A: I’ll assume you’re using the method which involves copper-plating the grains before placing them. But in case you’re not, here’s how to do it. This works much better than copper salt formulas when working with silver.

First put the grains in an iron container (A black iron pipe cap will do if you’ve got nothing better.) Then pour some used pickle over them. If you don’t have any, then use new pickle in which you’ve dissolved a little copper sulphate. Stir the grains around until they are well coated, then rinse and dry. They should be a dark copper color, though in actuality the copper layer is very thin.

Glue the grains down on your clean base piece with a mix of Batterns’ flux and hide glue. It should be diluted somewhat. Only use enough of the glue and flux solution to hold the grains on. You don’t want the grains floating on a layer of flux, after all—this alone could cause your problems. But a little bit of flux does help the process along, especially with silver.

I would guess that your adhesion problems are likely due to uneven or incomplete heating. Most especially, maintaining too oxidizing an atmosphere can cause problems. It can help to use two torches, one underneath the piece (or use a hot plate) and another from the top. Both torch flames must be very soft and brushy (reducing.) My favorite torch type for this work is the Prestolite air/acetylene type (Actually I use the Smith brand, but it’s the same fuel) This is one of the few cases in jewelry where acetylene gas, with its excess of carbon, is clearly superior, in my opinion.

The Prestolite type torch, with a large-size tip, but low gas pressure setting, produces a soft broad reducing flame with which you should attempt to completely cover the grains. The idea is to keep the copper from forming a black oxide. If it’s nicely metallic and shiny at the point when the temperature of the metal reaches the eutectic point, then the copper is free to join with the silver, bonding the surfaces. If the copper has become oxidized from too oxidizing a flame, then this bond won’t occur.

If you are heating both the grains and the base piece evenly, you should get good adhesion of all grains. If this seems to be the case, yet you’re still loosing grains, then reduce the flux, and heat longer, being certain to maintain that reducing atmosphere around the grains. The classical recipes (Littledale, etc.) for granulation relied on copper salts being reduced by the carbon in the glue. We don’t really need to depend on this, since that acetylene flame is quite capable of reducing copper oxide on the surface by itself, and the copper plating doesn’t need to be reduced, only kept from oxidizing. Thus you don’t need to use excessive glue or flux. You only need enough to keep the grains in place.

If you’re careful, there is little danger of melting anything, as the eutectic point is far lower than the silver’s melting point, and there is so little copper present in that plated surface that you can’t actually lower the melting point of the grains themselves. In this regard, the classical method—heating by firing in a furnace—is easier than using the flux/glue mix and a torch). If the temperature is even, and below the melting point of the parent metal, then it is easy to evenly heat-soak the item with no danger of melting anything. In addition to allowing enough time for the grains to adhere completely, a nice leisurely heating has another advantage. At these temperatures, copper is pretty mobile in the silver, and will nicely diffuse into the parent metal. Any that’s oxidized will stay at the surface, to be pickled off, but that leaves a poorer looking surface. So letting the copper diffuse into the surface with a longer gentle heating will improve the color of the finished surface and its finish. Also, by lowering the concentration of surface copper through this diffusion, you also reduce the chance of brittle joins due to excess copper or copper oxide at the joint.