RE: soldering
As to the question above, have you ever seen failures due to corrosion. The answer is an emphatic "YES" many times over. I've also seen where servos started glitching or show erratic behavior and, after demating and remating the connectors, the trouble disappeared. This was almost always traced to corrosion in the connectors, often so small as to be almost undetectable. The removal and re-engagement of the contact surfaces scrubbed off the corrosion. The current in the signal lead is so very small (microamperes) and the voltage so low that it takes very little corrosion to block the signal. If you are in an environment that occasionally has a high humidity, corrosion is an even greater problem than for those who live in an ideal climate. Even the gold plated ones can show this as the plating is very thin in most cases and not very durable. Even the poorest solder joints are usually a better risk. The solder pull test described above is not very pertinent, a solder joint is not noted for strength and should not be used for such. It is made for electrical continuity under normal conditions encountered in the field where tensile strength has little application.