Old Project - New Lute

With many interests and projects vying for attention, I am reluctant to start any new instrument building work until I have finished what is already in hand.
This is a project that has been 'on the back burner' , gathering dust, for a few years now so I am hoping that by posting on this forum I may gain renewed motivation (and prodding from other friends on the Forum!) to complete the work. Although it is a lute that I am building, the way I go about it may be of interest to oud makers as construction methods for oud and lute are very similar.
I do not know how long it is going to take me to get into the project because, for reasons I need not go into here, I shall likely be (temporarily) visually unable to do detailed work for the next couple of months - but we will see. In the meantime lets get started.

The lute is an example of a "Liuto Attiorbato" - a small extended neck lute used from the end of the 16th C as primarily a solo instrument in Italy. There are a number of fine surviving instruments from the early 17th C that fit the description of a "Liuto Attiorbato". I am basing this instrument on a lute in the Victoria and Albert Museum in London, England,
(Cat# 7756-1862), by Christofolo Choc, Venice, detailed on Museum working drawings by Stephen Barber in 1978. The instrument had been converted to a wire strung instrument at some time in its past so is not in its original state and has some other peculiarities of detail that I did not want to copy. I am, therefore, only reproducing the overall geometry, barring etc. for this project.
The lute will have a total of 13 courses.With a stopped string length of 570 mm (six double courses with a single top string) and six single diapason or open bass strings measuring 839 mm in length.

The attached image is of a liuto attiorbato by Matteo Sellas, Venice, 1638 in the Paris Conservatoire, Musee Instrumental No E 1028 C 1052 which appears in the definitive work on extended neck lutes " Chitarrone, Theorbo and Archlute" by the late Robert Spencer, published in "Early Music" October 1976. For those interested in this important article, it can be obtained as a free download (courtesy Oxford University Press) at
http://www.vanedwards.co.uk/spencer/html/spencer.htm

I have now prepared some images in order to take stock of where the project currently stands after all of these years.

The belly is essentially complete and fully braced but is oversize requiring final trimming when finally fitted to the bowl (yet to be built).
It is made from two pieces of 'bookmatched' Sitka Spruce with a pronounced "Bear Claw" cross silk figuring. I purchased the spruce - in days when it was much more difficult to obtain luthier wood - from the Kirby Glider Co. in Yorkshire, England. They used the high quality spruce for building - well - glider air frames. They were a great company to deal with and very kindly allowed me to select material from their kiln dried stock of heavy timbers which measured about 9 inches X 4 inches by about 12 feet long (as far as I can remember). Only two pieces of timber were selected from their whole stock as being judged suitable for instrument soundboards. These timbers were then quarter sawn to required thickness at a local mill (a large power fed bandsaw was used for this operation)
The selected pieces for the soundboard were first 'jointed', in the usual way, using a 'shooting board' and jointer plane. After gluing, the belly was then thicknessed to about 1.5 mm thick using a smoothing plane (worked at an angle to the grain to avoid 'tear out') followed by final finishing with a scraper plane.
NO sanding allowed!
The barring is spruce - from the same stock - with the grain - unlike on ouds and modern guitars - traditionally running across the bar, not vertically.
I have essentially followed the barring arrangement of the Choc Liuto Attiorbato with the exception of the double bar across the rose.

Here are images of the rose and bridge.
The original Choc attiorbato has an inset wood and parchment rose that may be a replacement for an original rose carved directly into the soundboard in the lute tradition. Inset roses were used on early wire strung instruments so this alteration may have been done when the original lute was converted from gut to wire strings?
I have cut a traditional lute rose design into the soundboard in place of an inset rose. The soundboard, where the rose is located, is first reduced in thickness to about 1 mm, to facilitate the piercing process. A paper pattern of the rose design is glued to the underside of the soundboard and the front of the soundboard over this local area is varnished with shellac to harden the surface to make carving easier. The rose is then cut directly through from the back using a fine, razor sharp chisel with the soundboard supported on a piece of 'hardboard'. ( I will post an image of the chisel design that I use a bit later). This gives a sharp, clear cut definition to the rose design. Jewellers saws are not used. When the 'piercing' operation is completed the front of the rose is carved in relief with a scalpel. This 'brings out' the design of the rose.
The other image shows some detail of the bridge. This has been made from African ebony. The grey material coating the front of the bridge is dust accumulation of the past 10 years or so!!

Peg blanks for this instrument have been cut ready for finishing and tapering with a peg shaper. Made from boxwood - the original pegs were made from ivory!
I need a total of 19 pegs - more than enough here! (see separate posting on 'More Making Pegs' to see how these were made)

This style of lute has a significantly flattened profile - that is the depth of the body is less than half the width of the soundboard. This makes the construction of the bowl a bit more difficult than for a lute (or oud) which has a (more or less) semi- circular section because the ribs are not geometrically equal and symmetrical and do not run to a central point at the end of the bowl. For this reason, a solid mold was constructed from pine hand carved to the required geometry. I have provisionally marked out the rib profiles in pencil for a 15 rib construction - just to get an idea of the rib geometry. However, I may decide to increase the number of ribs dramatically (possibly up to 35?) and deeply scallop the ribs by hand scraping after assembly - just to try out the technique. After all, I am not trying to make an exact copy of an original lute so I am allowed some freedom of expression as to the detail of construction.

..... a slight diversion but, continuing with the question of lute 'cut in' rosettes, here are a few designs taken from surviving lutes (except for the Costa which I derived from a painting of a lute by that artist). To my eye, the 'Arabic' influence in many of these designs is apparent so I have to wonder if they originated from rosettes of early ouds which may have been 'cut in'? It is a shame and a historical loss that no early ouds from the 16th and 17th C survive to confirm (or deny) this speculation.
Next, some images of the tools that I use to cut lute rosettes.

Rosette designs for lutes were often 'standardised' - the same basic pattern appearing in the lutes of many different makers. This fact might suggest a 16th C lute component industry - soundboards made in workshops specialising in this product to be sold to luthiers for assembly into completed lutes.
I prefer the 'Gerle' pattern - it is an elegant, well balanced, design and easier to cut than the delicate tracery evident in the Tieffenbrucker/ Harton rosettes. I have chosen this design for my liuto attiorbatto.
As this pattern involves straight cuts, I use a small chisel, made from a broken hacksaw blade set into a wooden handle. This blade is ground thin and with a curved edge so that a partial cut is made first by piercing directly downwards with the blade point and then with a rotation of the blade. This method minimises the risk of breaking the fragile rose during the piercing operation - particularly at places where there is cross grain. I always work from the centre of the rose outwards. I do not bother to try to 'clean up' the cuts afterwards with needle files as that can spoil the clean cut definition of the rosette. The front face of the rose is then cut in relief using a scalpel. Tools must be kept razor sharp to cleanly cut the soft wood which must be kept well supported at all times with a backing surface of 'hardboard'.
It is not current practice to cut oud rosettes directly into the soundboard but a rosette made in this fashion could just as easily be made separately and fitted to an oud in the conventional way.

I found some more suitable rib material yesterday among my stock of woods. It is a piece of dense, very hard, close grained maple with a small pronounced 'bird's eye' figuring - measuring 1 3/4 inch square by 6 ft long. I decided to cut it on a bandsaw into 2 sets of rib blanks, 2 ft long, and managed to obtain 19 ribs - a generous 3/32 inch or so thick rough sawn and cut with the end grain at about a 45 degree angle. Cutting the blanks revealed a number of small "flaws" often found in maple due to the high sugar content of its sap that can cause staining during seasoning if not controlled. These can be worked around or bleached out if I decide to use the wood - so not a problem. Working highly figured maple is difficult because of the danger of "tear out" of the irregular grain.
This beautifully figured wood would likely be best used alone with ebony lines but without another contrasting dark wood
So this is another contender for bowl material. Decisions, decisions!!

Wow. That would look outstanding as rib stock. Thanks for sharing this. It's a huge inspiration to see it.

Now that it has been decided to make the bowl with 15 ribs, like the original instrument, the rib profiles must be laid out on the mold as accurately as possible in pencil. The areas between the pencil lines, where the ribs are seated on the mold, must then be made flat. This is not an essential operation but removal of the mold curvature at these locations should make rib fitting a little easier and more controllable.
First task is to locate the centre of the mold at the widest point. This was done by trial and error using dividers. At the widest point, the rib widths should all be approximately equal with the exception of the first and last ribs on each side (next to the soundboard) that are made a little wider than the rest. The positions of the edges of the ribs at the widest point of the mold was marked by 'walking' the dividers across the mold surface.
A flexible metal straightedge was made up with a hole for a screw on the centreline. This device was temporarily screwed to the pointed end of the neck block so that the straightedge could be rotated around the form of the mold. Pulling the straightedge tightly over the mold surface, straight lines were drawn through each point marked by the dividers to produce the rib profiles. Hopefully the ribs, when assembled to this layout, will have relatively straight edges also - which should help the rib fitting operation. Note that because the bowl of the lute has a flattened profile the ribs do not converge to a single point at the end of the mold but 'run out'.
When building this type of bowl, care must be taken to fit the ribs as closely as possible to the mold - without any forcing - to avoid accumulated small errors that can cause 'ballooning" of the bowl and consequent unintentional enlargement of the designed profile - particularly at the transition between the flattened area and the sides of the mold.

The rib profiles have now been marked in pencil on the mold surface ready for filing flat. To make both the filing operation and future rib fitting easier, the mold has been mounted on a simple wooden support that will be clamped to the bench and allow me to rotate and clamp the mold in any position around a horizontal axis as well as allowing rotation and clamping about a horizontal plane.
The neck block, at this stage, remains untrimmed (running to a point). After filing the rib locations flat it will be removed and the angled neck joint cut prior to the rib fitting operation.

I made a start on completing the mold this afternoon by filing the rib locations flat. I first used a sharp scriber to define the rib outlines so that they would not be removed during the filing process. I used a flat rasp of the type designed for use on soft metals such as lead or aluminium as this cuts cross grained wood quickly and smoothly to produce a flat surface. As the surface of the pine mold had oxidised over time to a darker colour it was quite easy to judge the extent of wood removal as filing progressed. Flatness of the filed surfaces was checked often as work proceeded.
Once the filing operation is completed and all surfaces (including the neck block) smoothed with fine sandpaper, the mold will be sealed with a couple of coats of shellack and wax polished to prevent the ribs being glued to the mold during assembly of the bowl.
The adjustable mold mounting proved to be essential for this part of the operation.

After filing all rib positions flat, the rib outlines were clearly redefined with a scriber and marked in pencil as a future guide when building the bowl. There is still a little material that needs to be removed at the rib locations for perfect flatness but it is now pretty close to where I want it.
The central rib outline has been traced in pencil on to a strip of tracing paper taped temporarily to the mold and the tracing then transferred to card where the rib half outline has been cut out. This profile, in turn, has been traced onto a strip of tinplate (more stable than card which can expand or contract with humidity changes at this time of year) which was cut slightly oversize to allow for final trimming of the rib when being fitted. The half rib profile will ensure symmetry of the central rib which, however, is likely to be the only rib that will be perfectly symmetrical as individual rib fitting progresses.

I have now narrowed the choice of material for the bowl to two options - the first being the highly figured birds eye maple (image previously posted) the second being alternating ribs of lightly figured maple and yew wood. I purchase the yew from Oxford University (Parks) during a visit to that city many years ago. While walking through the parks I noticed a couple of workmen felling an English yew tree of fair size (about 9 inches in diameter). So - always on the lookout for useful wood at that time- I enquired if I could purchase the timber. They referred me to the Parks Superintendent and a sale was made. The
Superintendent then mentioned that they had recently felled a very large 'American' yew tree that might be of interest. This was about 24 inches in diameter and had been slab cut into boards about 2 inches thick and 'stickered' to dry. I purchased the central board from the stack and this is the material that I now have on hand - having given away the English yew logs to a couple of deserving luthiers in the UK who I trusted would make good use of it.
The yew is unlikely to be 'American' as the only yews native to North America and Canada are small, shrub like trees (and now protected). I assume,therefore, that I have a European variety of fairly straight grained yew of the type used to make the English long bow in ancient times - the native English yew being of rather crooked grain and unsuitable for bow making. I understand that today the surviving stands of yew trees in Europe are now protected.
OK - I still have a decision to make but am in no particular rush to do so!

I had not tried heat bending yew before so did a test on a piece of scrap to find out. This was easily heat bent - dry - into a circle of about 110 mm (4 1/2 inches) in diameter so I do not expect any problems. Yew is a very fine textured wood and quite hard for a "softwood".
I have only used this wood before for the flat back of one of my vihuelas. The attached image shows the bent wood sample on the back of the vihuela - to give some idea of the attractive colour of yew wood when finished.
The strip of yew alongside is a piece of unfinished English yew showing the contrast in colour between heart and sapwood (and the crooked grain). Contrasting heartwood/sapwood ribs
(of straight grain) were often used in the construction of lute bowls in the late 16th C. - which gave the visual impression of double the actual number of ribs.

Just for information for those interested, here is an image of a lute with a bowl made with shaded yew ribs showing the contrasting sap wood/ heartwood of each rib. It is a treble lute by Wendelin Tieffenbrucker, Padua, Italy late 16th C.
The yew wood that I have in stock that I may use for the liuto attiorbato project is not shaded but is only heartwood.

interesting looking oud on ebay UK

Hi JDownning,

Just saw this vintage looking lute on ebay, it looks like a carved one piece of wood if I am not mistaken.

this just in a bid to revive this interesting post.

best regards

Awad

Each of the ribs of the bowl will be separated by a strip of contrasting material - in this case either black-white-black 'purfling' or just plain black 'lines' Purfling or lines may be purchased from luthier suppliers but I make my own purfling from maple and dyed black maple veneer. Three sheets of veneer (about three inches wide and of the required length) are glued together to make a veneer "sandwich" using plenty of glue (to avoid glue starved joints), clamped between two flat pieces of wood. Modern knife cut veneer is quite thin and the glue can seep through the veneer under clamping pressure so the clamping blocks need to be protected with plastic sheet to prevent the veneer composite sticking to the blocks. When the veneer composite is dry, one edge is planed straight using a shooting block and hand jointer plane. I then use a bandsaw with a fine tooth blade to cut a purfling strip of the required width from the veneer composite. I then plane the sawn edge of the composite straight again and then cut off the next strip - and so on until all purfling strips have been cut from the veneer composite.
Solid wood 'lines' are made in the same fashion with veneer cut on the bandsaw and planed to the required thickness and then cut into lines of the required width on the bandsaw. If I use lines for this lute they will be made from African ebony.
The image shows finished purfling on one of my 19th C guitar reproductions.
The purfling or lines have to be hot bent,like the ribs, and are quite fragile and brittle so I will run a few tests to ensure that what I have in stock can be used.

My work on this and other instrument building projects has been temporarily delayed over the past couple of months while recovering from some routine eye surgery.
Picking up the threads again today, some final adjustments were made to the mold to remove a few remaining high spots on the rib locations (which need to be flat). A low angle, adjustable throat, block plane was found to be best for cutting the cross grain of the mold. To minimise grain tear out and blade 'chatter', the plane was set with minimum throat gap and razor sharp blade - the plane being held at an angle to the grain of the wood using a paring action. The planing was done with the neck block in position so that it matched exactly the contours of the mold.
The mold will next be given a coat of wood filler, smoothed with fine sandpaper and sealed with a few coats of shellac.

In order to shellac the mold, the neck block must be removed (it is held in place with two screws under the base plate of the mold).

With the neck block removed, the neck joint was cut to the required angle (it is a lot easier to do this now than after the bowl is complete). Some judgement is necessary as allowance has to be made for the additional thickness of the ribs and neck veneers as well as the offset of the neck to the bass side.
After checking the settings several times, the cut was carefully made on a band saw with the table set at 40 degrees and the guide offset by a few degrees to give the required tapered cut for the neck offset. This should give a fingerboard width at the neck joint of 105 mm and a depth of 25 mm (excluding the thickness of the fingerboard). However absolute precision is not necessary as adjustments can be made later as work proceeds with fitting and shaping the neck.
A pilot hole was also drilled through the neck block for the screw that will retain the neck in place.

Now to prepare some fresh shellac to complete the mold.

The evidence from the few surviving instruments of the guitar family of the 16th C (two guitars and three vihuelas), as well as from engravings of the period, indicate that a 'renaissance' guitar might have had either a flat back or a curved or vaulted back. Surviving guitars from the 17th C. are also made this way. The flat back might be built from two plates like a modern guitar or from multiple strips.
The vaulted back was constructed like the bowl of a lute from bent staves. The attached images show a vaulted back 'renaissance' guitar that I built in 1974 for research purposes and the mold that I used. The mold was mounted on a flat construction board - blocked up to the required height - and the guitar built 'Spanish style' i.e. with the sides slotted into the neck (which had an integral neck block). The neck was clamped to the construction board - in effect forming part of the mold. The sides were then hot bent and shaped to the profile of the mold after which the ribs were hot bent and fitted in place.

With the neck block detached I decided, at this stage, to cut and fit the neck blank to the neck block and instal the mounting screw - as it is easier to do this now rather than later after the bowl is complete. The blank has been made well oversize to allow plenty of adjustment later when the neck is trimmed, shaped and finally fitted. This will be done after the bowl is complete and the sound board temporarily fitted in place prior to gluing.
The angle of the joint was cut with the bandsaw table left at the same setting used to cut the neck block - in order to ensure correct neck block to neck alignment
The wood for the neck is a piece of English sycamore (a kind of soft maple), cut 'on the quarter', i.e. with the end grain running vertically for stability. I have several planks of this material that were bought as clearance stock from a wood yard in Northern England over thirty years ago. The planks are so old that they were black on the surface when purchased but they clean up nicely with a couple of passes through a planer. I figure that the wood, by now, will be stable enough for a lute neck!

The material for the bowl is now being prepared. The maple staves are finished to the required thickness of 1.3 mm (0.055 inch) - these are sequence cut as a matched set.
The yew has been planed both sides but has yet to be finished to size as the best grain patterns have first to be selected and matched as closely as possible from a fairly random set of oversized staves.

Each rib will be separated by a black-white-black purfling strip. These were made some years ago by gluing together a "sandwich" of maple veneer then cut into strips about 0.80 inch wide on a bandsaw with a fine toothed blade.
The thickness is about 0.075 inch - a little too wide - so the width was reduced to about 0.065 inch using a simple, easily made tool. A series of slots about 1/4 inch wide of varying depth was cut, with a router, in a small block of maple about 2 inches in width. Selecting an appropriate depth of slot, the purfling is drawn through the slot while holding a small scraper blade, at the required cutting angle, on the block - repeating the operation until no more shavings are removed.

The yew selected for the bowl is cut into 13 strips measuring from 3 ft to over 4 ft in length and about 2 inches wide. Although the strips were cut in sequence from a block, the grain patterns vary a lot between each strip and are quite widely spaced (for yew). A profile of the centre rib - traced from the mold - was cut from a piece of card. By placing the card "window" over each strip it is possible to better visualise the grain pattern and identify the best part of each strip from which to cut a stave so that all the finished staves will be a reasonable match and free of flaws. A total of 8 staves are required so there is ample material for the job. The off cut material will be used to veneer the neck and pegbox to match the lute bowl.

The upper pegbox is a complicated affair - difficult to visualise in three dimensions. There will be quite a bit of carving required to shape the pegbox so it will be made in two pieces - joined along the centre-line. This will make the carving of inner part of the pegbox a bit easier but is also necessary because I do not have any wood of sufficient dimensions to carve the pegbox from a single piece. Anyway, a 'built up' construction is to be preferred to solid for better stability.
The plan for the pegbox extension is to make this from figured sycamore (a wood similar to maple) - rather than using engraved veneered panel overlays like many original 'liuto attiorbati' - relying mostly on the beautiful natural figuring of the wood grain for decorative effect

Not having wood of sufficient dimensions to cut 'book matched' blanks, it was necessary to carefully select blanks (cut from the same piece of wood) having reasonably matching grain pattern. To check and assess the grain match, the outline of the pegbox - in plan and elevation - has been sketched on the blanks.

The wood selected is of medium hardness and density for better wearing properties with boxwood pegs. The interlocking 'flame' figuring of the wood grain also helps to improve wear properties.

I do not like the amount of 'waste' wood involved in making this style of pegbox but I guess that "you have to break eggs to make omelettes". In any case, the off cuts will be saved and eventually used for other projects.

Continuing with wood selection for the pegbox extension, book matched blanks for the extension and the backing plate have been cut from figured sycamore. The thickness of the extension blank is 15 mm and the backing plate just over 5 mm so there is very little excess material to work with.

"Book matching" is done by sawing a billet of wood down the middle and opening the two pieces "like a book" to create a symmetrical matching grain pattern after they have been joined together. Usually some small lateral adjustment of the blanks is required to obtain an optimum match of the grain.

Cutting a billet open often reveals hidden flaws in the wood that must be worked around or incorporated as part of the grain pattern. Maple and sycamore are prone to staining during seasoning unless the drying process is carefully controlled. The staining - caused by bacterial action on the sugars in the unseasoned wood - does not affect the strength of the wood. It is not a decay in the wood.
The wood selected for the extension has some staining at the upper pegbox end so the extension pieces will be cut to avoid the stain. The backing plate has two symmetrically placed 'pin' knots at the lower pegbox end. There is sufficient material in the blanks to avoid the knots altogether or they can be left as part of the grain pattern. This choice can be made at a later date.

A change of plan.
The upper pegbox is now to be made in three parts - two side blocks and a central spine. This will facilitate carving and shaping of the pegbox interior surfaces.
After squaring and fitting the three components, six pilot holes for the pegs have been drilled as it is easier to do this operation now than it would be after shaping the curved exterior of the pegbox.

The internal components of the upper pegbox have been shaped prior to assembly. Rough carving was done with a sharp chisel, shaped with a small wood rasp and then cleaned up with a file. The downside of working with highly figured wood is avoiding grain tear out.

Making the upper pegbox is time consuming so will work on it bit by bit between other jobs.
The sides and top have been cut and finished close to the required profile leaving final finishing after gluing the sides and core which is the next step. Shaping of the curved front and underside of the pegbox and trimming to length will take place after assembly. The final geometry is easier to visualise at this stage.
Drill bits positioned in the pilot peg holes will ensure correct alignment of the sides while being clamped for gluing.

At noon today the relative humidity in the kitchen of our house hit a high of 80 per cent. This unseasonal - and possibly record breaking - weather situation has put a stop to fine woodworking and music making. Hopefully it will all pass over in the next few days and conditions return to 'normal'.
In the meantime the upper peg box has been glued and carving complete - awaiting final finishing after fitting to the neck extension. The front and top section of the pegbox will be veneered with African ebony to match the fingerboard and planned inlay of the neck extension. The back of the fingerboard and neck extension will be veneered to match the alternating figured sycamore/yew ribs of the bowl.

Interestingly, the notes that came with the museum drawing of the original instrument state that "The carving in the upper pegbox is fairly crude" Surprising, perhaps, in an instrument that, in its day, must have been a very fine, highly decorated, presentation instrument?

As the weather has continued to be very wet and humid, work has been restricted to cutting and shaping a few components. This morning humidity levels dropped to 65 percent so I decided to move ahead with the first stage of the lower pegbox/neck extension assembly.
This a change of plan in that the lower pegbox/neck extension is now to be of 'built up' construction - made up of two side pieces, a tapered fillet infill piece, a tapered end block and a back plate. The neck extension and lower pegbox on the original lute was carved from a solid piece with a separate back plate.
As the assembly is intended to be permanent, it was decided to use epoxy glue - convenient to use and strong but irreversible once cured. There are different formulations of epoxy glues commonly available from hardware stores so it is important to select a brand that will set hard and not remain semi-hard in the cured state. The so called 5 minute epoxy glue varieties should not be used as they are lower strength and provide less setup time needed for more complicated or critical assemblies. Always run a test on scrap material before using any glue to ensure it is satisfactory.

The first step was to prepare the two side pieces and the fillet infill piece. The fillet was cut and planed to the correct taper and two plates of African ebony - about 2mm thick - were glued to each side. The ebony is to provide a decorative contrast between the natural figured grain of the side pieces and the fillet.
The end of the fillet infill that will be exposed inside the lower pegbox was decoratively shaped as in the original lute. The shaping was done with a file which gives a smooth finish on the end grain surface. The best files for this operation are those sold for sharpening chain saws - they are round section parallel sided (not tapered), smooth cutting, good quality, readily available in various diameters and relatively inexpensive.
Prior to gluing, the components were 'dry' assembled to ensure correct fit and their relative positions marked in pencil on the side pieces as a reference during assembly.
Here the fillet and side pieces will be assembled first and then followed by shaping and fitting the end block as a second, separate stage. This will ensure exact relative positioning of the end block.

In order to glue a tapered assembly like this, a simple gluing jig is required to hold each component in its correct relative position until the glue has set. The jig was made from a flat piece of particle board as a base with blocks B nailed to the board to locate the side pieces and block A to hold the fillet piece in place. Without these blocks the fillet would simply slide in the direction of the arrow under a clamping force.
A piece of polythene sheet was placed on the jig surface to prevent the assembly being glued to the jig by any glue 'squeeze out'. Glue was applied to each side of the fillet piece which was then clamped in place with spring clamps. Two clamps were also used to ensure that the assembly was held flat on the base of the jig. After checking the assembly to ensure nothing had moved out of place, everything was left until the glue fully cures (in about 24 hours).

At last the humidity today dropped below 60 per cent. Still too high for more critical assembly work but there is other work that can proceed and I need to finish this lute this year.
The peg blanks were turned as a batch a few years ago. About a hundred of this particular style were made using a profile cutter on a metal turning lathe - so I have plenty in stock to choose from.
I need 19 pegs for this instrument - boring work - so will progress accordingly. This might take some time!
The pegs are made from a close grained hardwood - similar to boxwood - that I purchased many years ago, about 60 kilos, as small billets from a warehouse clearance in the North of England. Described by the seller as 'South American Boxwood' it is likely that it was originally stock intended for making spindles for the cloth weaving industry so may be a kind of dogwood or similar species. I have a couple of logs of genuine 'Persian Boxwood' purchased from a British importer in the 1970's so I know it does not match the quality of that. Anyway it turns well and makes good pegs.
The image is the batch of peg blanks in their unfinished 'rough' state straight off the lathe.

Completion of the pegs is done in two basic stages - shaping and finishing the peg head and cutting the tapered shank. Preliminary shaping of the peghead assists in the cutting of the tapered shank by providing a positive grip - so this operation is to be done first.
Early lute pegheads were generally made with flat faces - not with a curved profile like modern violin pegs. Preliminary shaping is done by slicing away all surplus material with a woodworking chisel followed by finishing with fine sandpaper. Each completed peg will later be 'fine finished' by burnishing and polishing. At this stage, the objective is to remove most of the machining marks left by the lathe cutter.
The original pegs were made from ivory and had an ebony finial or ball in the peghead. An ebony finial might be the way to go to finish these pegs as well.
More work to contemplate - but, hopefully, all worth it in the end! One peghead done only 18 more to go - yawn!

The end block of the lower pegbox/neck extension is now fitted and glued in place. This will be covered completely by the neck joint so I used a plain piece of sycamore for this part.
To level the assembly and to bring it close to finished thickness, I ran it through a thickness planer - many passes with very fine cuts. Not to be generally recommended but saved time. This is guaranteed to tear out the grain of heavily figured wood - so plenty of material has been left for final dimensioning and finishing. The finishing work is done using a scraper plane to eliminate grain tearout. This tool cuts a very fine shaving. The blade is sharpened like a regular plane blade - with a single bevel - and the edge is then burnished to a slight 'hooked' cutting edge with a burnishing steel. Here I am using a scraper plane with a fixed blade angle - so the blade has to be sharpened exactly at the correct corresponding angle. A better design is a plane with adjustable angle. This is one that I bought at a local flea market for $15. I found out why it was so cheap when I returned home - it has a crack in the side where a previous owner had unsuccessfully tried to remove the pivot spindle with 'brute force'. My eyesight is not as good as it once was - never mind. The crack has been stabilised with super glue and the plane will eventually be restored with additional reinforcing brass side plates and a rosewood sole. The handle is made of Brazilian rosewood and the fittings are brass - so still a bargain.
Highly figured wood may also be finished with sandpapers - if all else fails - but I prefer to use cutting rather than abrading tools wherever possible.
The lower pegbox inside dimension has been made about an inch (2.5 mm) longer than the original as I wanted to provide more room for a slightly wider peg spacing.

The next stage is to layout and drill the pilot peg holes in the lower pegbox. This is done prior to carving the peg box interior and finishing the neck extension to size.
A tracing in pencil is first made of the peg box interior onto a piece of card (by laying the pegbox over the card). The centre line is marked, the peg positions laid out with dividers and the peg centre lines marked at 90 degrees to the peg box centre line. Having counted the peg positions to ensure that there are, in this case, 13 pegs (it is easy to make a mistake and end up with one peg short - once bitten twice shy! Hence the layout on card. Card is cheap).
With the correct layout established, the peg centres are transferred in pencil onto both the outside face of the peg box and also the opposite inside face (by laying the pegbox over the card). The marking of the centres on the inside face is just to provide an additional check of the alignment during the drilling operation.
Now to make a simple jig to support the peg box at the correct angle for drilling.

A correction - for the record. Referring again to the museum drawing, I note that the original ivory peg heads were not flat sided but have a slight curve - but not as pronounced as on a violin peg. Nevertheless I shall make the peg heads flat sided - as most pegs of the time period were.

The drilling jig for the lower pegbox will be a simple, quickly made affair - used once and then sent to the scrap box for recycling. Looking through my box of useful bits, I found this jig that I could re-use and save some time. I have no idea what I used it for previously (always write on a jig to identify it , just in case it might be needed again in future).
A suitable piece of pine (from the scrap box) was screwed to the upright block (using recycled screws!) and checked for 'squareness' against the base.
The layout card of the lower pegbox was pinned to this 'back plate' and the marked centre line on the card made parallel with the base using dividers. The pegbox extension assembly was then correctly positioned over the card and a strip of pine - to act as a supporting surface - was positioned under the peg box and firmly screwed in place.
Now to drill the pilot holes for the pegs.

It should be noted that it is not absolutely essential to make a jig for drilling the pegholes - it can be done, freehand by eye - but this is best done with two persons, one observer to check alignment of the drill and one to drill.
If done single handed, a jig is the best way to go - especially if your eyesight is not what it used to be!

With the jig set up on a drill press and checked for squareness against the drill bit, a couple of holes were drilled to verify alignment. These were within the level of tolerance, so will go ahead and drill the rest.
I am using an undersized drill bit for the pilot holes so these will be increased in diameter later - this time freehand - using standard drill bits. A wood boring bit with centre point - 3/16th inch diameter - is being used to cut the pilot holes. This type of bit does not cut as cleanly as a similar bit with side 'spurs' to scribe the hole diameter but works well enough using a slow 'feed'. A small block of pine is used to temporarily support the inside face of the pegbox to minimise any damage as the drill 'breaks through'.

The lower pegbox/neck extension has been finished to the required depth of 16mm and the sides trimmed in the bandsaw to within about 1mm of the finished dimension. The depth of the peg box is quite shallow on the original instrument - presumably to keep the weight of this component to a minimum and so provide a better balance to the finished lute.
A cut out in the inside of the pegbox provides clearance for the treble strings. This cut out is easier to make before the back plate is fitted (an advantage of the built up construction rather than being carved from the solid as the original was). Before shaping the cut out, all of the peg holes have been reamed undersized. This preliminary operation has been carried out at this stage as maintaining the reamer alignment would be a bit more difficult if undertaken after the cut out had been completed. The peg hole reaming will be finished, and pegs fitted after the back plate has been glued in place. There is plenty of material yet to be removed with the peg reamer.
Before reaming the peg holes, the large diameter of the tapered hole is marked in pencil. This is to ensure that no mistakes are made. As a right handed player I would be less than pleased if I cut a left handed pegbox in error at this stage!
Having previously drilled the pilot holes, each hole has been increased in diameter using ordinary drill bits, like reamers, in a hand cranked drill. Slow but ensures that control is maintained. This operation reduces the amount of excess material that the reamer must remove and so speeds things up a bit. Be sure to double check that the correct diameter of drill is being used as work proceeds - oversized holes at either end are not an option!

With humidity levels back up to an uncomfortable 75%, work has been restricted to preliminary shaping operations.

The upper pegbox has been faced at the front with African ebony and brought close to the finished shape by planing and filing. Ebony is hard and brittle so care has to be taken to avoid chipping the edges.
The slot for the upper pegbox nut has been cut on a router table using a sharp, solid carbide cutter 4.5 mm in diameter and by making several shallow cuts of increasing depth. For this operation the pegbox was inverted on the router table and traversed by hand against a guide with end stop - a painstaking task as any slight mistake here could ruin the work. The slot should be a couple of millimeter deeper but it was decided to quit at this point - while the going was good.
The peg holes have been reamed roughly to size.

It has been decided to fit ebony banding along the edges of the lower pegbox/neck extension - to provide a contrast at the joint between the upper pegbox and neck extension, to provide a harder edge to the neck extension and also as a decorative feature to complement the inner fillet of the neck extension. The rebate for the banding was cut on the router table using a solid carbide bit - after running a couple of trials to ensure that the router would give a clean cut without chipping the highly figured sycamore of the neck extension.

It will be best to glue the banding in place before shaping the interior cutouts of the lower pegbox interior - to avoid the possibility of damage to the edges of the freshly cut rebates.

For those interested, here are some images of my router table - home built specifically for my general woodworking requirements. Not a thing of great beauty - but just as accurate (or more so in some cases) as any commercially available equivalent. It is basically in two parts - a simple wooden clamped holder for the router that can, in turn, be held in a bench vise, and the table itself made from a piece of 3/8 " (9mm) 'Plexiglass' or acrylic sheet. This readily available material is stable and perfectly flat - essential for accuracy. The router fits inside a plywood ring screwed to the table, fitted with wooden 'toggles' to hold it in place. To release the router - if required for freehand use - the toggle clamps are simply rotated to one side and the router removed.
The router table guide is infinitely adjustable! - just a piece of straight sided pine from my scrap box clamped to the table. The settings of the guide and router cutter depth are always checked for accuracy by first cutting a piece of scrap wood.
The rebate is cleaned up after routing, to remove any stray 'whiskers', using an "emery board" (here 'borrowed' from my wife's manicure set!) a low cost tool and perfect for the job.

The banding was cut to size - from an accurately planed strip of ebony - on a bandsaw. I did not take the trouble to change to a fine toothed blade - a general purpose 4 tpi blade was used - so the banding was cut well oversize to allow for the inevitable chipping of the wood during sawing (ebony is very hard and brittle). The rough saw cuts will later be removed with a plane.

The banding has been glued in place with a good quality PVA general purpose, woodworking glue (Lee Valley 2002 GF). The rebates were coated with plenty of glue so that "squeeze out' was visible along the glue lines. The banding was held in place with standard white paper masking tape - pulled tight by stretching it a little. Stronger stuff is, apparently, available but I have never used it. If a joint is well made and fitted - fairly low clamping forces are all that is necessary to hold things in place until the glue cures.
The banding has been brought to size by planing with a fine set plane.
The interior pegbox cut outs were made first by removing most of the waste with a sharp woodworking chisel - used with a slicing/paring action to avoid 'tear out' of the highly figured wood. A small wood rasp then quickly brought everything close to size followed by fine cut files to remove the coarse marks left by the rasp. Rasps and files (cutting tools) were used to avoid grain tear out.
Everything is now very close to finished size. Humidity levels are still too high so fitting of the back plate and upper pegbox will have to wait.

congratulations jdowning
big project

Hi Mr. jdowning,

I am sure tis will turn out to be a grand oud, judging by your meticulous work and atention to details. I'm sure everyone is following your progress and indeed it is an educational project to all of us.
I wish you all the best.

regards,

Awad

Still too hot and humid for instrument assembly or anything other than light duty physical work - so spent half an hour planing the strips of yew for the ribs to remove the rough saw marks - before settling down to recover from these exertions with a refreshing glass of beer! No rush - this instrument is for my own use.
Nevertheless - "All the time in the world - but not a moment to lose" is a saying always to keep in mind!

The strips were cut from boards years ago when I only had a cheap, inaccurate table saw to do the job - so lots of waste unfortunately - and quite a rough sawn finish as well.
To save time, each strip was run through my 12-1/2 inch thickness planer - a low cost, light duty machine that does a reasonably accurate job for my woodworking needs around the home. Planing sections less than 1/4 inch (6mm) in thickness is not recommended on this machine. To get around this limitation, a 1'' thick pine board - a bit longer than the yew strips - was prepared in the planer. Each strip was then planed by placing it on the board and running the board and strip together through the planer - taking very light cuts with each pass. This way it was possible to remove all of the saw marks, make the strips of uniform thickness and reduce the thickness to 2mm. This is the minimum thickness possible with this planer at which point there is gouging (sniping) over the last couple of inches of each end of the strip - accompanied by an alarming 'crackling' noise - the signal to quit planing!
With the strips 'rough planed', sections of each strip can now be selected for best match using a thin sheet metal form with the central rib shape cut out. A form made of card is not as good for this as it tends distort out of shape in use. Once selected, each over sized section will then be hand planed to the required thickness of about 1.5mm - ready for cutting to shape and forming into the ribs.

At the time that I purchased the yew, I was told that it was "American Yew" but had some doubts due to the width of the boards (about 2 feet wide). I have since learned that a species of yew native to the Pacific coast of Canada and America (Pacific Yew) typically reaches diameters of 2 ft or more - the largest tree of its kind growing to 60 ft tall and 6 ft in diameter. So this is likely the species of yew that I have - a specimen tree culled from the parks of Oxford University. Nice to know the history of the wood that one uses.

Continuing with work that can be done in humid conditions.
The mold for the bowl has been finished with sealing coats of shellac. The rib layout - as reported earlier in this thread - was marked out using a flexible metal straight-edge rotating about a fixed point at the neck-block end of the mold and the rib positions carved flat. The neck block - with the neck joint already cut at the required angle - has been screwed in place. The mold is mounted upon a 3/4 inch (19 mm) plywood base. I prefer this arrangement so that I can make the outer ribs a little oversize - the raised mold then making trimming of the edge of the bowl easier. More on that later.
I am not satisfied with the rib layout, as marked out on the mold, which does not seem quite right at the bottom end of the bowl. With this configuration the ribs on either side of the first, central rib will need to be excessively curved to meet at the end of the mold. The central rib must, therefore, be made a bit wider at the bottom end of the mold. The rib layout, from the widest point of the mold to the tail block, however, should be reasonably correct. The geometry of the central rib would seem to be critical but how to determine the optimum geometry? I guess that trial and error is the only way (for me).
A scrap rib blank of sycamore was reduced to size on a thickness planer and then brought down to 1.5 mm -by hand planing. I used the thickness planer method, previously mentioned, because the scrap stock was rough cut many years ago and I wanted to save time. However, although it has worked, this is not a method of thicknessing rib stock that I would recommend. Rib stock - accurately cut on a bandsaw - should be hand planed to finished thickness. Easier and far less wasteful.
The rib profile - traced from the mold layout - was cut out, well oversize (in the flat) - on a bandsaw. This rib blank was then hot bent to the required mold curvature until the fit was close without need to force the rib onto the mold surface. The test rib was held on the mold by pins at each end, pushed into the mold (by the time this project is finished the mold will have as many holes in it as a pin cushion!). In preparation for the next stage of the operation, pins were also positioned along the length of the rib - on one side - to locate the rib on the mold and prevent any radial movement.
It is interesting to note the concave curvature of the test rib at this point. This is natural 'fluting' often to be found on lute ribs - the wider the rib, the greater the concave curvature. Lute ribs are usually around 1mm to 1.5 mm thick, (about half the thickness of oud ribs) - and are left flat ,not shaped to a semicircular exterior profile like oud ribs.
The fluting of the rib is due to the natural balance of stresses in the rib after bending (known as anti-clastic bending, in engineering terminology) where the stresses across the rib must balance the stresses induced along the length of the rib. This results in the concave surface observed. The rib then only contacts the mold along a single line - unless the mold is carved to compensate - which I have not done. However, this will help to ensure that the ribs do not stick to the mold as work progresses.

Playing around further with the rib geometry - to find the best fit - the second rib profile was tested. This time a strip of card was pinned into position on the mold (card is cheap and convenient to use). The profile of the edge of the centre rib was then transferred onto the card using compasses drawn along the length of the rib joint. The card was then removed from the mold, cut along the pencil line and replaced on the mold with this edge matching the centre rib. The outer edge of the second rib was then determined by transferring the width of the centre rib to the card at various points. A smooth curve was drawn through these points to give the profile of the outer edge and the card was cut along this line to give the complete rib geometry. The card rib was then used to check the next rib position.
After a few trials of this nature I should be able to get a better understanding of how the ribs should lie on the mold for the best overall fit. It should not be necessary to modify the mold but the ribs when finally fitted will likely not conform exactly to the lines as previously marked on the mold at the tail end. The lines at the neck block end should be pretty close. I expect that only the centre rib will be symmetrical in profile.
As the ribs are quite narrow in width at the neck block, the plan to use b.w.b. purfling between the rib joints may not be practical due to the overall width of the purfling. It has been decided, therefore, to use thin lines of African ebony instead. These will be cut from ~1 mm thick planed stock.

Playing around with card ribs revealed a few small errors in the mold geometry - the centre line was found to be offset by about 4 mm requiring adjustment of several of the rib 'flats' on the mold. The mold has now been corrected and adjusted by filing and sanding. Absolute perfection is not the objective because - apart from the first central rib - all of the other ribs will be hand fitted and non symmetrical in profile, due to the extreme 'flattened' section of the lute. Anyway perfect symmetry is not a feature of surviving lutes made by the master luthiers of the 16th and 17th C. Well, that's my excuse!

As we now (thankfully - and lets hope that it continues!) have a break in this Summer's unusually wet and oppressive weather conditions and as today is clear and sunny with relative humidity in the low 60 percent - the corrected mold was taken outside to be given several coats of (fast drying) shellac to seal the wood grain. The mold will then be waxed and polished to make it as 'glue unfriendly' as possible.
I intend to try building the bowl using hide glue - following the centuries old technique still practiced by master oud luthiers like Dincer - i.e. glued paper strips scorched and 'ironed' in place with a hot iron to hold the rib joints together as work progresses.
So now I am a bit concerned about the possibility of the ribs being glued to the mold (less likely if a toast-rack or bulkhead mold had been used) but reckon that the thin lute ribs - that should have a natural concave section after bending - will, therefore, not be in contact with the mold surface along the length of the rib joints (except at the tail end of the mold which will be protected with a paper overlay). We will see! Nevertheless, I plan - as an additional precaution - to 'break' any superfluous glue leakage on to the mold with a flexible spatula as each joint is completed.
Allons-y!

Moving on to prepare the African ebony rib fillets.
I am working in unfamiliar territory here as I have always made lute bowls previously without "between-rib" fillets or purfling.
The original lute has parallel sided fillets - rather than tapered fillets - so that is how I plan to proceed.
A strip of ebony veneer was cut on a bandsaw to about 2mm thickness and the saw marks on both sides removed with a plane and finished with a cabinet scraper. The target thickness of the finished fillets is to be about 0.8 mm. The depth of each fillet will be about 2 mm - to be cut on a band saw.
African ebony is brittle so - in order to minimise chipping of the wood during cutting - the band saw blade was changed to a fine tooth (14 t.p.i.), metal cutting blade and the standard fairly wide gap between the blade and the table was minimised by running a strip of scrap pine into the blade to provide, temporary, zero clearance, support during the sawing process. This worked quite well although some slight chipping is inevitable.
Before each depth cut in the bandsaw, the sawn edge of the ebony veneer was smoothed with a plane by "shooting " the edge - that is by running the plane on its side along the edge to remove all saw marks.

Each fillet strip was then reduced to the required thickness using a simple thicknessing jig - subject of an earlier post in this thread. Each strip was placed in a slot in the jig, of appropriate depth, a plane was placed over the strip with the blade held at an angle of about 45 degrees, and the strip repeatedly pulled under the sole of the plane to shave the strip to the required thickness. Note the spiral shavings produced.
The final finish was produced by drawing the strip under a scraper blade held at the correct angle.

A minor set back. Yesterday, when taking images for posting on the forum, I managed to drop the plane being used to thickness the fillets on to the concrete floor of my workshop. The result is a broken cast iron, lever style blade tensioning device. Fortunately, that appears to be the only damage in evidence.
The plane is an old style Stanley 601/2 low angle block plane with an adjustable throat that I purchased new in the 1960's. This plane is not particularly well made or machined but was made to work with a bit of adjustment and I have since used it quite a lot over the years. It is the only adjustable throat plane that I have so repairs were undertaken this morning.

As repair of the broken component is not feasible, the plane has been modified by providing a brass screw in place of the tensioning lever. The knurled screw is from Lee Valley cat# 00M91.01 with standard 1/4-20 thread that I happened to have in stock.
The old pivot pin on the plane was removed, the casting drilled out to 1/4" diameter and a steel nut was cemented in place under the casting with epoxy glue. As the nut will always be subject to a compressive load - the epoxy is only required to permanently hold the nut in position so should be fine for that duty.
With the plane dismantled, the opportunity was taken to lap the sole of the plane flat. This should be done with the blade locked in place (to take care of any distortion of the casting) so will check this again once the epoxy has set for 24 hours and the plane is ready for use again.