I am only familiar with the Balbach-Thum silver refining cell (nitrate solution, horizontal cathode on bottom of rectangular cell) so my comments only apply to that system. I'm not a silver refining expert either, but here are my thoughts.
Silver dendrite formation is the norm in Thum cells. Compact silver deposits are unusual (and often the result of serious impurities in the system). Dendrites are often considered a "good thing" for that cell since it makes it easy to harvest the silver. A serious drawback of dendritic formation is the "wild dendrite" where a dendrite grows vertically, very quickly, until it reaches the anode and short circuits the system.
Silver dendrites usually form because of hindered nucleation - silver metal particles want to form but their initial appearance requires a lot of energy to create that surface area, so they don't form until there is a large overabundance of tiny nuclei at that location and a large local concentration gradient. Then the crystals form too quickly in haphazard ways, and grow randomly onto other needle tips that were previously formed, making dendrites.
What is worse, when silver nucleates at the cathode surface, it creates a nucleation exclusion zone which prevents nearby silver ions from nucleating. So it is a bit of a vicious cycle - nuclei don't form easily, and when they do they prevent other nuclei from forming on the cathode. The silver instead nucleates on other silver micro-needles, which eventually grow into big needle-like structures aka dendrites.
If you are interested, there are a ton of academic papers on hindered nucleation and dendritic growth for silver that you can find by searching any of the academic databases.
People have tried to solve this issue in the past. Two approaches that work include:
1. stirring - there is a large concentration gradient in a Thum cell between the anode and cathode, and there is little chance for nuclei to contact other nuclei and form a compact nucleus that can then deposit on the cathode as a compact deposit. So people have tried various stirring schemes. Pumping was, as far as I recall, the most successful - a tube near the bottom corner draws solution up, through a pump and exhausts at the surface. Unfortunately stirring can also allow impurities that wouldn't deposit with the silver to be trapped as the silver deposits, reducing final purity. Too much stirring can cause actual particles from the anode to be released into the cell, rather than remain in the anode bag. A pump that withstands the Thum solution is expensive. It can also introduce oxygen and extra dust into the solution.
2. adding nucleation agents or precipitants - things like small phosphate additions have been successful in accelerating nucleation in Thum cells. Again, this introduces impurities, and the normal silver refining system doesn't particularly tolerate phosphate ions in cathode silver. It can also make very hard deposits that are difficult to harvest.
Industrially, most silver refineries using Thum cells will use "rapping" or vibration to control dendrites (so no wild dendrites form). Rapping means hitting the cell with a hammer to jar it. Vibration uses a vibrator that shakes the cathode from time to time, causing wispy dendrites to collapse.
Hopefully others more familiar with the other silver refining cells will chime in with their expertise.
Best Regards, Geraldo
