Any silver has the potential to encrust and trap some gold. Generally with karat gold anything with 9% or less Silver content I would do a direct dissolve in aqua regia. But I say that from the perspective of a refinery equipped to properly assay the incoming lots and segregate them accordingly.
All scrap coming into a properly run refinery is melted and sampled, this way the customer, if they are there to witness, gets a sample as well as the refinery gets a sample. Your question pertains to a gold and silver alloy so only gold and silver will be tested for. When a refinery collects enough bars to process a refining lot, the bars are melted together based on the assay of each bar. All bars are melted and mixed well, sampled one more time, and poured into water to produce a shot which looks like popcorn or cornflakes and has a large surface area to aid in digestion. The thing you are trying to attain is a lot with 9% or less silver content.
When you are trying to refine a bar with say 20% silver, aqua regia will not work. But if you have enough other bars with less than 9% silver, you may be able to add this high silver bar to the melt because what you need is the end alloy which will be put into aqua regia to be 9% or less silver. If you make up a spreadsheet for your bars to be tracked, you can mathematically determine which bars can be added to a melt and still be safely under the 9% as long as you have gold and silver assays for each bar.
I have seen tumbling systems actually be more effective against silver chloride encrustation because the tumbling and grinding action actually breaks up any potential encrusting lumps of silver chloride. These systems can actually produce acceptable results with up to 12% silver content.
The silver chloride accumulated from these digestions is rinsed well and reduced to a metal which is melted into anodes for a silver cell. This is how most refineries recover any entrapped silver.
Usually a refiner will establish an "acceptable" accountability for a refining lot and if the lot produces the accepted quantity of expected gold, the silver chlorides are not processed individually. But until the results of a refining lot are quantified and accepted by the boss, the residues from the refining lot should be kept separate in case the accountability is not met. This makes isolating and finding any missing values easier.
In any refining operation of even moderate size, there is a quantity of gold that isn't worth chasing on a lot by lot basis. With the method stated above, large "losses" can be isolated quickly and the smaller "expected" losses can be processed in due time to recover any gold lost to the silver encrustation. That produces the most cost effective use of a refiners labor and time.
