Some specifics of the Torres model construction

The beauty and simplicity of Torres' rosettes remain an inexhaustible source of inspiration.

As Torres did, I assemble my rosettes directly in the soundboard, which makes it easy to vary the design from guitar to guitar.

I use traditional animal glues throughout construction, as period guitar makers did. I mix hide glue and bone glue in different proportions depending on the operation.

For top bracing, I use mostly bone glue with about 20% hide glue. Hide glue exhibits greater shrinkage during curing and can introduce tension that may distort the thin spruce plate. However, this same shrinkage is advantageous when gluing bindings, as it helps pull them tightly into place during drying.

For most neck and body glue-ups, I typically use one part hide glue to three parts bone glue.

For the solera, I use black walnut because it is stable and easy to work. The neck lies in the same plane as the soundboard, and the proper action is achieved by planing a taper into the fingerboard.

I used a convex-sole plane followed by a scraper to scoop out the cavity that receives the domed top. I did not use templates; instead, I aimed to achieve a smooth, even curve in all directions.

One of the characteristics of the early Spanish school that appeals to me is the so-called free-form assembly method. The instrument is assembled on the solera without a rigid mold to force the sides to follow the plantilla precisely. The sides are bent directly to the outline on a hot pipe, which results in slight asymmetry — almost invisible to the eye — but, in my opinion, contributes to the instrument's character and individuality.

Although more demanding to execute, this method offers great flexibility and makes it easy to vary plantillas, scales, and other parameters.

In guitar making, the thickness of the top, back, and sides — together with the wood species — largely determines the final tonal quality of the instrument.

After studying the data in Romanillos' catalogue and available plans (including guitar SE 114, which belonged to Tárrega), I observed that Torres tended to use relatively thick backs compared with other makers. The sides, by contrast, are very thin — probably to facilitate bending. The soundboard shows variable thickness: thicker in the central area and thinner toward the periphery.

In effect, Torres created a light, responsive structure supported by a solid foundation.

• A solid back contributes to power and projection
• Thin sides help lower body resonance
• A thinner top provides exceptional tonal richness

The challenge is finding the right balance in the top thickness. Each piece of wood is unique, and tactile feedback combined with experience helps determine optimal flexibility.

I glue the bridge "in the white," which allows me to play the guitar and evaluate the tension and tonal response before finishing.

One of the key features of Torres' architecture is the doming of the lower bout of the soundboard. The added rigidity allows for a thinner soundboard and lighter bracing. Doming also improves structural stability and helps reduce the risk of cracks during climatic changes.

Radial struts, glued to the soundboard in the scooped solera, help maintain this dome.

In Torres' work, the transverse bars are not scalloped at the ends; their height remains constant along the entire length. All bars are anchored to the sides with triangular blocks.

• Increased torsional strength
• Reduced stress from rib deformation
• Improved structural integrity, especially with thin sides

Today, the Spanish method of attaching the top commonly uses individual blocks (tentellones) glued one by one. Torres, however, used continuous kerfed linings for both top and back in most of his guitars.

I prepare the kerfed linings using simple hand tools and glue them with homemade clamps.

Most modern luthiers first glue the transverse bars to the back, then fit and glue the back to the sides. Torres' method differs.

In his approach, the back bars are first positioned on blocks already glued to the sides. This stabilizes the structure. Once the bars are glued, they can be shaped to define the back arch. Glue is then applied simultaneously to the bars and linings, the back plate is positioned, and the assembly is tightened using cotton string — typically about 100 meters of it.

When working with animal glue, the joint must be reheated for proper bonding. Traditionally, this is done with a small alcohol flame applied briefly to the joints. The burning time is sufficient to warm the wood and activate the glue.

Closing the box is critical to the final sound. This method respects the wood's hygrometric state without introducing unnecessary stress. The structure is already strong enough to prevent deformation.

The downside is that the process is technically demanding, and glue squeeze-out is nearly impossible to remove afterward.

A single full-depth groove for both purflings and bindings increases the rigidity of the body. The purflings are made from multiple hardwood veneers laminated with animal glue, creating a very strong rim structure and offering a good alternative to solid hardwood linings.

For very large purflings exceeding 4 mm, however, a separate ledge is preferable.

Both the back and the bindings/purflings are glued using the traditional string method.

This traditional method uses string and wedges to clamp the bridge. The top dome is made slightly greater than the bridge arch, allowing perfect contact across the full length when pressure is applied at the wings.

I lightly scrape the bridge's gluing surface with a convex scraper to create slight concavity. This helps prevent the thin bridge edges from lifting. The same principle can also be applied when gluing the fingerboard.

The fingerboard is glued with hide glue.

The pores are filled with pumice, and the guitar is fully French-polished with button shellac.

The finished instrument reflects my ongoing dialogue with Torres' legacy — respecting traditional methods while interpreting them through personal experience and modern understanding.