Tuesday, June 18, 2013

Checchucci Guitar - A Vaulted Back Body


 I have finished assembling the side ribs and vaulted back of my model of the Checchucci guitar. It was a delicate operation. The ribs and back of the original guitar are very thin, about 1 mm. Although the original instrument is intact, I was able to discern its over-all lightness and to estimate the thickness of the back and sides while examining the restoration photographs. These include a view of the vaulted back with a thumb size hole. Other views show the top edge of the side ribs with the soundboard removed.




In my last post I mentioned that I band sawed the face contour and retained the waste edges of the mould block in one piece as I was going to need them for support in cutting the longitudinal profile of the mould. They also proved useful for preserving the contour of the side ribs once they were bent. Sue and were going to be traveling in Europe for several weeks in April. I had finished contouring the vaulted back of the mould but I didn't want to start assembling the vaulted back because I knew that I couldn't finish it before we left and I didn't want to leave it half finished. But I did have time to prepare and bend the side ribs. I clamped them into the waste edges of the mould and left them until we returned home and I finished assembling the vaulted back.

For the vaulted back I prepared rib stock of macassar ebony about 1.1 mm thick. I also made  2mm wide white/black/white spacers by laminating 20 mm wide holly and ebony stock and then slicing off individual spacers. The ribs were bent to the necessary contour and shaped on an upside down jointer as seen in the top of the adjacent photo. I secured the center rib with positioning pins at each end and then applied one rib after another, alternating sides.


Each newly shaped rib was held in place with clear plastic tape and push pins. Throughout the assembly procedure the vaulting always felt fragile and as I neared completing the vault I left the plastic tape in place.

The work went quickly, in part, because Checchucci designed and constructed the vault in such a way that the shaping of each rib was simplified.








MFA Boston Restoration photo
The ribs are laid out so that the center facing edge of each rib is a straight line. I suspected this when I saw a photo of the back vault on the museum's website. When I visited the museum I arrived equipped with a straight flexible plastic ruler that I laid on the vaulted back along each rib line to test my hypothesis. Although the vaulted back was imperfectly constructed and distorted many of the ribs exhibited the center facing straight edge that I expected.

I laid out the ribs lines on my mould assuming that every rib conformed to my hypothesis. Having now assembled the vaulted back and side ribs I don't think I was completely correct.



During construction the vault always seemed fragile but when I finished gluing the vault to the side ribs I was satisfied that once I papered over the inside of the rib joints it would be stable in spite of its lightness.

I was to be disappointed. When glue dries on a papered joint it not only pulls the two sides of the joint more tightly together but it also pulls the joint down toward the inside of the instrument's body. On a arched body like a lute bowl the curvature of the ribs counteracts the downward pull and the ribs find an equilibrium.

This was not the case with the Checchucci vaulted back. The vault does not follow a uniform curve in the lower bout but flattens out over the last three ribs. I built this feature into my mould with the result that, when the glue dried on joints that I papered over, the joint between the third and fourth ribs from the edge collapsed.




The interesting point is that the original guitar suffered the same fate. In the photo it appears as a very slight deviation. If you follow the line of the joint between the third and fourth rib and the joint between the second and third ribs in the lower bout you can pick out the collapsed section just to the right of the glare.




I noticed this feature during my study of the guitar in Boston but I assumed that it was the result of distortion taking place over the long life of the guitar, not an inherit design flaw. The amount of distortion on both my guitar and the original Checchucci was only 1 mm, but it was a noticeable blemish.

Fixing the problem proved to be troublesome. I removed the three ribs that were affected by the collapsed joint and began to rebuild the area. By trial and error, in attempting to fit a new rib in place, I found that I needed to prop up the adjoining rib as seen in the photo, by about 1mm. This was possible because of the flexibility of the vault and side ribs.  The procedure not only increased the curvature of the vault slightly but also changed the shape of the new rib from one with a straight edge to one with a concave edge. This created a stronger joint that would resist the downward pull after papering over the inside of the joint. I was able to accommodate the last two ribs by filing down the side rib by a millimeter. This allowed each rib a little space to cant over, creating a slight sideways arch to an otherwise nearly flat slope.







So is this a faulty design?  I think it is a design that leaves little room for error. In instrument making there is a fine line between being in balance and being out of balance and success often depends on a finding the elusive balance.




Building replicas of historical models adds another dimension. One that is seems to be measurable, ought to be, but isn't quite.


















Wednesday, May 8, 2013

Building the Jacopo Checchucci Guitar - The Mould


Jacopo Checchucci, MFA 2001.707, Boston.  Author's photo














Last April I measured the Checchucci guitar and posted a report on my work in this blog (April 25, 2012). During the summer I prepared a working drawing of the instrument and posted it on November 6 in a format that you can download and print. I always intended to build a model of this guitar and last month I got started. So far, I've built the mould, bent the side ribs and today I finished assembling the vaulted back.  Over the next weeks I will describe the techniques, discuss the problems and show you the results.

First the mould. I assembled a suitably sized block made up of  pine
and basswood boards salvaged from an old bench that I no longer needed. The template contour lying on the block is the outline of the face of the guitar that I made at the museum. I had taken particular care in making the template as accurate a representation of the original as possible.  To do so I removed the strings from the bridge and prepared a foam board with a recess cut in so that the guitar could lie flat on its face. A piece of card stock with a corresponding cut out completed the set up. I was then able to accurately trace the contour. I used the template to lay out the guitar body in my drawing and it serves again to lay out the mould.


As I noted in previous posts, the side ribs angle in from the face to the back in varying degrees, from 4 degrees at the waist to 2 degrees at neck and 3 or 3 1/2 elsewhere, as part of an intentional design.
However, I think Checchucci intended a more uniform outcome.  During assembly with the back and top the side ribs of any guitar have a tendency to splay out at the waist. This results in a greater angle, as was observed with this guitar.
 The neck joint area has distorted over time, accounting for the lesser degree of angle in this location. A crease in the smooth curve of the vaulting of the back behind the front block results from the upward pull of the strings.

Given the variation, I decided to make the mould with a uniform side rib angle of 3 degrees.

I band sawed the face contour and retained the waste edges of the block in one piece. I was going to need it for support in cutting the longitudinal profile of the mould.









Once I had the basic shape of the mould I started to contour the vaulting.

If it had been good weather outside I would have taken the mould out to the back deck, fastened it to a support and used a belt sander to make quick work of it. But, there was snow on the ground so I decided to work indoors. I used a block plane to form the vaulting, checking periodically with the templates that I made in the museum.





Once I was satisfied with the shape of the vaulting I used the cross-section template of the highest point of the vault to mark the location of the ribs.




For other locations on the surface of the mould I drew lines across the width at the upper and lower bouts and at the waist. I used the same number of ribs as the original so I divided the line at each location into the suitable number of equal parts. At each end of the mould I used photos to determine the most accurate termination points for the rib ends.

Stretching a thin, straight-sided  strip of wood across the surface of the mould and scribing a line took three hands. Sue helped out. The lines did not always intersect the points that I had marked on the mould, missing by a millimeter or so. In such cases I followed the line of the straight-sided strip of wood and allowed the rib widths to vary occasionally.

To finish the mould I planed facets for each rib with a miniature
plane and applied a coat of varnish to seal the wood and to help prevent the ribs from becoming glued to the mould.


Next time I'll show you the results of my work and describe what I've learned about the design of the vaulting and the technique for putting it together, but more importantly, how the two go hand and hand.

Friday, April 19, 2013

The Christoph Koch Theorbo Berlin



This past week I visited the Musikinstrumenten Museum of Berlin's Staatliches Institut für Musikforschung to study, photograph and measure the theorbo by Christoph Koch, Venezia 1650, No.3581.

This is an important example of a medium size, mid-century theorbo. Its construction combines features typical of early 17th century Italian theorbos with elements usually associated with later instruments. Also, the fingerboard, neck and extension are stunningly and masterfully inlaid with ivory arabesques, whorls and engraved figures.


I have been interested in building a model of this lute for some time. It is slightly smaller but with a flatter bowl than the Kaiser theorbo (Musee de la musique, E.24) that has proved so popular for me.  However little information about it, other than the essentials, is available. See Klaus Martius' lute database:


http://www.cs.dartmouth.edu/~lsa/associated/database/dbdetail.php?PID=228

Therefore, a trip to Berlin was essential.

I recorded details of the belly, neck and theorbo extension with measurements and photos which I will describe later in this post, but my principal interest was in the design in the construction of the bowl. This is the feature that tonally distinguishes one model of theorbo from another.

The bowl is built with fifteen ribs of kingwood (dalbergia cearensis) separated by 2mm wide spacers composed of narrow bands of ivory/kingwood/ivory. The rib wood is "flat-sawn" rather than quartered in order to display the its intense figure.
 Using the same technique, as I did with the Unverdorben lute in Dean Castle, Scotland (post from October 2,2012) I traced the face contour of the bowl on a paper grid and made templates of the central axis and five cross-sections.




The bowl is noticeably flattened as is the norm for larger Italian theorbos. It is constructed with wider ribs while its compatriots, the multi-rib bowls, are assembled from many more narrower ribs -- a technique that actually simplifies construction because narrow ribs can be bent side-ways to conform to the odd rib shape that results from flattening the bowl.


I noted the size of the end clasp and the width of the ribs as they disappeared under it.

Also, I measured the width of the ribs at the neck joint and the angle of the joint itself. So by combining the measurements from both ends of the bowl with the  positions of the ribs derived from plotting the cross-sections, I will be able to assemble an accurate mould of the bowl.





The soundboard is 375mm wide and 594mm from the end to the neck joint and is constructed from very fine grain spruce. The contour is pleasing; full in the lower third and sloping gracefully toward the neck. The triple rose whose size, top to bottom, equals one quarter of the length is placed in the third quarter of the total length. The border is a combination of thin ivory and kingwood (in places repaired with ebony) bands with a wider strip of ivory on the outer edge. However, there are two disruptive elements. The contour in the  lower third of the face is asymmetrical; the treble side contour forms a sharper curve than the bass side. And the bridge is off-center. This in itself is not unusual, but the placement is extreme. The total string width on the bridge is 161mm. The first treble string is placed 60mm from the center line of the bowl, leaving the last bass string significantly close to the edge.


 Whether this was by design or error is difficult to decide. The fretted neck has been shortened which would have resulted in repositioning the extension. Perhaps it was repositioned at the wrong angle, necessitating moving the bridge. However I could discern no marks on the soundboard to support this hypothesis. Presumably, Koch chose this placement expecting a particular tonal reward.
Contrary to the appearance of the quality of work in the photo of the triple rose it is well done. A patina, which appears to be an accumulation of wax has collected between elements of the design and in the incised areas obscuring the detail.
One of the features of this lute that separates it from its earlier predecessors is the noticeably cambered fingerboard. I couldn't remove the strings in order to measure the camber accurately nor could I take an adequate photo due to the glare produced by the strings, but I estimated the camber to be nearly 4mm at the nut, diminishing to nearly 3mm at the joint to the neck. To put this in perspective however, the neck is 108mm wide at the joint to the bowl and 91mm wide at the nut,

The ebony fingerboard with the inlay is thin (around 1mm) stretched over a cambered spruce (?) core. You can see that the whorl has been truncated at the nut leading to the speculation that the neck had been shortened.

The arabesques on the back of both the neck and theorbo extension are superb. This is the most beautiful inlay work I have seen. Usually I find such things too elaborate, but the lines are delicate and graceful. The veneer for the back of the neck
and the extension, front and back as well as the lute bowl is a light colored kingwood. The tiny gap that remains between the ivory and the wood veneer is filled with a darker colored mastic glue that accents the over-all design. I think the combination of these factors adds to a perception of lightness of design. As a point of interest, Joël Dugot, Musée de la musique, notes that this design is the reverse (dark wood inlaid into an ivory panel) of the one that Koch created for his archlute built in 1654 and conserved in Paris, E. 546. 
The theorbo head is constructed in the recognizable style that folds back on itself. The front is decorated with an attractive checkerboard. The unusual feature is that it contains nine holes. But that is not all.

There is also a cluster of plugged peg holes at the end of the peg cavity. The vertical black lines are not shadows but an inlaid design. The tops of these pieces are visible in the previously shown photo of the fingerboard inlay. 

An additional plugged hole is located mid-way along the length of the extension level with the inlaid emblem.
The top surface of the extension shows a restored area contiguous with the cluster of holes shown previously.
Obviously, this area had been used to expand the compass of the shorter neck.  Whether this represents a legitimate or an invented use could be investigated further.

Dr. Otterstedt lends a hand for a final photo
I have only touched the many interesting features found in this instrument. But my trip to Berlin and this report is only the beginning of a lengthy project to document and to build replicas of this important theorbo. I plan to report my future progress in this blog.

I would like to thank Heide von Rüden for her assistance during my study and Dr. Annette Otterstedt for arranging my visit and graciously agreeing to extend it for an extra day when I found that I needed more time to complete it.


All photographs by the author.
















































Friday, April 5, 2013

Jerry Willard's 1831 Panormo Guitar

Jerry Willard Photo

New York City guitarist Jerry Willard sent me his 1831 Louis Panormo explaining that it had serious intonation and tuning problems.  In spite of this it is a dynamic and lovely sounding guitar.
Here are two of Jerry's recordings with his Panormo.




Once I had the guitar in my shop I spent a considerable amount of time analyzing the problem before I decided what to do about it. It was complicated.



Intonation problems are not unusual for nineteenth century guitars as many builders did not adjust the design of their bridge to compensate for the intonation problem that arises because of the physical difference in each of the guitar's six strings. Generally, the stopping point of each string has to be positioned on the bridge so that the string length becomes progressively longer from treble to bass by a total of one or two millimeters.  Otherwise, the guitar will not play in tune. Panormo built a raised edge, a raised wood saddle, on the front of this bridge to provide a clear stopping point for each string but no attempt was made to "compensate" for the characteristics of string physics. This fact would account for a portion of the tuning problems but not for all of the false notes. 

Next, I turned my attention to the relationship between the length of the strings and the fretboard. I measured the the distance from the nut to the twelfth fret and compared this measurement with the distance from the nut to the bridge.  


For a guitar play to pay in tune the latter measurement must be twice the distance from the nut to the twelfth fret plus one or two millimeters. It wasn't. In fact, the distance from the nut to the bridge was less than twice the twelfth fret distance.  I was astonished. Legendary guitar makers don't make this type of error. This guitar could never play in tune.



 I continued my analysis using an electronic tuner. By testing one string at a time and muting the others I was able to get accurate readings free of string harmonics. The results gave me a clear idea of the scope of the problem. Not only did every fret position on every string sound slightly out of tune, which was expected because the string length was wrong, but some fret positions were worse than others. Using the measured distance from the nut to the twelfth fret I calculated the position for each fret and compared this with the fret positions on the guitar. I found many discrepancies of misplaced and even misaligned frets.


I wasn't happy with this discovery. The guitar appears to be in nearly original condition and I believe it is important to preserve original work. But there were fundamental problems with both the bridge and the fretboard that prevented the guitar from being used in professional concerts and recordings.

 There was enough wood in the peaked front edge of the bridge so I knew I could treat it like a classical guitar saddle and reshape it to a contour that would accurately compensate for intonation. As you can see in the photo, the side of the front edge of the bridge has previously been altered as a glue joint is visible where the peaked front piece joins the curved contour where the bridge pins are placed. Either the piece had broken off and been re-glued or some other remedial work had been done to it.















But re-shaping the front of the bridge would only  be useful if the fretboard were substantially altered. Since there were so many problems with fret placement I decided there was no other alternative than to replace it. The fretboard was 6.5mm thick at the nut tapering to 4.5mm at the sound hole and had been re-fretted with modern, medium size T frets. My plan was to plane the it down to a minimum thickness and then use a warm iron to loosen the glue and remove the remainder of the board without damaging the neck material or its finish. As I worked down through the board I could see that it had most likely been slotted for bar frets because the slots were shimmed with ebony slivers, as you can see in the photo. Original bar frets are about 0.8mm wide and that is what the slots measured when the shims were removed.
As I continued planing I was surprised to discover that a very poor quality piece of Indian rosewood had been used for the fret board. Would Panormo have used such material? And would he have made such a mess of the fret placement? I would like to think not. But modern assumptions about historical instrument building are often at odds with the evidence. I don't have an answer but at least this is an opportunity to describe such findings and to ponder their significance.

Before I completely removed the fretboard I strung and tuned the guitar. I did this in order to acoustically determine the position of the twelfth fret. Using this information I calculated all of the fret  positions.
I prepared the new fretboard and made a caul that would fit under the neck so I could use spring clamps positioned on the edges of both boards with the neck and fret board sandwiched in between. Other clamps placed through the sound hole secured the end of the board.  I then cambered the surface of the fretboard to a 25 inch radius and hammered in the frets.  I use small modern fret wire that is intended for mandolins as it is about the same size as the various types of frets that were used in the nineteenth century.




In order to complete the alterations to the guitar I  reshaped the front of the bridge in a way that compensated for the physical characteristic of each string as I discussed earlier in this post. In the photo I highlighted the crown of the saddle with chalk to display its contour. The first and second  strings are closest to the front while the third abruptly needs to be longer. The fourth and particularly the fifth move back toward the front while the sixth is about in the middle of the range.

Jerry told me he intends to spend the summer preparing and recording with his Panormo. I'll look forward to hearing the results.

















Sunday, March 10, 2013

Building an Early Lacôte Guitar - Part 4

My replica
The original 1823 Lacôte has a mahogany body with black lacquered neck. The peghead, which is a later replacement, is a light colored wood with an ebony top and back plates. Although I built my model in curly maple I wanted to retain the black lacquer scheme for the neck. This made the finishing procedure a little tricky.
1823 Lacôte

You'll remember that in my last post I explained that the neck and body had to be assembled before I could complete the purfling because of the latter's delicacy. I knew these would create a problem in finishing because the guitar body, varnished with oil varnish, would share a joint with the neck that would be sprayed with black lacquer. I decided to proceed in the following manner even though the two finishes are incompatible.


When I build an instrument in curly maple I like to  give it a strong stain of roasted chicory tea. The maple for this guitar is actually English sycamore - I use the term "maple" generically - is quite light, so it needs some color, otherwise it apppears too washed out. I boil several teaspoons of chicory in a small container with about 60 milliliters of water for as long as it takes to produce the intensity of color that I want. I apply the stain in one coat with a new white shop cloth. Since the stain will raise the wood grain it is necessary to prepare for this in advance. After each sanding session I raise the grain by wiping the wood surface with a warm damp cloth. After the last sanding with #600 paper I scrape the area next to the binding with a thin steel scraper to remove any dis-coloration.
I have always varnished my instruments with oil varnish but it  will soon be impossible to buy oil based products because they are being phased out for environmental reasons. Last fall, a large old family run paint supplier closed and I was able to buy two gallons of varnish.
I find that oil varnish dries slowly on ebony, rosewood and other exotics so I seal these a with several coats of shellac.
To prepare the varnish I filter it through fine cheese cloth and add 5% thinner. Using a 1 inch wide natural bristle water color brush I cover a quarter of the back at a time using strokes across the direction of the grain and then brush out with the grain. In the photo you'll see that I masked the neck heel. Since black lacquer will be sprayed right to the edge of the varnished side rib I made sure the mask was accurate and firm.

 I sand lightly between coats with #400 paper or #1800 Micro Mesh. Heavy brush strokes or sags are removed carefully with a sanding block. Three coats will usually be enough. This is left to dry thoroughly in a cheese cloth tent. The maple dried well but even though I sealed the purfling it proved to be absorbent and needed an extra separate coat of varnish.


I had removed the masking after the last coat had dried over-night but I waited three or four days for the varnish to harden. Then I accurately masked the side ribs and protected the rest of the guitar from over-spray. There is always the risk of bleeding under the mask with sprays so I sealed the joint with clear lacquer by spraying several light coats (misting) right at the joint. If this does bleed, the clear coat will protect the under-lying varnish and make the clean-up easier. I proceeded to spray light coats of black lacquer at about 15 minute intervals until I had built up a thick enough finish.

The peghead of the original 1823 Lacôte had been replaced by one with side mounted tuners. The peghead core was capped top and back with thin plates of ebony. A maple like wood that had developed a beautiful patina  contrasted with the ebony. Chicory stain when heavily applied will collect in the pores of the wood and create an unattractive appearance on broad surfaces. But on small areas like those visible on the peghead it has an antiquing effect. I followed the same procedure in masking the appropriate areas of the peghead as I did around the neck joint. I finished the lacquering with several clear gloss coats. After allowing this to harden overnight I fitted the elegantly simple Rogers tuners and strung and tuned the guitar and then allowed the finish to harden for a week or more.

I varnished the soundboard of the guitar separately from the back and sides and the neck simply for the reason that I find it less cumbersome in handling the instrument.  While three coats of varnish were necessary to cover the back and sides I think that is too much finish on the top. Some makers would criticize me for varnishing my tops in the first place but soundboards need protection and I believe this can be achieved without inhibiting the response of the top. I thin the varnish a little more than 5% with solvent. This allows the varnish to penetrate the soft porous spruce enough that it doesn't lie on the surface. It also makes it easier to brush around the bridge and embedded frets.

The second and last coat is thinned by 2-3%. In comparison to the varnish on the back and sides, the two thinned coats creates a thin flexible film. If my hand is steady and my eye accurate, I can apply a smooth even coat in spite of the obstacles the bridge and belly frets present.  When all of the finishes have hardened I rub them out with rottenstone lubricated with mineral oil using a firm cotton ball wrapped in cheese cloth.

Pat Bianculli's original 1823 Lacôte is a beautiful, lovely sounding guitar. I enjoyed every minute that I spent building the replica.