Old Project - New Lute

With the plane repaired, tested and back in commission, blade razor sharp and the throat set to minimum gap, the remainder of the first test batch of eight ebony fillets was completed. Apart from the first two fillets, the thickness of the remainder was held to a consistent 0.035 inch, +/- 0.001inch (0.89 mm) using the plane. A satisfactory result. The first two fillets showed a slightly wider variation in thickness (undersized to 0.003 inch maximum) - due probably to using a scraper blade for the final cuts, which results in removing too much material in places. Only the plane, which provides much better control, than the scraper blade, will be used in future.
The depth of each fillet was reduced to about 2 mm using the same process i.e. using the thicknessing block and plane. The objective here was to remove all traces of saw marks and produce a smooth finished edge (as well as a uniform depth). Any remaining saw marks might act as a stress concentration point that could induce splitting of the brittle ebony during hot bending.
Some scrap ebony fillets from this batch were used for preliminary hot bending tests.
My new, electrically heated bending iron (see 'Waste not Want Not" on this forum) is provided with slots for bending the fillets. The slots are designed to support the sides of the fillet during bending - to minimise 'buckling' of the fillet and to provide a better heat distribution on both sides of the fillet in order to facilitate the bending process.
An adequate curvature of the test fillets was achieved by this means. However, I am not happy with the new bending iron as it is 'under powered' and needs a higher wattage heating element. Warm up time of about 30 minutes was too slow and bending of the fillets was also very slow. Insufficient heat may cause 'forcing' of the work and result in breakages. Also this bending iron is only 2.5 inches (64 mm) in diameter - which is a little on the small side).
I have, therefore, decided to modify my propane heated bending iron to provide slots for bending the fillets. This bending iron is fast to heat up, easy to control and has a quick heat recovery. Also the heating surface has a diameter of around 4.5 inches (115 mm) which will allow better heat distribution, a more uniform bend and , consequently, reduce the risk of fillet breakage (hopefully!).
As a result of these tests I shall proceed with my plan to use ebony inter rib fillets.

My 'heavy duty' propane heated bending iron has now been modified to prepare for bending the ebony, inter-rib fillets.
The heating surface of the bending iron is a piece of thick walled (3mm thick) seamless brass tubing that has been formed into an oval section by squeezing it in a vice.
The channel, designed to support the fillets against buckling during bending (they are 0.9 mm thick by about 2mm wide), was made in two parts from 3mm thick copper. The first part - about 12 mm wide - was formed to match the curvature of the heating surface by hammering on a tinsmith's anvil (an iron tube or round bar would have done just as well). The inside edge was then filed flat and square and the piece rivetted to the end of the heating tube with three copper rivets.
The other part of the channel was made about 22 mm wide, formed to the curvature of heating tube and the inside face made square and straight to match the fixed component. This part is adjustable - held in place by two brass screws - the heating tube being drilled and tapped to accommodate the screws. The screw slots allow the gap in the channel to be set from zero to about 8 mm. This will allow the fillets to be bent either singly or in bundles.
The attached images show the modification prior to filing and final finishing of the outer surfaces - a purely cosmetic operation but I have my pride!

The modified propane heated bending iron is now re-assembled and ready for testing.

The 'average' profile of a fillet was transferred from the mold using a draughtsman's "Flexicurve" (In practice each fillet profile will be different due to the flattened section of the bowl). The curve was marked onto a piece of pine, cut out on a bandsaw and smoothed to form a jig for the fillet profiles.
The small nails around the edge of the jig were intended for holding the fillets against the jig using elastic bands - in order to retain their shape after bending. However, I do not think that this will now be necessary - just an added complication.
Three of the prepared fillets were held together as a bundle for testing (with masking tape applied at the extreme ends) and the gap adjusted in the bending channel to just fit the width of the bundle. Bending three fillets at one time should be faster and provides some extra stability to the thin, deep fillets during bending. The edges of the fillets were examined for any flaws or saw marks that might act as stress raisers and result in breakages during bending. The idea was to position these flaws on the inside curve where the wood would be subject to compressive loads and be less likely to fail during bending. However, after messing around with the tape and other preparations, I still managed to end up with some flaws on the exterior of the bend - the worst case scenario! Also, a slight grain 'run out' was noted in the fillets which will not help either!

After bringing the bending iron up to temperature (about a minute) bending commenced - matching the curve bit by bit against the jig to obtain the required profile. Despite the concerns about possible breakage there was no problem in hot bending the fillets - although a bit of practice is going to be necessary for best results.
Despite having the support of the bending channel side walls, some slight sideways buckling of the fillets did occur in places. However, it is judged that there is sufficient flexibility remaining in the bent fillets - both laterally and longitudinally - to allow precise matching of the rib profiles when the fillets are finally glued in place.