The familiar mid 15th C diagram of a lute by Henri Arnault de Zwolle is a composite giving a geometry and barring layout for a lute as well as
information on how to construct the body of the lute. According to a translation of the latin text by Ian Harwood, Lute Society Journal, 1960, the
parallel lines on the diagram depict the placement of 5 wooden 'bulkheads' of a lute mold. As the lute has only 9 ribs or staves, the semicircular
sections are flattened at the rib positions. To construct the body, thin strips of wood are shaped to conform to the 9 divisions on the bulkheads. The
first, central rib is first glued to the neck block and bottom block. The next rib is glued to the blocks and then attached to the neighbouring rib
with a long piece of paper coated with glue - and with a rather warm iron. When all of the ribs have been joined together, the mold underneath is
taken away leaving the ribs joined to the blocks. Next the joints inside the body are glued with parchment (reinforcement strips) and the paper
outside the body is then scraped off. Then the bars and belly are put on "as you know"
Arnault de Zwolle is describing the familiar "toast rack" mold used today to make ouds and lutes.
He also is describing a method of joining the ribs together (in days before the convenience of modern self adhesive tapes) using glued paper strips
'scorched' into place with a hot iron.
Thomas Mace, writing in 1676 (Musick's Monument) describes this technique for 'do it yourself' lute repairs in his chapter 4 "Concerning the
Mechanical Order of the Lute" in which the belly is held in place during re gluing with penny sized pieces of glued paper. An iron hot enough to
scorch the paper (but not the wood of the instrument) causes the glue - apparently - to set rapidly and 'crust' (and shrink the paper?) pulling the
joint tight. The paper can then easily be removed later by moistening with a damp cloth.
I have never tried this method but imagine that it may have some advantage over more modern methods using tape - possibly one being hot glue drawn
into the joints by capilliary action without the need to apply glue first (very quickly!) along the joint edges?
On website http://www.oudmaker.com there is a slide sequence of luthier Dincer Dalkilic making an oud. Unless I am mistaken, he appears to be making the body
of an oud using this technique (except that he is using an electrically heated iron). Is this technique commonly used these days among oud makers? If
so what are the advantages?
Johnjdowning - 12-21-2007 at 10:29 AM
I finally got around to contacting Dincer Dalkilic yesterday and he confirmed that he does indeed use this early method of construction for making oud
bowls. On his website Dincer has posted an interesting series of images in "Oud Making Photo Show" depicting the making of an oud including images of
the bowl assembly process.
It can be seen that he glues the ribs (and 'lines' between the rib joints - after hot bending to shape - by first applying hide glue along the edges
of the ribs (and lines). Each rib is then placed upon the mold and glued to the neighbouring rib in stages using glued strips of paper first laid over
the joint followed by a hot iron which re-melts the glue in the joint and then rapidly cures the glue on the paper which also shrinks a little so that
it holds the rib joint tightly together. To prevent the ribs sticking to the mold, strips of paper are also placed under the ribs in places where the
ribs make contact with the mold surface.
When complete, the entire outside surface of the bowl is covered in paper strips which helps to temporarily reinforce the finished bowl as it is being
removed from the mold. The paper strips are then removed by scraping, after first softening the glue with a damp cloth.
I am going to try this method on my next instrument project as it seems to have a lot of advantages.Oud Freak - 12-21-2007 at 01:08 PM
I think more care could have been brought to the bracing. Although a bracing is "invisible" it should be nice to look at, it should be perfect, with
perfectly shaped pieces of wood, and the parallelisms have to be respected. The bracing is the brain of the oud. Supposing this doesnt affect the
sound at all, the sound is not everything.
for example in this thread is shown a clean bracing. Good work Raby
Even though I know none of Raby and J. Khalaf, I appreciate their care for their precise and clean work. Congratulations for you folks. Keep it up!jdowning - 12-21-2007 at 02:30 PM
Perfection in anything man made is an illusion. No harm in dreaming though.Oud Freak - 12-21-2007 at 02:48 PM
Quote:
Originally posted by jdowning
Perfection in anything man made is an illusion. No harm in dreaming though.
Downing, look at the bracings of Khalaf and Raby, they are very real and excellent, consequently they are not a dream, and u'll understand No one is perfect, but there is a minimum. A musical instrument is not a scratch
paper.jdowning - 12-21-2007 at 05:36 PM
Oud Freak - What have your opinions expressed in this thread got anything whatsoever to do with early oud construction methods?Oud Freak - 12-22-2007 at 03:59 AM
Jdowning: I think it is worth to be tried. But what is particular about ouds, is somehow a kind of freedom in the design and in the constructing
methods, as well as many potentials for developing the instruments, without deying the old methods. Violins, flutes and pianos for instance cannot
tolerate such changes and are more rigid.
Despite the possible evolution of the oud, and regardless of the building methods, the inside and outside finishing must be the closest possible to
the excellent i.e Jameel Khalaf. Really beautiful work
Ibrahim Sukkar's finishing is also not bad, despite the affordable prices of his instruments.
If there is some oud which is built with the early methods, I'd be most delighted to try it!
I had a phone conversation with Nazih Ghadban this morning and he told me he was making a oud in the traditional way, but I ignore if it is the early
method or not.
Best Regards and Merry Xmas wishes
OFjdowning - 12-27-2007 at 03:21 PM
I am not aware of any step by step, 'hands on', detailed description - ancient or modern - of how an Arabic luthier might go about constructing an oud
bowl using this ancient method. Therefore, until such time as such a description is posted lets look at what 17th C Thomas Mace has to say in an
attempt to uncover a few more clues about the old ways. Mace gives a detailed account of how to remove, repair and replace a lute belly which
includes use of the glued paper/hot iron technique for making joints - a technique just as applicable for gluing the rib joints of a bowl.
In Chapter 4 of the second part of "Musick's Monument entitled
"Concerning the Mechanical Order of the Lute", Mace first gives a list of the tools needed to carry out the repairs and maintenance of a lute. The
first item is a glue pot of about a pint capacity (0.57 litre) and ".... some of the clearest and best made glue, together with isinglass (both which
mixed together make the best glue)" .
Isinglass is a type of fish glue, the best of which is made from the swim bladders of the sturgeon fish - now generally unavailable due to
overfishing and destruction of the environment of that species - although isinglass is also made from other unprotected species of fish. It is less
viscous and stronger than hide glue but also brittle in nature. Both hide and fish glues were being used in Egypt at least 3500 years ago.
Cennino d'Andrea Cennini in his "Il Libro dell' Arte" published in 14th C Florence, describes both glues - fish glue being good for "mending lutes"
and hide glue, made from goat skin, being good for "making lutes" - both available then in leaf or sheet form.
Mace does not give the proportion of isinglass to hide glue needed to make 'the best glue' so it might be assumed to be a 50/50 mix.
A source of isinglass in powder form may be found at home brew suppliers. It is used for clarifying wine and beer by the brewing industry.
Both isinglass and hide glues are prepared in the same way - by first soaking in water and then heating the glue in a water jacketed glue pot until
liquid. Both glues are used hot. I have never used isinglass glue so cannot verify its properties through experience but it is reported to be good for
gluing porous materials (such as parchment) and presumably its low viscosity facilitates its penetration into a glue joint by capilliary action. Also
mixing isinglass with hide glue may help reduce the brittleness of the isinglass if used alone.
If I can obtain powdered isinglass from a local brewers supply store I shall conduct some experimental trials to see how it performs when used as a
glue.jdowning - 12-28-2007 at 01:51 PM
The sturgeon is now globally threatened with extinction and is listed by CITES (Convention on International Trade in Endangered Species of Wild Fauna
and Flora) as endangered due to destruction of spawning grounds (building of dams, reservoirs etc.) and overfishing - both legal and illegal for the
eggs of the fish (caviar). Sturgeon isinglass is a byproduct of this international trade so - as I would never use elephant ivory in the construction
of a musical instrument - regardless of whether or not it had been legally acquired - the same must apply to sturgeon isinglass. Therefore, I will not
be attempting to carry out any trials or experiments to test a combination glue containing locally purchased brewers isinglass finings as previously
proposed.jdowning - 12-28-2007 at 04:01 PM
Apart from a glue pot and a quantity of 'best' glue, Mace's tool kit for repair and maintenance of a lute includes two heating irons - a large one,
in section measuring about 3/4 inch (19 mm) square and about 3 inches (76 mm) long or "the length of your middle finger" made of iron filed flat and
smooth at one end and at the other end fitted with a long shank with a sharp end run into a handle of wood to hold it by. The other iron, although not
essential, should be a quarter of the thickness of the large iron - useful for finer work.
Mace goes into detail about how to remove and repair a lute soundboard and then describes how to re-glue the soundboard to the bowl as follows:
"Now to your glue pot with back and belly, and begin with which(ever) you please, and anoint all the edges carefully around (with glue) where you know
they must join........ then, having a good fire ready, bring both to the fire and warm them a little and (then) clap them together quickly.......
Now having in readiness your great (large) iron red hot, heat the edges (of the soundboard) thoroughly all over ..... until you perceive the glue (in
the joint) has become warm and thin. "
Mace then describes how small pieces of paper "wet with glue" are place over the soundboard joint spaced so that there is a small gap between each
piece of paper ("a straw breadth or two .... so that you may see how the joint joins"). Using the end of the hot iron and "leaning pretty hard", each
piece of paper is then scorched in place - one by one - working gradually around the soundboard joint equally on both sides as work progresses, inch
by inch ("to avoid any unevenness"). The iron "must not be too hot for fear of burning the belly, yet hot enough to scorch the papers and (turn) the
superfluous glue into a crustiness" Having completed the joining of the soundboard in this manner the instrument is set aside for the glue to dry
until the following day.
To clean off the papers, a rag dipped in water is used to moisten all of the scorched papers and crusty glue "often renewing the moisture" a little at
a time so that within the space of about half an hour the glue will be so soft that the papers can be gently scraped off using a fingernail. Mace
cautions against using any tool other than a fingernail to avoid damaging the soundboard. The glued joint is then allowed about two days to cure.
This procedure is the same as would have been used for gluing the ribs of the bowl during construction the only difference being that the paper strips
remain on the bowl exterior (for temporary support) until the bowl has been removed from the mold.jdowning - 12-30-2007 at 01:20 PM
Now for some trials to test the method of using the paper and glue - a la Mace and Arnault de Zwolle. I am not sure how well it will work so a bit of
'hands on' experimentation is in order.
I figure that the glued paper strips will shrink when scorched with a hot iron - drawing the joint tightly together and possibly allowing sufficient
glue to seep into the joint as well.
A bit of cross technology research to find out about paper shrinkage revealed that bookbinders know all about it. Paper has a grain direction - like
wood - as well as built in directional stresses, due to the manufacturing process. When soaked with water on one side, paper will expand across the
direction of the grain (but very little along the grain) causing the paper to curl. It will shrink again when it dries. Running a few tests it was
found that paper samples shrank between 3% and 5% across the grain direction from their fully expanded state when saturated with water to their fully
over-dried state when rapidly heated (on a woodstove surface). Shrinkage along the grain was found to be almost negligible.
The best paper from these limited tests was found to be that from "junk mail" (you have just won a million dollars or other 'special deals' etc!).
This kind of paper is of a fairly heavy grade (to feel impressive no doubt) but contains no fillers or sizing so the paper absorbs moisture quickly.
Whether or not it will be strong enough to shrink without splitting under test remains to be seen. Other paper may also be suitable on testing.
Two test are planned using jointed walnut plates, with a piece of purfling between, mounted on a wooden block held in place with pins. Paper covering
the block will prevent the test pieces sticking to the block itself.
The first test will be with a dry joint, the glued paper strips being scorched over the joint - to see if glue will seep into the joint - and the
second test will be first to apply glue to the joint surfaces and purfling followed by scorching the glued strips in place.
I shall be using an old household electric iron set at maximum temperature to do the paper scorching operation. This setting is just enough to scorch
the paper a light brown colour.
The paper strips - about 1" x 3/4" (25 mm X 19 mm) - will be applied with their grain in the same direction as the joint (or the cross grain direction
across the joint) so that when the paper strips shrink, the joint surfaces will be pulled together.
So, with some good quality pearl (hide) glue soaking overnight (no isinglass glue added!) the trials will be carried out tomorrow. Lets hope it all
works!
Jesse Frank - 12-31-2007 at 05:26 AM
looks lke a neat technique. I'm really interested to see how it works.
We have sturgeon down here that you have to watch out for..... they don't bite that I'm aware of, but they do jump, and they are really big fish!
boaters have been seriously injured by them on the Suwanee! It's mostly the jet skis that have to worry about it, though....
Anyway, thanks for letting us all benefit from your research.Jameel - 12-31-2007 at 05:45 AM
Fantastic post John! I can't wait to see the results of your test.
By the way, I'd like to hear your thoughts on adapting the butt-joint and screw/nail joint for attaching necks.jdowning - 12-31-2007 at 01:41 PM
I carried out the gluing trials this afternoon. The results were not quite what I had anticipated but I have something to report in my next
postings.
Jesse - Sturgeon are a primitive looking fish and can grow very large (if allowed to live!) but beautiful to my eye. There are sturgeon on the West
coast of Canada but these are now on the point of extinction with only about 500 individuals currently surviving as I understand it. It is now illegal
even to attempt to fish for them. I will have a look into the Australian species as they seem - from what you say - to be plentiful but I do not know
of Australian sturgeon isinglass (or caviar). Perhaps it is a different species with the same name?
Jameel - The nailed butt joint is the way it was done on early lutes where "speed was of the essence" in fixing the neck to neck block when hide glue
was being used. The nails used had large heads to apply maximum pressure to the glue joint and the nail also prevented the neck sliding out of
alignment until the glue set. This was particularly important on lutes of the late 16th C and 17th/18th C which had significantly sloping neck joints
- unlike earlier lutes and ouds where the joint is at a right angle to the fingerboard. Of course, the nail also added extra structural strength and
security to the neck joint assembly. Screws can be used if slower setting modern adhesives are employed for the neck joint. Screws would not have been
used for this purpose in earlier times as they were then individually hand made and would have been costly - as well as being too slow for fast
setting hide glue.
I very much like the traditional longitudinal dovetail joint (or the modern spline adaptation) used by oud makers today as this allows the set of the
neck to be accurately established in a 'dry run' before final gluing and would also allow use of either hide glue or modern adhesives. I am not aware
of this type of joint ever being used on early lutes. However, it would likely be a difficult joint to make accurately on a sloping, lute neck joint.jdowning - 12-31-2007 at 03:30 PM
Gluing Trials - Test 1 to determine if scorched paper shrinks to pull a glue joint together.
The glue used was pearl hide glue prepared in the usual way. A sufficient quantity of glue was placed in a small, clean glass jar, covered with water
and allowed to soak overnight to absorb all of the water. The jar was then placed in a saucepan of boiling water to melt the glue for use at the
proper temperature of about 140 F (60 C) - above this temperature the glue loses strength. The glue is used hot and is of the correct consistency when
it runs off the glue brush in a steady stream making a slight 'spattering' noise as it falls back into the glue pot. If it is too thick, a little
water is added to bring it to the correct consistency.
A simple test rig was made up of two strips of wood (Hickory in this case - no particular reason, they were just handy). The strips were hinged at one
end with masking tape so that they were separated with a gap of about 0.025 inch (measured with feeler gages) at the other end and the gap could be
closed with light finger pressure. A strip of paper was first saturated with water, coated with glue and then laid across the gap - the paper
orientated with its cross grain across the gap for maximum shrinkage. The paper was then scorched in place with a hot iron causing the water in the
glue to instantly boil off and the paper to quickly set in place within a second or two. The glue on the paper turns to a "crustiness" as noted by
Mace.
The gap was measured after this operation but no significant closure was detected indicating that paper shrinkage (if any) is of no consequence in
closing the joint surfaces. The scorched paper strips simply quickly lock the joint firmly in place and do not draw the joint together. The grain
direction of the paper is. therefore, not important if this method of gluing is to be used.
jdowning - 12-31-2007 at 03:55 PM
Test 2 - does the glue from the scorched paper seep into a dry joint?
Arnault de Zwolle's treatise of the 15 th C does not mention glue being applied to the rib joints - only glued paper being applied over the joint (as
a temporary measure) and the ends of the ribs glued to the neck and end blocks while the bowl is being constructed. Does this mean that some of the
glue on the paper strip will seep into the rib joints when the paper is being scorched in place and that it is not then necessary to apply glue to the
rib joints prior to applying the paper strips?.
This was tested by joining two pieces of walnut with a strip of purfling between. The joint was left 'dry' and glued paper strips scorched over the
joint. Paper was placed under the joint so that any glue seeping through the joint would not cause the test sample to stick to the mounting block
underneath.
A small amount of glue did find its way into the joint but not sufficient to form a consistent strong joint so that the joint started to separate in
places as the sample was flexed.
Part of the problem also was that the purfling was made many years ago using modern PVA woodworkers glue. The glue started to degrade when the paper
strips were being scorched causing the purfling to delaminate. Had hide glue been used to make the purfling in the first place this would not have
been a problem as hide glue would have simply melted under the heat and reset again when the heat was removed.
Final test comes next.
jdowning - 1-1-2008 at 06:12 AM
Oops! Sorry Jesse, wrong country for sturgeon - Florida not Australia!
Nevertheless the Gulf sturgeon of Florida are also threatened with extinction and are listed under the Endangered Species Act 1991. Sounds like there
are too many boaters in that part of the world.
For the historical record, a very strong combination glue of hide and isinglass glues can be made using spirits of wine (i.e. ethyl alcohol) instead
of water when preparing the glue. A formula for this is given in "Cyclopedia of Useful Arts" London & New York, 1854.
" made by infusing common glue in small pieces with isinglass in spirits of wine, just sufficient to cover the mixture. Heat is then cautiously
applied and when melted powdered chalk is added making the whole of an opaque white"
Presumably this glue could be made without the chalk 'filler'.
The same book gives the following preparation of so called 'Diamond Cement' or white fish glue. "Isinglass is dissolved in dilute spirits of wine or
common gin. The two are mixed in a bottle loosely corked, and gently simmered in a vessel containing boiling water (i.e. a water bath); in about an
hour the isinglass will be dissolved and ready for use. When cold it should be an opaque, milk-white, hard jelly: it is remelted by immersion in warm
water (i.e. in a water bath while in the bottle) but the cork should be loosened. After a time a little spirit should be added to replace that lost by
evaporation"
Other sources recommend use of cheap brandy or common whisky in place of gin. No doubt vodka would also serve for making glue from Russian isinglass!
The alcohol content of these glues would, no doubt, help to preserve the glue from rotting and decomposition.
These strong glues would likely be useful for high strength joints on instruments - gluing bridges, neck joints etc. - joints that could presumably be
taken apart if required by application of moisture and heat.jdowning - 1-1-2008 at 08:41 AM
Test 3 - apply glue to the joint surfaces before scorching paper.
This test follows Mace's instructions to first "anoint" the joint surfaces all around with glue - otherwise it is a repeat of test 2.
This time - as expected - surplus glue was found to have soaked into the protective paper under the test piece indicating that the joint was properly
saturated with glue throughout (see image 2)
The paper strips were removed according to Mace's instruction, by dampening them with a moistened cloth to re-soften the encrusted glue and then
scraped off using a fingernail. The light glue residues remaining on the surface were then removed with a cabinet scraper.
The finished joint was destructively tested this morning, after the glue had dried overnight, by flexing the test piece. The joint separated (snapped)
cleanly along one side of the purfling strip but this time the purfling, unlike in test 2, remained intact. It should be noted that the thickness of
the walnut test pieces was 0.055 inch or 1.4 mm - a typical rib thickness for a lute. Also walnut is an opened grained wood so that the strength of
the joint may have been affected by the open pore structure of the joint surface further reducing the effective area of the joint.
Another test will be run using a close pored sample of hard maple 0.095 inch thick or 2.4 mm - with glue applied to the joint surfaces but without the
purfling strip to see if that makes any difference to joint strength.
It should be remembered that the rib joints in an oud or lute bowl are always reinforced from the inside of the bowl with paper strips permanently
glued in place.
Conclusions to follow after the final test.
jdowning - 1-4-2008 at 07:10 AM
Further comment about isinglass glue made using alcohol can be found in Volume 1 "Turning and Mechanical Manipulation" by Charles Holzapffel 1843.
deals in great detail with 'materials, their choice, preparation and various modes of working them'. It is a very useful reference on how to work
materials such as bone, ivory, tortoise shell, sea shell, horn etc and, as such, is of interest to the luthier as well as a turner. A modern facsimile
reprint of the book is available from The Astragal Press, Mendham, New Jersey 07945-0239.
Holtzapffel, in chapter VIII dealing with ivory, gives the formula for white fish glue or 'Diamond Glue' previously posted in this thread saying that
it is very often used for ivory work - for attaching ivory to ivory or ivory to wood. So this may well have been the glue used by early luthiers for
building instrument bowls with ivory ribs. He seems to confirm that the alcohol (in this case common gin) helped to preserve the glue as he states
that "When the isinglass is dissolved in water alone, it soon decomposes" Again this is mentioned for historical interest only as, hopefully, few
luthiers today would even contemplate building instruments from materials derived from endangered animal species.
It should also be mentioned that true isinglass glue should not be confused with a material called water glass which is also sometimes called
isinglass. Water glass (sodium or potassium silicate) was once used for preserving eggs that were immersed in a solution of the mineral. Water glass
can be made into a cement for repairing glass but is not recommended for use on wood.
As part of these on going gluing trials, I shall make up a glue from pearl glue dissolved in alcohol instead of water to investigate how that performs
(strength, shelf life etc.) compared with glue dissolved in water in the usual way. I shall use cheap brandy (40% alcohol by volume) for this purpose.jdowning - 1-9-2008 at 03:33 PM
Due to a major power outage caused by adverse weather conditions the gluing trials could not be completed so instead the relative strength of pearl
glue prepared with water and with alcohol was investigated.
Test samples of pearl hide glue were made by covering the glue in water, in the usual fashion, and, for sample #2, by covering the glue in commercial
brandy of 40% alcohol by volume in strength. Both samples were allowed to soak for 24 hours to fully absorb the liquids. The glue samples were heated
in a water bath until melted and applied hot.
A simple test rig was made - comprising two rods of pine measuring 1/2 inch X 3/4 inch X 22 inches long, glued to a mounting block of pine clamped to
a bench. The glued areas measured 1/2 inch by 2 inches or 1 square inch. An increasing torsional loading was applied - by moving the load inch by inch
along each rod - until each joint failed. The load was simply a couple of 2 litre water bottles filled with water weighing in total 8.9 lbs.
Loading the glued joints in tension - perpendicular to the wood grain - was chosen for simplicity as wood is weakest when loaded in this manner.
Each glue joint failed when the load was positioned at
20 inches from the pivot point (at the edge of the mounting block). Both of the test joints failed due to structural failure of the wood and not the
glue joint although the alcohol based sample was marginal - probably because the glue was a too viscous when applied and should have been diluted a
little more. Nevertheless, the test indicates that both joints were of equal strength in tension perpendicular to the wood grain. Therefore, the use
of alcohol rather than water has no significant effect on joint strength although it may help preserve the glue from decomposing in storage.
Out of interest, a rough calculation of the stress at failure of the joints gave a figure of 267 p.s.i. which compares with the standard equivalent
stress limit of dry white pine of 280 p.s.i.
jdowning - 1-12-2008 at 06:38 AM
As the first test gave marginal results, the alcohol based glue trial was repeated using the same test rod inverted to provide a new test surface.
This time the glue was diluted further with a little brandy to make it of more fluid consistency when hot. The joint was then clamped and left to cure
for 22 hours before being tested to destruction.
With the load at maximum distance along the rod (22 inches), the glued joint remained firm. Failure of the rod itself was then achieved, at an
undetermined load in excess of this, by pushing down the end of the rod.
This limited test appears to indicate that the relative strength of the alcohol based glue is maintained (or possibly increased?) after the glue had
been reheated (usually hide glue weakens with each reheat so should be used fresh for maximum strength). However, a series of controlled tests would
be needed to confirm glue strength after repeated reheat cycles.
jdowning - 1-12-2008 at 01:12 PM
Test 4.
The final part of the gluing trials was to repeat test 2 and 3 using 2.4mm thick maple test pieces. The paper strips used for the previous tests were
found to be a bit fragile after soaking - tending to stick to the iron and tear apart when being scorched in place - so a slightly heavier grade of
paper of paper (about 40lb) was used and the glue applied without soaking the paper strips. This was more satisfactory. The strips were applied with
the grain of the paper running across the joint to take full advantage of any slight shrinkage that would pull the joint together before the glue
set.
The attached image shows the test piece with a dry joint - no glue being applied to the edges. Although the paper strips held the joint edges firmly
together, the joint remained dry with no glue seepage from the paper strips into the joint after scorching. This is not surprising as scorching the
paper causes the glue to set hard in a crust within a second or so.
A second test was made with glue applied to the edges of the joint. Scorching the paper strips in place caused the glue in the joint to re-melt and a
firm close fitting joint resulted. After the glue in the joint had cured overnight, the paper strips were easily removed by wetting and scraping and
the joint was then destructively tested by bending the test piece. The resulting joint was strong enough but failed along the glue line with a clean
fracture. It is possible that the glue may have been weakened due to overheating during the paper scorching process so prolonged scorching should be
avoided - only sufficient heat to set the paper in place should be used.
This method is for making a bowl using hide glue. Hide glue gels quickly and, therefore, must be remelted bit by bit as the ribs are joined. The use
of glued paper strips to temporarily hold the individual rib joints together, as well as the complete bowl assembly, is both viable and practical.
I shall use this method for my next project.
jdowning - 1-18-2008 at 07:36 AM
For completeness, another test was carried out to compare the relative strengths of hide glue prepared with water and with alcohol. This time both
glues had been reheated - the water based being subjected to two reheat cycles (with a little water added to achieve correct consistency) and the
alcohol based glue to three reheat cycles (without need to add more brandy).
This time both joints failed at a calculated stress of
240 lbs/square inch - that is 10% less than on the previous combined test. Although the tests are not statistically significant, the result is as
expected - indicating that hide glue becomes weaker on each reheat cycle and should, therefore, be used freshly made for maximum strength. It can be
seen that both joints failed marginally at the glue line and that the water based glue appeared to be slightly weaker and less tenacious than the glue
prepared with alcohol.
jdowning - 1-20-2008 at 03:32 PM
Mace gives the sizes of the pieces of paper for scorching over the glue joints "... then with with little pieces of paper (so big as pence or
two-pences, wet with glew)...."
Initially I was not able to establish dimensions of the coinage referred to by Mace - the English silver pence and two-pence coin of his time - but,
after a little research, have now been able to find the data. A complication is that the coinage of the period may have been made either by stamping
(hammered) or machine made (milled). The hammered coinage was not only uneven in diameter to start with but was also subject to fraudulent "clipping"
which made the diameter of a coin less than when originally minted. To cut a long story short, the diameter of a silver penny or two-penny coin of the
period was between 12 mm and 15 mm - a bit smaller than I first thought - so about 1/2 inch square was the size of the glued paper strips used by Mace
for his repair work.
As an aside, Mace also refers to the dimensions of a coin to describe the height at the nut of a lute - i.e. the distance between fingerboard and the
underside of the strings. In this case he says that it was the thickness of a half-crown. From the dimensions and weight of a half-crown coin of 1642,
I was able to calculate a thickness of about 1.5 to 1.8 mm which is about right for a (fretted) lute but might be a bit lower for an oud.
Although Thomas Mace in his "Musick's Monument" of 1676 was writing about the English and French lutes of his time - as well as other instruments like
the Theorbo, viol etc. - some of what he wrote is also relevant to the oud - because of the structural similarities between lute and oud. I have
quoted some of the relevant passages from Mace's work in this and other threads on the Forum.
For those interested in reading, first hand, about what Mace had to say in greater detail, there is an excellent, cloth bound, facsimile of his book
published by C.N.R.S. (Centre National de la Recherche Scientifique, Paris), now out of print but available through on line book sellers like Addall
books. The complete facsimile is Volume 1 of two volumes. Volume 2 is a commentary (in French) about the life and work of Mace and includes
transcriptions of the music.
An alternative source of the work is through 35mm microfilm or on line digital copies available at "reasonable cost" from many of the major libraries
such as the British Library or la Bibliotheque Nationale de France.
Mace originally published his work, unbound, in loose leaf, to subscribers at a cost of 12 shillings. He complained that - at this price - he could
not make a profit and was going to have to raise the price accordingly. Interestingly, the sum of 12 shillings sterling in1676 is today equivalent to
about $152 - according to http://measuringworth.com/calculators.jdowning - 1-20-2008 at 06:23 PM
I should add that the scorched paper strips that I used for the gluing trials measured roughly around 30X20mm so were about three times the area of
the strips used by Mace. The strips in my trials, therefore, required more scorching heat to set than those of Mace with the possible consequence of
potential overheating and weakening of the glue in the joints under test.Jameel - 1-20-2008 at 06:58 PM
Again, interesting tests and photos John. Thanks also for you opinions on neck joints. This is a lot of reading, and a lot of effort on your part.
Thanks very much. Incidentally, using alcohol for hide glue is new to me. Never heard of that before. Is it explosive? How does it not evaporate very
quickly?jdowning - 1-21-2008 at 06:20 AM
Thanks Jameel - I suppose that the alcohol alters the chemical composition of the glue in some beneficial way so that any residual alcohol vapor
remaining is unlikely be a safety concern - although the early recipes recommend corking the glue container after use, loosening the cork when heating
the glue each time as well as adding more alcohol if necessary to make up for losses due to evaporation. This is no different from a water based glue,
however, as the water is, of course, also prone to evaporation.
Note also that I am only testing an alcohol based hide glue not the isinglass/hide glue combination recommended in the early recipes. The alcohol
based combination glue might prove to be significantly stronger.
At least the alcohol based glue is no weaker than the water based and may prove to have a longer 'shelf life'.
Note that only dilute ethyl alcohol is specified in the early recipes. This is most economically and readily available (as it was in the 19th C and
earlier times) as alcoholic spirit beverages like gin , brandy etc - usually sold at 40% alcohol by volume. Other commercial alcohols like methyl
alcohol (methanol) or propyl alcohol (rubbing alcohol) may not work but I don't know.
Interesting what can be found by exploring other early technologies.jdowning - 1-23-2008 at 08:46 AM
I came across this reference the other day confirming use of alcohol based hide/isinglass glue for instrument making by luthiers Stephen Barber and
Sandi Harris. See under 'Glues' at http://www.lutesandguitars.co.uk/htm/cat.htm
I am not clear from their website posting if they use an alcohol based combined hide/isinglass for all of their instrument building work or just for
specific purposes such as inlay work or perhaps high strength applications such as gluing bridges. Therefore, I have contacted them for more
information. If I get a response I shall report any further clarification on this thread.
Looking again at the video clip recently posted on this forum, showing Labib Habib in the process of making an oud, it is interesting to note that in
constructing the bowl he uses a solid mold but that the ribs are not assembled in contact with the mold except at the neck block and end block. In
effect, despite use of a solid form, he is building the bowl using the open, mold free, method recommended by DR. Oud in his book "The Oud
Construction and Repair". So why does he use a solid mold when a simple frame to hold the neck and tail blocks in position would do?
The video shows Labib Habib applying hot hide glue to the edge of a rib (using a wooden spatula - not a brush) and then a shot of him 'ironing' the
completed rib joint with a hot iron - presumably to remelt and consolidate the glue in the joint. Interestingly there is no sign of any paper strips
or other clamping devices (pins, nails, elastic bands etc) being used to hold the ribs in position until the glue sets.There are, however, what seem
to be numerous pin or nail holes in the surface of the mold indicating that pins may have been driven into the mold to temporarily hold the ribs in
place as bowl construction proceeds. The scene in the video may, therefore, have been 'staged' for the camera and does not represent the complete
assembly procedure? Has anyone seen Labib Habib (or other Arab luthier) at work to confirm this assumption?
Richard Hankey has posted another video clip of an oud maker at work on his website. The video quality is not very good but this time the luthier is
using an open bulkhead or 'toast- rack' style of mold and appears to be hot ironing the rib joints without use of any temporary paper strips or pins -
he just seems to be holding each rib in place until the hot iron causes the glue to quickly set? Again the ribs do not seem to be closely fitted to
the mold bulkheads - perhaps because this would avoid risk of the ribs becoming stuck to the mold?jdowning - 1-23-2008 at 11:03 AM
I have been able to view the video clip of Labib Habib in finer detail by using a flash video player with a screen capture facility. (A nice free
viewer of this type is FLVPlayer4Free). Examining the still shots it is now clear that the Labib Habib drives nails into the mold along the edges of
the ribs so that the nails press against the ribs temporarily holding the joint together while being, fitted, 'ironed' to reheat the glue and until
the glue is fully cured and hardened.jdowning - 1-31-2008 at 07:05 AM
Here is another reference to alcohol being used in preparing hide glue - from an earlier source than the 19th C.
In his book"Mechanick Exercises or the Doctrine of Handy-Works" London 1703, Joseph Moxon - in the section dealing with the art of Joinery, page 105,
section 34 'Of Chusing and Boiling Glew' describes the preparation of hide glue. He finishes by stating:-
" When (with often heating) the glew grows too thick, you may put more water to it; but then you must make it very hot, lest the glew and water do not
wholly incorporate.
Some Joiners will (when their glew is too thick, put small beer into it, thinking it strengthens it: I have tried it, and could never find it so, but
think it makes the glew weaker, especially if the small beer chance to be new, and its yest (i.e. yeast) not well settled from it, or so stale, that
it be either draggy, or whit mingled with the settlings of the cask".
"Small beer" in Moxon's day was a weak beer - a common drink for everyone, regardless of age, at a time when water was often polluted and dangerous to
drink. It was likely about 2% by volume of alcohol (or less) in strength so (being mostly water) it is perhaps not surprising that it had little
effect on glue strength compared to simply adding water. Nevertheless, it would appear that English woodworkers of the time believed that adding
alcohol to reheated hide glue made the glue stronger. Speculating, it is possible that these woodworkers were following a traditional method handed
down from an earlier period but had forgotten two critical details, in the passage of time - that the alcohol used must be distilled spirits of wine
(of much greater strength than that of small beer) and that addition of the alcohol acts not to strengthen reheated glue but prevents it getting
weaker with repeated reheating- alcohol being an effective sterilising agent preventing bacterial decay.
While historical record confirms that hide and fish glues were used for instrument construction since ancient times, we do not currently know if
alcohol modified glues were ever used for this purpose. Nevertheless - even in Muslim cultures where alcoholic beverages have always been forbidden -
ethyl alcohol has been prepared from wine by distillation since at least the 8th C - for medicinal use or employment as a chemical. It is, therefore,
quite possible that alcohol modified glues may have been in use since ancient times.
To divert a little, to complete the picture about early glues, Cennini in his book "Il Libro dell' Arte" of 15th C Florence, Italy mentions
preparation of a glue made from cheese used by woodworkers. Cennini does not say, however, that cheese glue was used for instrument work - as he does
for fish and hide glues.
An earlier reference to the making of cheese glue is recorded in the 8th C. "Kitab al-Durra al-Makmena (The book of the hidden pearl) - a book about
glass making by the Arabic alchemist Jabr ibn Hayyam.
A modern version of cheese glue - casein glue, made from milk protein - was used (with some unfortunate and deadly results) in the manufacture of
wooden aircraft frames during the first part of the 20th C. These glued joints were found to fail under load if exposed to moist conditions due to
bacterial decomposition of the glue. The solution was to treat the glue with an antibacterial agent to prevent this decay.
On the other hand, the cheese glue of the ancients, does not appear to have suffered from this problem. Perhaps these glues were treated with alcohol
as a sterilising agent to help kill any bacteria and perhaps this is the origin of alcohol based non synthetic glues? Just more speculation on my
part, however.jdowning - 2-1-2008 at 10:23 AM
FISH GLUE - some more information.
It is important to note, in considering fish glue, that there are three main types - isinglass made from fish swim bladders, glue made from fish skins
and liquid fish glue. Liquid fish glue is made by boiling fish waste - skins, bones, heads etc - to produce a concentrated liquid. It is not comprised
of pure gelatin fibre like isinglass and consequently, is flexible and weaker. Like liquid hide glue, liquid fish glue - which is readily available -
is unsuitable for making or repairing instruments.
Early historical records about fish glue refer primarily to isinglass - made from the swim bladders of sturgeon or other fish - except where use of
fish skin is specifically mentioned.
Pliny the Elder in his 'Naturalis Historiae' of 76 AD mentions both types in Book XXXII Chapter 24.
Checking out other early technologies, the powerful and compact composite reflex bows used by the Turks, Persians and Mongols were made of layers of
wood/sinew/horn glued together. The ancient Mongol bows (used in warefare up until the end of the 18th C) used fish glue made from the swim bladders
of freshwater fish.
According to L.F. Salzman ("Building in England down to 1540", Clarendon Press, Oxford) fish glue was used by English woodworkers during the 14th C.
Records show that in 1348 the sum of 18 pence was paid..." for 100 soundes for making glue for the carpenter" and in 1352 the sum of
3 shillings was paid ... "for 100 greylyngsondes for joining boards". In 1358 the hefty sum of 2 shillings was paid for "25 fisshesounds". A 'sound'
is a fish swim bladder. A 'greyling' is a species of English freshwater fish. Perhaps the expensive swim bladders of 1358 were imported sturgeon
sounds?
Thorpe's Dictionary of Applied Chemistry of 1941, Longmans, London, Volume V under "Gelatin" states that isinglass could also be made from fish
species other than sturgeon - such as Cod or Hake. The article goes on to say "In preparing Russian isinglass, the bladders are first cleansed from
blood in hot water, cut open, washed and exposed to the air with the delicate silvery inner membrane turned upwards. This membrane (i.e. the
isinglass) is stripped off, kneaded in damp cloths and dried."
Presumably, the swim bladders of fish other than sturgeon may also be prepared in this straightforward manner to make isinglass?
Now to find a local 'ice fisherman'!!jdowning - 2-1-2008 at 01:37 PM
Post script.
Csaba Grozer - a Hungarian bow maker - makes replicas of early composite reflex bows using traditional materials and techniques. http://www.grozerarchery.com
From experience, he claims that isinglass glue is the best glue in the world for making this type of bow quoting the properties:
1- it is very strong.
2- it does not become brittle with time.
3 - it has a long "open time" - that is it stays fluid and workable without gelling quickly when being applied (this allows the bowmaker to thoroughly
saturate the sinew filament layers before the glue sets).
All properties that make the glue suitable for instrument making it would seem to me.
Incredibly, some 19th C researchers immersed many old and valuable bows in water so that they fell apart - allowing the method of construction to be
"investigated". Bad news for history but at least it confirms that isinglass glued joints are reversible like those of hide glue.jdowning - 2-2-2008 at 03:09 PM
Post- post Script.
Continuing with my investigation into glues used in early Asian bow technology, the early 16th C Arabic treatise "Book on the Excellence of the Bow
and Arrow" translated to English by Faris and Elmer as well as a study in English of the early 15th C "Kitab ghunyat at-tullab fi marifat ramy
an-mushshab" are fascinating and detailed accounts of archery practice, bow construction etc. of the time.
For those interested, these and other texts can be freely downloaded from http://www.sacred-archery.com
Neither of the texts gives details of the glues used - with one exception. In Chapter XXXII of the 16th C treatise "On the Variations in the Length
and Construction of the Arab Bow", the anonymous author states that "Others have recommended that in countries of excessive heat the sinew should be
saturated with glue made of the best parchment, which is characteristically moist and, therefore, suitable for hot regions but not for those which are
cold and humid".
Hide (ie parchment) glue would , therefore, seem to be an exception to isinglass glue which would seem to be more suitable in cold and humid
conditions.
Not much to go on but every little piece of data can help in building a better understanding of the overall historical picture.
It is unfortunate that the surviving early texts about the oud do not go into the same immaculate, informative detail recorded by those who wrote
about weapons of war.jdowning - 2-15-2008 at 09:54 AM
Just an additional thought on glues.
The early technology of Arab composite bowmaking was - like many early technologies - highly sophisticated. The ratio of materials used in the
construction of a composite bow - wood, sinew and horn, and quantity and type of glue - was adjusted by the bowmaker depending upon the climatic
conditions and season of the year when a bow would be used. So, for example, in regions where the climate tended to be cold and damp, "like Syria and
Andalusia", suitable bows were made with a greater proportion of horn, a moderate proportion of sinew and glue and a small proportion of wood.
It is interesting to speculate, therefore, if the bowmakers also mixed hide and isinglass glues in different proportions to arrive at an optimum glue
for a particular climate. Perhaps this could be the ancient origin of the hide/isinglass compound glue recommended by Mace as being best for lute
repairs in the cold and damp climate of 17th C England?jdowning - 3-1-2008 at 08:54 AM
To change direction a bit - there has been some discussion about edge banding styles on the forum over the past year or so. Here are some rough
sketches to summarise the situation. The sketches represent sections at the edge of a soundboard and show five banding possibilities. There may be
others that I am not aware of.
Banding is not only decorative but reinforces the edge of the soundboard against damage and may also improve the response of a soundboard due to the
weakening at the edge due to the banding - just as the edge purfling on the violin family of instruments seems to, for example.
jdowning - 3-1-2008 at 10:33 AM
Sketch A and B, of course, represent soundboards without banding.
A - is to be found on surviving lutes of the 16th C. so, no doubt at one time it was to be found on ouds as well. I have not found the lack of banding
to be a disadvantage on a lute as far as edge protection is concerned.
B - here the edge of the soundboard is protected by a thin strip of material glued in place which reinforces the glued joint and helps protect the
soundboard edge. This is found on lutes of the late 17th C/ early 18th C - known as a "lace" - and is mentioned by Thomas Mace as being made from thin
parchment or silk.
There is some evidence that a type of lace was also used on ouds - at least covering part of the sound board edge.
The attached image shows an old oud soundboard of unknown vintage where the banding does not - for some unknown reason - extend to the lowest part of
the soundboard, this area being covered by a narrow strip of 'leather cloth' glued in place (here the cloth strip has been removed). Ronny Andersson
has also recently mentioned coming across examples of this type of edge reinforcement.
jdowning - 3-1-2008 at 11:10 AM
C - here banding is glued into a shallow rebate cut around the edge of the soundboard. This type of banding arrangement can be found on surviving
lutes of the late 16th C and 17th C. - usually the rebate is cut to half the thickness of the soundboards or about 0.8 mm.
Dincer Dalkilic recently commented that he uses this method of banding and that it was traditionally used by some of the old oud makers. The depth of
the banding can be half the thickness of the soundboard - like a lute - or cut to almost the full thickness of the soundboard so that the banding
appears to be full depth.
The attached image of an old oud soundboard is an example.
D - here the edge of the soundboard is completely cut away to receive the banding which is the same thickness as the soundboard. A method used by
Richard, Jameel and, no doubt, many other makers today. Was this method used by oud makers before, say, the 20th C?
F- this is mentioned by Jameel as being used by one maker (if I have interpreted Jameel's description correctly) - so may be a unique modern
application?
Note that for examples C, D and E it may be impossible to determine - from an external examination alone, without removing a section of banding -
which method is being used.
jdowning - 3-19-2008 at 12:59 PM
Turning now to the oud 'tail block' - this is the block or plate to which the ribs are attached at the bottom or tail of the bowl usually made from
pine or other resinous soft wood. There would appear to be two possibilities which, because they occur in early lutes, are an indication that they
were likely also used in early oud construction.
The attached images are some rough sketches - more or less to scale - showing vertical sections through the tail end of oud and lute bowls and
including the bridge and first bar above the bridge.
Sketch A - is the familiar Arnault de Zwolle lute of the 15th C which has a fairly massive tail block. The tail block forms part of the construction
mold to which it is temporarily attached. This form of construction is found in late 19th C ouds and is used today - although the thickness of the
block is usually, proportionally, about half that of the de Zwolle lute.
Sketch B - is a section through an early 17th C lute. There is no tail block. Instead, the bottom of the bowl is reinforced, on the inside, with a
fairly shallow, thin strip of wood glued in place - the end plate. The end plate extends over most of the circumference of the bottom of the bowl.
Further reinforcement is also provided with an 'end cap', glued to the outside of the bowl - covering approximately the same area as the end plate.
The end plate does not form part of the mold but is fitted and glued in place after the bowl has been removed from the mold. The thickness of end
plate and cap is typically about 1.5 to 2 mm.
Sketch C - for comparison, shows the construction of an old oud that I own - of unknown vintage - which has a lute like end plate that appears to have
been glued in place after the bowl had been removed from the mold. There is a semicircular end cap glued to the outside of the bowl. This cap is
primarily decorative.
jdowning - 3-19-2008 at 01:42 PM
The attached images show the oud end plate and cap represented by sketch C. The end plate is about 2 mm maximum thickness (tapered at the ends), 72 mm
maximum depth and about 240 mm in length. Note that - unlike on a lute - the end plate does not extend to the top edge of the bowl so is not glued to
the soundboard.
The semicircular cap is 30 mm in diameter.
jdowning - 3-19-2008 at 01:46 PM
Just for comparison, the end cap on this replica of an early 16th C lute is 45 mm maximum depth.
jdowning - 3-29-2008 at 05:33 PM
During 1980, while living in Scotland, I had the opportunity to examine in detail a number of early instruments in the Charles van Raalte collection
at Dean Castle, Kilmarnock. Instrument cat#32 in the collection is a 16th C ivory lute by renowned luthier Marx Unverdorben. Although converted to a
lute-guitar during the 19th C, the bowl and neck of the lute would appear to be original. The instrument is unusual in having a relatively narrow neck
- of oud like proportions - and ribs decorated with an 'arabesque' design in gold leaf.
jdowning - 3-29-2008 at 06:04 PM
Some distortion of the bowl and soundboard were recorded in my notes and sketches taken at the time the instrument was examined. In particular, there
was a noticeable 'dip' in the soundboard - amounting to about 6.5 mm at the bottom edge of the rose - measured from a straight edge laid across the
belly. At the time, this 'dip' was assumed to be due to distortion over time.
However, in "Historical Lute Construction", Robert Lundberg notes that this 'dip' is to be found in many surviving original lutes and was intentional
- the edges of the bowl being 'scooped' to cause the soundboard to deflect downwards when glued in place - the maximum deflection occurring at the
lower edge of the rose. Richard Hankey (DR. Oud) in "The Oud Construction and Repair" also mentions this as a feature of oud construction.
antekboodzik - 2-28-2017 at 03:22 PM
May I refresh this topic and ask some questions even if it is nine years old?
Recently, I was taking hard time figuring out, what is going on with "paper strips method", and trying to reproduce it working to some of my work.
Every time when gluing lute ribs I found it difficult to have the edges of two ribs being joined perfectly levelled to each other along the lenght of
the joint (gluing line). These were very small unevenness, but hard to overcome even with stretching sellotape across the joint, or with putting
gently barrel pins at an angle along the joint line.
What I understood from this topic, is that those paper strips are literally "pinched" to the one rib, then the iron is rolled immidiately to the
second rib, putting the jointline as level, as it can be, and it "'pinchs" the strip to the second rib, holdig them together. Am I right? Please,
corrct me, if I got it wrong.
But what I got, it was paper strips glued every time to the sole of my iron, not to ribs even...
Please, help jdowning - 2-28-2017 at 06:56 PM
You will find details of an oud bowl being built step by step using this technique here, posts starting at 6-2-2009
Use heavy gauge paper and do not soak the paper, just apply glue to one side of the paper and allow to dry partially to identify grain direction -
otherwise soaking wet paper will tear and stick to the iron!antekboodzik - 3-1-2017 at 02:57 AM
Shall I wait some time for every strip to be set or proceed one by one, gluing and "patching" whole lenght of the joint at once?jdowning - 3-1-2017 at 06:05 AM
As mentioned in the link - apply hot hide glue along the rib joint and glue the end of the rib to the neck block first. When working with hide glue
the glue will quickly gel so must be reactivated bit by bit as you work using a hot iron (it is the same situation when gluing a sound board in
place). Once the end of the rib is fixed to the neck block - as indicated in the thread - move along the rib joint bit by bit. Re-melt a short length
of glue in the joint, press the joint firmly together and hold in place with paper strip 'scorched' in place. Once the paper is scorched in place it
is already set and does not need to dry further so move on right away until the whole rib joint is completed (it takes just a minute or two) - covered
in paper strips along the whole length of the joint. The paper strips remain in position - covering the bowl outer surface as temporary reinforcement
when handling - until the finished bowl is removed from the mold and the usual internal paper strip reinforcement is glued in place.
Of course all ribs must be well fitted together and on the mold so that no forcing is required.
No one is perfect, but there is a minimum. A musical instrument is not a scratch
paper.