top of page

The Nylon-string Harp Guitar


Excerpted from "Sixty Years of Listening to Wood Sing," by Michael O'Brien, 2019.

Introduction by Gregg Miner

        This nylon instrument is purely classical, sleek and understated.  It’s a modern take on the traditional Austrian Kontraguitarre, but intended for contemporary practice.


        It’s also multi-scale with an elevated fingerboard, side port, and cutaway.  The four sub-bass strings are scaled to accommodate the LaBella ten-string sets, which are widely available. Sides and back are curly Claro Walnut. The top is Engelmann Spruce, and the necks are straight-grained, quartered Walnut. Walnut makes excellent necks, although it’s heavier than Mahogany or Spanish Cedar. A continuous bone saddle accommodates all ten strings, and the bridge design, borrowed from Tom Bills, wraps the strings around the elevated portion of the bridge, back under the saddle where they are simply tied off.

















        The twin necks of Schrammelguitarre traditionally join the body with a single, large joint, but I dovetailed each neck separately into the body. The necks join at the guitar body and at the headstocks in a slightly “updated” way.  By configuring the tuners in a 4-2-4 pattern rather than the usual 3-3-4 the headstocks could be joined elegantly. On so many older instruments with sub basses the design of the headstocks is awkward, at least to my eye.  But that’s just personal taste.


        The building process was straight forward, and a lot of it was done by hand. The initial plan called for routing the neck dovetails on a jig, but the complications of devising one that would work on this body’s sloping shoulders made me change my mind.  I have a lot of experience with hand-cut dovetails, having made so many clavichords. I once hand-cut dovetails into the cheek/bentside joint of a solid walnut double-bentside German harpsichord for Albert Fuller.  Because of the geometry of the bedside curving away from the cheekpiece, a router jig was next to impossible.  Tom Wolf and I had a lengthy discussion about it, and in the end he let me cut them by hand. It was his name on the instrument, after all!  Honestly, I sweated bullets the whole time, but in the end they came out perfectly, all nine inches of them, and so far as I know, that instrument is still in Albert’s Manhattan flat, for all to see, some 30 years later. I even repeated the idea in the ceiling of my house, just for fun. So, I felt more comfortable cutting the neck dovetails by hand than routing.

nce cut, the necks were dry-fitted at the correct angles, and then joined at the body and headstocks.  The details and curves were cut with a simple coping saw and finished by hand. Eventually they were glued to the body with hot hide glue.













        A couple of years ago, I actually measured the commercial fret-slotting templates I had.  Using a rule that divides the inch into 100ths I found a few, very small errors.  So I went to laying out and cutting the fret slots by hand, and have never looked back.  Marking out and cutting by hand is admittedly impractical for even the smallest production shop, but I’m not interested in making production instruments, and there are many great shops doing that, making great guitars.  Besides, I really love drawing with a pencil, cutting with a hand saw, and slicing through wood with a good chisel or plane.  (I’m old-school and after all these decades I still love getting my hands on the wood itself. When one begins to notice that the light up ahead is growing brighter and the light behind is growing dimmer, one begins to reflect carefully on how to spend one’s time.)





















        For accuracy, I use a .3mm mechanical pencil sanded to a sharp point, like mechanical draftsmen used to do. The Starrett rule divides the inch into 100ths and with a magnifier it’s possible to eye-ball the thousandths.  (BTW:  there are many fret calculators on-line. StewMac’s is free and easy to use, so I make a print out of the calculations I want and go from there.) The line that I draw represents the center of the fret slot which is thus the center of the fret crown. It’s important to triple check everything, not only nut-to-fret distance, but fret-to-fret. Also. I mark the fret positions along the string path of lowest and highest strings (which spread), NOT along the parallel sides of a fingerboard billet or along a center line. By using the first tooth of a pull saw (which is hooked on most saws) to lightly score the saw’s path and to allow you to see the width of the actual cut, it’s possible to center the cut on the line, against a guide.  For depth control, I simply check each cut as I go rather than attach a stop to the saw blade (for better or worse, I have a fundamental distrust of gadgets). The whole process of laying out, checking and correcting for errors takes about 45 minutes. Cutting the frets takes about half an hour. I’m fine with that, but I’m running an artisan shop, not a production one.






















        The shorter e-string is 645mm or comparable to a short-scale classical guitar, and the long E-string is 655mm, slightly longer than “standard” classical scaling. The master fret is no. 5. So, again, the fan-fretting is fairly mild, laid out for ergonomic reasons not acoustical ones. 





        The top bracing is a basic Kasha-style pattern but modified slightly to accommodate the long, angled bridge and the body’s cutaway.  The back is radially braced. 


        The greatest challenge of a nylon guitar is getting the higher positions and upper strings to sing out. So I monitor the thicknessing process of both the top and back with a caliper, but I voice by ear.  


        I can’t give a recipe for this, but by really listening one can hear fundamentals and partials in the so-called tap-tones.  So I listen a lot for these, very carefully, if instinctively.  Since 1978 I’ve heard a lot of wood sing to me.  I don’t know a better way to express it.
































        The bridge of a guitar transmits vibrations throughout its entire length. It can only move in certain ways, either up or down (more or less as a whole), by pitch and by yaw, and very slightly side to side with the dimensional fluctuations of the soundboard. These movements roughy correspond to the monopole and dipole patterns of the top. The bridge does not, however, move along the length of the guitar since the top is virtually stable along the grain. That means that the bass end is not entirely isolated from the treble end — which seems to favor a single-bridge design for harp guitars rather than a multiple bridge design. Two or more bridges, vibrating independently of each other could possibly conflict. Many great instruments do in fact have separate bridges for the sub bass strings and/or super trebles, and they are truly great instruments. So I’m not entirely convinced by my own thinking on this. Still, every luthier has to make decisions, and I’ve chosen a single bridge design, and I love the homogeneity in sound I think it encourages.  Either way, God and the Devil both lie in the details.



































        Similarly, the wrap-around method of typing off the strings theoretically favors a more efficient transfer of energy from the strings through the bridge to the soundboard. And I used it on this guitar.



        I use a very traditional method of applying French Polish, with pumice, grain alcohol, button lac, droplets of olive oil during bodying up (which is spirited off), and in the later stages a very small amount of walnut oil emulsified directly in the shellack. Walnut oil is a drying oil and becomes hard over time, and since miniscule amounts remain in the finish, it adds a small measure of hardness. The whole process takes about 3-1/2 weeks with upwards of 30 applications. 





































        I like the color possibilities of button lac, although it has to be de-waxed. For this I strain the dissolved shellack through coffee filters, and to save time (which can take a day or two), I run several at once, and it does a more thorough, cleaner, job. Then, I adjust the color by adding activated carbon pellets to the de-waxed shellack and filtering it out after a day or so, depending the desired color. I’ve sometimes used super-blond flakes, but I prefer the unifying effect of a slight amount of color in the finish. Preparing the shellack takes about a week.

































        All in all, this kind of French Polish is a fussy, but very pleasant process. It can be almost zen-like. No spray booth, no spray guns or noisy compressors, no hypersensitivity to dust, no toxic or volatile fumes, and a finish of exceptional beauty.  It’s only real drawback is that a French polished guitar is not really suitable for gigging an open-air Farmer’s Market on a hot, summer day, when sweat runs and few people are listening anyway.  A factory instrument is probably better for this.




















































bottom of page