"Although I have never had the need to produce silver powder as a product, I commonly produced silver powder as an intermediate refining step. The powder was then melted into bars. In general, this involved dissolving the impure silver in nitric acid and then precipitating the silver as silver chloride (AgCl), using either hydrochloric acid or a solution of table salt (if no mercury or lead are present, this is quite selective for silver, but the AgCl must be well-rinsed to remove any traces of dissolved contaminants). The AgCl was rinsed well and then it was reduced to silver powder using a solution of sodium hydroxide (NaOH) and Karo syrup. I have no idea what the particle size of the powder was. Here are some threads on this method.
http://www.goldrefiningforum.com/phpBB3/search.php?keywords=karo&terms=all&author=goldsilverpro&sv=0&sc=1&sf=all&sk=t&sd=d&sr=posts&st=0&ch=300&t=0&submit=Search
Many schemes for making silver powder involve first making AgCl (as above) and then using various means to reduce this to silver powder. Most use chemicals, but a dry method for this, using heat, is in this patent. To use this method, you would need an electronic furnace with accurate temperature control.
http://www.google.com/patents?id=Mj06AAAAEBAJ&printsec=abstract&zoom=4&source=gbs_overview_r&cad=0#v=onepage&q&f=false
On the forum, 4metals outlined a method for producing the powder directly from the silver nitrate solution without going through the AgCl step. Although I haven't used it, it sounds interesting.
http://www.goldrefiningforum.com/phpBB3/viewtopic.php?f=50&t=1275&p=39590&hilit=formate#p39590
Some methods are quite simple, but I think your main problem will be controlling the particle size, as it will probably be dependent on the variables, such as solution temperature. Also, different methods will likely produce different particle sizes. You could screen the powder and use what passes through, but the smallest common screen is 400 mesh, which is 37 microns. They do make smaller screen sizes, I think, but they are probably expensive and hard to work with."