Dolceola Restoration

So, a Dolceola. Well, it doesn't look to be in TOO bad a shape, right?

Ha! To quote the teachings of Waylon..."WRONG!" The keyboard part was actually in good condition, but the zither part was a MESS! Oh well, it's almost a hundred years old and it looks like it's been thrown out the door after its owner a few times. My sense is that this latter wasn't the Dolceola's fault, so let's see if we can bring the poor old thing back to life.

As our documentation opens, all the internal work has been finished and the box is back together. This involved making new braces, a new back, and 3 of the 5 pieces of the frame. Let's start with a brief review of the damaged parts, in no particular order. First up, here's the short side of the frame. As you can see, it was a few grain lines from the two big cracks meeting in the middle and leaving it in two pieces.

Here's the head end of the long rail. Hey, isn't there supposed to be a tenon there?!

Despite its mass, the tail block was busted like this in a couple of places.

Here's a shot of the inside of the old back. Yes, the large script lettering that runs the length of middle piece does in fact say "Honeydew". It’s a piece of a melon crate! In all fairness though, it looks like it's probably not original, and for that matter it's not a particularly unacceptable material for the purpose, but it's still entertaining. As far as damage, notice the crack at the far left; that's where the braces had come through the back and were seeing daylight.

Speaking of braces and seeing daylight, get a load of this. Here's one of the braces, against a steel straight-edge. That's hard maple, and that's a lot of swag.

The bridge suffered from missing wood in nearly all of the chord groups. The pins slant toward the direction of the lateral force of the strings, and there is no reinforcement under the bridge on the inside...the bridge pins go through the top and that's IT. The missing wood would leave the pins free to move at the ends that show, pivoting at the point where they go through the top...gee, that wouldn't crack the top...would it? (And yes, it did.) Needless to say, I reinforced the area under the bridge before putting it back together.

But the bridge had another just-as-serious problem. It was separated at the bend.

Doesn't seem serious? That's what I thought too: "Big deal; I'll squeeze it back together." Again, enter Waylon. Somehow it was badly twisted...I mean as badly as I am for tackling this project! Getting the two laminations to meet up required misaligning the two halves by about 3 inches at the end.

After considerable thought, I pronounced the old bridge unsalvageable. It just made more sense for the long haul to make a new one. So I planed up some 1/8" x 5/16" x 29" hard maple strips, and went about the matter of trying to bend them.

I tried s-l-o-w, patient bending with dry heat, using a fully regulation fiddle rib bending iron, and even sandwiching the strip between that and a hot flat-iron. 20-30 minutes later, I'd gained about 5 degrees of curvature, and a few minutes after that, that sickeningly as-irreversible-as-pulling- the-trigger sound..."plik". #$^&*(%@!!! Yes, I broke it. It became evident that bending 1/8" thick hard maple required different tactics from the dry heat method used to bend c. 1/32" thick fiddle ribs. So...off to the web to learn the finer points of bending wood. Following two or three sets of directions which consistently outlined the general direction I needed to head, here's my new steam box!

Yes, that's a 5-gallon galvanized gas can, a Coleman camp stove, some radiator hose, some PVC, and some dowels that form a rack for the wood to sit on. I think you can imagine how it works...it just makes lots of steam. The one detail that's not so obvious in the picture is the 1/2" dia. vent hole (red arrow) where the steam (faintly visible) comes out. The strips were inside it when this shot was taken.

So does this thing work? I admit I had my doubts. But I have to say, it appears that it *does* actually work. After an hour in the sauna, the strips of hard maple (determined by the USDA to be THE WORST wood for bending) conformed to the ready-and-waiting mold quite nicely. I let it spend the day in the sun to dry, then glued it up. After a day in the mold, I took it out, wondering if it would spring back out as straight as an arrow, or at least enough to really mess things up. But no, I was pleasantly surprised; it didn't straighten out...any, at all. Here's a shot of it in the mold, just after bending.

Now as for the dolc. proper, the body is all back together. Here's a series of shots in which I took a trip around it, showing the new frame pieces and joints. Glimpses of the new braces are visible too. The tail view from the treble side...

...from the bass side...

...and a head end shot. I think that shows all the joints, so you can see how the thing is framed.

Here's a shot of the new back. It's 1/4" poplar, same as the original. I haven't planed it yet (or "cleaned it up", as the aulde-time carpenters would have said), but that'll only take a minute; not a big job.

The top of this instrument presented some interesting challenges. In the following paragraphs, I refer to the image below.

Mainly, there was the big V-shaped shrinkage crack down the middle (black). I thought I would just have to plane the sides (it's all but impossible to do a good job of fitting a piece of wood to a natural break) and fill it with new wood. As always, I didn't like that idea because (a.) cracks are never straight, neither along their length nor from the outer to inner surfaces of the wood, so it always requires enlarging them to make them straight, and (b.) new wood never matches old wood perfectly. But over the course of my multi-volume effort to try and flatten the top, all the joints had let go. One of them was near the bottom half of the big crack, just to the right of it (blue line). This resulted in the piece between the crack and the joint being detached and free from the rest of the top. (I should say from the rest of the pieces of the top, at that time.)

I realized that when I took this little piece out, the halves of the long top part of the big crack mated back together nicely. And when I put this piece against the shorter bottom part of the crack, the same was true. So I decided to plane the already-straight joint edge of this small piece at an angle that would make everything come back together (red dotted line). Both the angle and the amount removed had to be perfect, so I took the precaution of making sure I was in perfect mode when I did it. (Actually, I settled for really, really careful mode.)

Here are a couple of close-ups of the results, from where the crack was worst. It doesn't necessarily look like it in the photos because of the old finish and excess glue, but the halves are together tightly. The piece of new wood in the image of the top part of the crack was to fill a chip, not the crack itself. Here's the scene at the top of the sound hole...

...and the bottom.

Now while this bit of plastic surgery did miraculously almost erase the big crack, it also created a couple of new problems. First, the top was now narrower at the bottom than when it left Toledo (white line). This is a big deal because the action has to fit, and its width isn't adjustable. So I had to add a wedge-shaped piece of wood on the long side. Here's a close-up of it.

Also, the tail edge of the top was now swallow-tail-shaped (yellow line). The drawing is exaggerated, but the actual measurement of the discrepancy was a full 1/8"...couldn't ignore that; it definitely needed to be reckoned with.

The fillet for the tail wire had to be re-cut. I got to use one of my old-time tools for this. (I love it when I get to use them!) I keep them in good tune and they work really well. I had one that was the perfect tool for the job, a "moving filletster". This plane has a skewed iron for cutting across grain, a little slicer that travels ahead of the iron to prevent the side of the fillet from getting all torn up, and both width and depth stops, to stop you when you're there. You can't go wrong with Movin' Phil! Here's a shot of this great old tool at rest.

...and in action.

Here's a close-up of his work. Not bad for a couple of minutes' work that involved no machines.

So finally, here's an unobstructed view of the top at this point. Looks like hell, doesn't it? Oh, I'm dying to plane it and get rid of that old finish, but there's work to be done before I get to do that.

As you can see in the image above, the area of the top around the chord string pins is a mess. The pin holes don't line up with those of the block, and lots of little pieces are missing. The wood of the top around all the melody string pin holes is wallered out too, but at least they are in alignment with the holes in the block.

So I undertook to inlay strips of new wood, one per chord, and re-drill these holes. To begin with, I chiseled away just down through the top for the patches.

But I knew I'd never hit every one of the 35 pin holes in the block dead-center, and that if I didn't I'd have egg-shaped holes in the block. This would obviously be bad; not only would it make it hard to keep in tune, but if the gap was on the bridge side, it would set the stage for the pin pulling forward and breaking out the little piece of wood between it and the next pin, like it did before.

So I filled the pin holes with dowels. I used epoxy for the adhesive, since it fills gaps. Once it was all set up, I leveled them off flush with the block. I figured this way it wouldn't matter (as much, anyway) if I missed the original pin holes in the block by a little with the drill; at least there wouldn't be a void down inside there. I tried a pin in a couple of the melody string holes, and the fit was too loose to suit me. I thought I'd better treat them the same way as well, while I was at it.

Finally, the patches are laid in. They fit tightly enough that I had to clamp them in, which yielded the hydraulic effect. I could feel them being supported by the cushion of (non-compressable liquid) glue, then finally it would find a way out, and sometimes it had some force behind it when it did. So yes, it looks like I got a little carried away with the glue.

Now...finally...it was time to plane the top. It was like Christmas...it seemed like I'd been waiting forever to clean up that old top, and then it was done before I knew it. I brought my jack plane to surgical sharpness and went at it. The first trip across did away with most of the old finish, the second got the rest. Then I "cruel lighted" it, as I call it. I turn out all the lights and sidelight the work with a single lamp, so it shows every ridge the plane has left. Then I start backing the iron up, go a trip across, then back it up more, until finally it's taking an almost imaginary amount of wood. I won't sand the top, or even scrape it. I saw some micro-photos once, showing wood that had been planed, sanded, and scraped, and it convinced me that wood dressed with a super-tuned plane is at its most reflective and alive-looking. This thing can use all the life it can get! Besides that, sanding abrades the soft grain more than the hard grain, leaving the latter raised, and it rounds off edges (at least when I do it), both of which I find objectionable. So I hardly ever use sandpaper on wood. Anyway, after about 5 long-awaited minutes of pure joy, here's what I had.

Now it was time to mount the new bridge and the keyboard action to the body and locate all the bridge and hitch pins. I forgot to get a photo of my set-up for this, but I included a rough sketch of it below. It amounted to a piece of .012" music wire, tied at one end to a large steel brick (about 20 lbs.) that sat at the head of the instrument on the bench. The wire ran the length of the instrument, off the tail end, and off the end of the bench. From it dangled another smaller steel brick (about 10 lbs.), to keep the wire tight and straight. I temporarily mounted the keyboard assembly (shown in semi-transparent "see-through" view in the image). Then I sighted from the tail end, aligning the wire with each hammer and damper and marking the locations of the bridge pins on the bridge and the hitch pins on the tail block. 46 keys/60 strings worth of this only took about...oh...forever! Well, it seemed like it, anyway. At any rate, locating both ends from scratch wasn't a quick little job.

Once that project was done, it was time for another long-awaited event, namely finishing. This included the new sound hole decal I had made in the computer from a photo of the old one and then bonded to foil. I admit I probably took the liberty of brightening up the colors a bit, though those of the old one may have been faded. But then, at this point, am I worried about absolute adherence to the original? Over half the zither body is new! On the other hand, I won't be refinishing the keyboard housing because unlike the instrument's body, it wouldn't improve it. The finish is in plenty good enough condition to keep, so it stays. The finished wooden parts of the housing are of a deep brown color. I was pretty careful in trying to match the color, and it came out pretty good. Here's a shot of the freshly finished zither.

The only close-up of the new decal that came out in focus was the one where the flash went off independently. Oh well, you get the idea.

Here's a stretch of the new bridge after finishing. I shaped it with the simplest of tools, just pieces of spring steel shaped into scrapers, and they worked great on the hard maple. I scraped in the trough for the bridge wire first, while the sides were still flat, then rounded off the corners. It's pretty well identical to the original bridge.

And here's a micro-inspection of the chord string pin patches.

Well, once the 180 pins were all in place, it was at long last time to start putting the wire to this box. Incidentally, as you may have noticed, I decided to wait and drill all the pin holes after finishing. It turned out to be a good plan. I used a new, sharp “pilot point" bit, and the holes are slick as a whistle...not a trace of tearing out around the rims, and no slop whatsoever around the pins. I also decided to drill them at a slant, like a "normal" zither. They were set square originally, and I didn't like the look of it. Does anybody know why zither pins are usually slanted? Is there some mechanical advantage to setting them slanted, or is it just for appearance's sake? Is there some mechanical disadvantage to setting them slanted? (Please say no!) Well, for better or worse, here's the first look at the Dolceola with the strings installed, showing the now-slanted tuning pins.

And finally, here’s a full-view shot of the finished Dolceola. This project consumed nearly every scrap of my spare time for two months, but I suppose having a nice, functional Dolceola out of the deal justifies this expenditure of time and effort. One thing’s for sure; I don’t ever care to see the inside of another one!

If you would like more information on the Dolceola's stringing configuration, go here.

If you would like to see and hear evidence that the Dolceola was not the instrument played by Washington Phillips, go here.

And for even more great info on both the Dolceola and Washington Phillips, visit Gregg Miner's Dolceola Pages.

Back to Restorations Page