"Arfa Daily" hath wroth:
Ignoring acid flux, the rosin fluxes come in two flavors. Water based and those that require some solvent to remove. Water based fluxes can allegedly be left on the board and will not eat the copper traces. The inspiration for this was not to reduce cleanup costs, but because of environmental and workplace regulations which proscribed the use of chlorinated hydrocarbon solvents for board cleaning.
The problem is that water soluble fluxes require rather hot water to remove properly. They're mostly used wave solder machines and not hand soldering. However, we were using them on the production in rework stations and of course, I ended up with a few rolls of the stuff. That's when I discovered that if you leave the flux on the board for more than a few days, it turns rock hard and no amount of hot water will get rid of it. The residue is water soluable, but the big lumps left in rework are not. I experimented with some household cleaners and found the ammonia cleaner sorta works. It's often easier to scrape the ossified flux off the board than to wash it as it's quite brittle.
Another fun experiment was to make my own flux. I managed to accidentally purchase a roll of solid core 63/37. No flux inside. I could purchase a tub of Kester rosin solder paste, but that's too easy and no fun. Rosin is just tree sap and I live in a forest with pine trees and lots of sticky gooey sap. I just walk outside, scrape off some pine sap, and I have instant rosin. I melted a big lump of the stuff to clean it and boil off the volatiles, and voila, instant sticky rosin flux.
When testing it, I found that the soldering iron was insufficient to set the flux on fire, but did an adquate job of converting it into noxious fumes. Rosin works by vaporizing into a cloud of inert smog, which protects the tin and lead from oxidation. The soldered connections were successful, but difficult to see under the carbonized sticky mess. Removal required trichlorethane (auto brake cleaner) and some elbow grease. Once cleaned, the connections looked quite shiny and bright. Obviously, I missed a step in the production process, but in general, it worked.