bongosnotbombs wrote:The way that works best for me is Titebond glue, I use Titebond II because it is waterproof.
The function of a wood glue is as good as you get the pieces pressed together. But(!) both surfaces have to be wood - not old and dried glue.
I do not know if you get Titebond in the Czech republik. As I know there is one sort …
… which is said to be similar to traditional hot glue (don´t know if this is the correct word) while having all qualities of modern wood glues. It´s instruction says that you can also use it for repairing chairs without removing all rests of the original glue (because the glue needs rather long to harden and therefore is able to creep into all spaces). But those informations are often wrong.
bongosnotbombs wrote:Then use the rope with a stick inserted to cinch the rope really tight.
bongosnotbombs wrote:I've closed probably 20 cracks this way and never had one open back up, nor have I experienced cracking in another location. I also dove right into the repair with only a little bit of research. I came up with the rope and stick method on my own.
These experiences make your arguments rather strong. Nevertheless I want to describe the constructive problem (as I see it) with two graphics, which are probably interesting apart from Vasik´s concrete damages.
The doves´ shapes, if not bent, are cutted at a band saw. Then every single dove gets it´s angles from a circular saw, while being fixed on a sliding bed. This sliding bed holds the doves horicontally in the exact position as they later will be glued together (as parts of the conga corpus).
When you cut every dove separately at the band saw there will always be tolerances. So the thicknesses vary, while the inner width of all doves are identical [- b -] (because of their fixation on the sliding bed). But when the doves get glued together, they get pressure from the outside (the principle of constructing vaults). And this can cause shifts like my second graphic shows [- A2 - B2 -]:
You can see, how already one shift causes tension. In my graphic I just show one gap on the opposite side [- C -], but such a tension disperses all around the whole corpus. In an advantageous distribution the doves´ variancies neutralize each other more or less.
The described rope-and-stick-method reconstitutes the tension. And possibly there is another bonding, which is the (next-) weakest. Therefore epoxy would be a good solution, because it adheres perfectly, can be mixed very fluidly (in order to disperse in smallest gaps - I am not sure, but there happens also a capillary action), fills gaps, is a resin substance and can be dyed.
p.a.dogs1