the frame takes shape

I've started to lay out the design, on the bottom panel, similar to "Italian" construction (in that regard at least). I'm going with the two-line-segment bentside. I was going to draw the layout in CAD first, but instead I have ended up drawing it in real-life on the plywood, because that is how I can visualize it best: decisions about the angles and proportions, etc.. Then I'll have to carefully measure what I've drawn, lengths and angles, in order to do the lumber cuts; it won't be at all surprising if the measurements end up going into a CAD drawing after the fact: a little backwards, but what can I say!

(Just as an aside, I should note that I realized: I'm not doing something quite as offbeat as I thought, with the compass. I mentioned, I'm starting from E instead of the more-usual F; however, that is E based on A=440 Hz tuning, aka F in the "early music" community where A=415.3 Hz is the more-common consensus tuning. So, interestingly, I have found my way right to the compass which was often used, even though my materials (no brass strings) and design constraints (over 6' long OK, but no foreshortening) are different. Many times, instruments are made to transpose between the two by sliding the keyboard, but I'm not planning to do that with this one, made only for my own "rock and roll" use, I'm just going with fixed A=440 Hz tuning. My harpsichord will have a transposing keyboard.)

I've cut the oak plank I plan to use for a wrestplank in two, glued it up, and that is now drying; I'm making a double-thick sandwich of the 3/4"-thick planks, final thickness 1.5". Given the density of the grain, I am letting it dry for a "good long time" before unclamping. Seems like it might take a while for the moisture to find its way out of there (or redistribute, at least). I haven't worked with oak or other hardwoods much...

Fortunately, in addition to my drill press, I have access to a circular saw, configured as a "chop saw". It is able to cut angles up to 45 degrees away from a right angle. A couple of the lumber joints I had planned exceeded 45 degrees, so I redesigned the frame to eliminate these. It is better without them.

Here I've laid out the cut lumber, on top of the plywood back. Nothing is glued or fastened yet; the lumber will be held together with countersunk screws, just in order to firmly clamp the glue joints. Same goes for the bottom plywood and the soundboard (also plywood): sheetrock screws will be used to clamp the surfaces in place for gluing. As opposed to the "go-boards" used in conventional instrument-building: other than avoiding visible screws, it's not clear to me that the go-board approach has any real advantage, and it basically requires a proper woodworking shop, with an assembly table and a matching surface on the ceiling for the go-boards to brace against. Perhaps the go-boards (especially with an assistant or two) allow the soundboard to be clamped into place more quickly, which is a critical consideration when using traditional hide-glue, because it sets so quickly. I am planning to use synthetic adhesive with a longer setting time.

Before I glue the frame lumber together, I will need to carve away part of the top surface of the interior diagonal beam, to leave the soundboard free to vibrate there. About half of the beam closest to the front and to the spine-side, will not be carved away, so that part of the soundboard will be "masked off" and prevented from vibrating.

Also, I may fit one more smaller brace diagonally across the long tail section, and this brace would also be partially carved away on top. In my original design for the frame, I had the shorter segment of the "bentside" continue past the junction with the other two beams, ending at the spine-side. That portion of the beam would have functioned as the smaller brace and would have been carved appropriately. However, I realized that the beam did not need to continue in that fashion for strength reasons; the structure as you see it should be more than strong enough to handle the tension. I hope!

Also, the "spine" and "cheek" pieces, i.e., the parallel straight sides, will need to be carved for the wrestplank, which will be inset approximately flush with the top surface of the beams.

Finally, thinner and wider softwood planks will run around the outer perimeter of the frame, extending mostly above the top surface of the beams (which will form the bed of the soundboard), and extending downwards enough to cover the edges of the bottom plywood panel. These planks could be used to hide the screws fastening the frame together, but actually I'll probably use screws to attach the planks as well.

frame and case design

The basic frame of the pandalon will be constructed of softwood 2x4s, creating a box structure along with a plywood back panel and plywood soundboard.  The plywood panels will form a sandwich, attached to the front and back faces of the 2x4s, except for strategic regions of the soundboard where the 2x4 underneath will be carved away, to allow the soundboard to vibrate freely in those regions.
Wider flat lumber will cover the perimeter outside the 2x4s, forming the sides of the case and adding additional stiffness.  The wrestplank will be formed of two oak planks, glued together to double their thickness.

I was originally planning to use a very simple perimeter outline for the case, a narrow triangle ("shard of glass"), with a straight side on the right, where most pianos and harpsichords have a curved "bentside".  However, given the high length-to-width ratio of my non-foreshortened, 4-octave design, it is desirable to truncate the acute angle at the tail end, which otherwise would extend, dramatically but inconveniently, more than an additional foot beyond the length needed to accommodate the longest string.  And given the desire to have the string endpoints located at a consistent distance from the places where they pass over the bridge (in order to obtain consistent down-bearing force for each string, on the bridge), it is convenient to have a shape to the "bentside" which is more concave than a straight line.  However, I'm still not going with a curved bentside; instead, my "bentside" will be composed of two line segments, allowing better string endpoints while still preserving ease of construction with my limited tools and shop facilities (which is one of my main design priorities; not only for myself, but to make these designs easier for others to build as well, if they should wish to do so).

I was planning to use long bolts and hardware through the 2x4s, basically pinning them all together such that the bolts would support all the tension even in the absence of glue (which would simply serve to firm-up the structure).  But given how strong this new formula of Elmer's wood glue seems to be, I think that as long as I can obtain good glue joints, I don't need to count on additional strength from any metal fittings.  Hence, fittings can be smaller, though still present: I will use long screws to firmly attach the lumber together and hold everything effectively "clamped", while the glue joints dry.

pandalon: hammered dulcimer plus keyboard

So here I will document design and construction of my first full-size keyboard instrument utilizing what I've called my "dulce-melos" hammer action.  This will be a 4-octave (49) note clavier, in "grand" format, i.e., strings horizontal, parallel to the keys.  It will be double-strung (two strings per note, both at unison pitch).

In my research, I've found that another name which has been applied in the past to instruments of this type, i.e., hammer-action keyed zithers, is "pandalon".  I adopt this name, because it has fewer syllables than other ways I've tried to name these instruments.  I purposefully steer away from anything related to "piano" and/or "forte", even though this is technically a pianoforte.  I wish to emphasize the difference between this dulce-melos hammer action, and the actions of any of the instruments which have been called pianos or pianofortes or fortepianos, since about 1700.  The sound of this instrument, if I am successful, should be a lot closer to that of the hammered dulcimer, than to anything which has "piano" in its name.

At the same time, I am also building a small 15-note dulce-melos action, as an attachment to my existing 2-octave zither; that project (see separate blog) will let me prototype and refine aspects of the action (i.e., the keyboard, and the mechanism attached to the keyboard), before building the action for the larger instrument.  I will construct the basic frame of the large instrument, or what I call the "harp", leaving a gap for the action.  The action will be a separate project unto itself -- and if I don't like the action, after I build it, I may choose to replace it wholesale, perhaps even with a harpsichord action.

Probably the compass will be E2..E6: i.e., the low E of a guitar, up to the e''' 4 octaves above that, often the highest note on electric guitars.  Many harpsichords, fortepianos, and other archaic keyboard instruments, tend to have compasses which begin on F or G, if not C.  E is less common.  However, I am a guitar player; and there are other benefits to this E-based compass, such as E2 is the lowest note of the average male voice (by no coincidence correlating with the guitar).  But the real reason is, I have determined that E2 is the lowest note I can reasonably get, using iron strings, without incurring inconvenient length to the body of the instrument (it will be rather long as it is).  Harpsichords reach lower notes in shorter spaces, by using brass strings, at least in the bass (and also, in most cases, by foreshortening the bass notes, but I am avoiding that entirely, as a basic design principle).  However, I want to use iron across the entire compass, so that I can fit an electromagnetic pickup to this instrument.  (I may also put piezo elements on it in different places, and mix the different types of signal in different proportions.)

If the sound and versatility turn out to be what I hope they may be, I intend to make this instrument one of my main melody sources, in the new type of music I am creating.  Bass and other sounds will be provided from other sources at the same time; in particular, often from simulated pipe organ played on my pedalboard.  So the relatively limited compass of this pandalon, the lack of deep bass notes, is expected to be compensated by the rest of the "ensemble"; instead the priority is for an expressive instrument with singing tone and lots of harmonics, to play the upper lines of the music.  I.e., a role very similar to the one which a guitar would fill: so again, the "guitarish" compass seems right.

One important element of stringed keyboard instrument sound to me, is the damper pedal.  I.e., it is important, both that individual notes have dampers available for articulation, but also that the dampers can be globally disabled, for that harp-like (or dulcimer-like) sustaining sound.  I differ with the current popular opinion on harpsichords, in that I think harpsichords should also have damper pedals; and I have some novel harpsichord action designs which provide this (and I am aware of the past efforts along these lines).  Damper pedals -- or other means to disable the dampers -- have always been a part of the pandalon tradition.  My pandalon will of course have a damper pedal.  As a practical matter and for portability, the instrument will be built as a flat "tabletop" unit, with a separate optional stand.  So, my plan is to make the pedal a separate, detachable unit which sits on the floor, connected to the instrument with a flexible bicycle brake cable.

In addition to the damper-lift pedal, there will be a number of hand-stops to modify the sound in different ways.  Basically, any kind of gadget which significantly changes the sound in some way which is likely to be musically useful or interesting, I want.  Likely candidates are some or all of: buff stop (leather or felt touches the strings close to the end, giving a muffled or pizzicato sound); moderator (cloth strip moves in between hammers and strings, giving a muffled tone similar to -- but different from -- the buff stop); sitar stop (small metal pieces move into position such that they touch each string near the end, giving a buzzing sitar-like effect); bassoon stop (paper or parchment touches strings close to end, giving a different kind of buzzing effect which is not much like a bassoon but still a useful tone); octave harmonic stop (curved wooden piece slides over and brings metal pins into contact with one of the two strings in each pair, touching it at the centerpoint, thus effectively fretting it at the octave).  Additionally, there will be stops or a selector lever to disable one of the two string choirs at a time (by muting farther from the end of the strings; both strings will still be hit by the hammer), for tuning or for different tone quality.  And, there will be a hand-stop to latch the damper pedal down, i.e., dampers always-off: the essential "pandalon sound".