers 8" each; turns out the width of the full keyboard is exactly 8" as well, once the allowances are put in for space on the ends, so all the main basic wood pieces of the action are 8" long.

The first phase has been a whole lot of fabrication of parts.  There are 15 keylevers, 15 strikers, plus 3 different rails (others to come in the future).  Multiple holes of different diameters to drill in each piece, including countersinking and holes which end at a known depth (i.e., need the adjustable stop on the drill press).

For much of this, I used a "fence" to determine one dimension of the hole positions, and then carefully lined up markings on the wood to place the other dimension.  I drill the holes first, before cutting up the 24" basswood strips into their final pieces.

Trying to make my portable jigsaw behave like a saw that cuts straight lines, which it normally is not...

Cutting the notch in the end of each keylever for the guide nail.

And here is drilling the long holes into the ends of the strikers, for the striker guide nails.  This project very much depends on a drill press or other means to drill straight holes, it's probably obvious.  (The camera flash washes out the visibility, in later photos I've turned it off.)

Here you can see a couple of strikers positioned on their guide nails, on the keylevers.  The strikers slide up and down on the guide nails, and long straight wire springs (not fabricated yet) hold them down seated against the keylever, when at rest.  The springs also keep the (rectangular section) strikers facing the right way.  Initially, the strikers will be bare wood.  This is how pianofortes started out, too...  Perhaps I will decide to cover them with felt (and maybe in the larger dulce-melos and not the zither).  This is one of numerous things I want to experiment with, on a working model.

As noted in my notebook, I discovered a problem upon test-assembling a few keylevers with the balance rail, for the first time.  The levers were too tippy, left to right.  Once keys are attached to the ends of the levers, especially with the wide natural keys applying torque to the sides, the motion would be unacceptable.  I had thought that the guide nails at the back of the keys would keep them up straight, but given the crudeness of my slot-cutting technique, these guides are very loose.  And I don't want tighter guides anyway, I want a free and fast action, responding only to gravity, with no intrinsic friction, to the extent possible.  So I realized, an axle through the center pivot-point would probably hold the keys with adequate firmness while also minimizing frictional effects.  It means the keys are not so easy to remove, although unlike my pedalboard it *is* possible to slide the axle out to free up the keys for removal.  After drilling the keylevers for the axle (which is 12 guage steel wire left over from the pedalboard project), I took care to ream the holes out just a bit using a scrap segment of the wire, so that while being a close fit, the holes allow the keylevers to swing freely.

Here's how the axle fits through the four eye-hooks.  These eye-hooks fit in between the keylevers when the whole thing is assembled.

I was planning to use the same wood glue which holds the rails to the underlying plywood "floor", to seal the nails of the balance rail and the guide rail into place.  However, I need to assemble the keys to the balance rail first, with this new axle in the picture, and since the nails were a somewhat loose fit, I tacked them into place with superglue.  I've used superglue on wood before, with adequate results, but this time I have to say I don't like how quickly and voraciously it wicked far into the basswood, and mostly away from the nails -- though it seemed to do the job enough for this purpose.  I think I'd use wood glue for all such tasks in the future; maybe, also, the "gel" type superglue would have been better; I used the plain, extremely low surface tension, liquid type.

Ta da!  So that's what I've got so far.  Keylevers, balance rail, and guide rail.  Practically done!  (Not.)