When you make a solder joint you are making a gradually-changing alloy between the solder on one side and the metal wire or terminal on the other. To do this the solder has to be hot enough to melt - and the terminal or wire has to be hot enough that the surface melts into the solder. This means the terminal or wire has to be at least up to the melting point of solder contaminated with a nontrivial amount of the metal of the wire.
It's REALLY easy to not get the terminal hot enough. Then the solder molds itself into the surface irregularities of the wire or terminal, but doesn't alloy with them. This is one type of "cold joint". It will show as connected to an ohmmeter at an ordinary setting (though it might show a noticable voltage drop in the drive-current-measure-voltage-drop resistance measurement configuration).
With this kind of cold joint, the resin's thinner may work its way into the microscopic crack and re-insulate it, corrode it into non-conductivity, or shrink away and pull the wire just off the surface.
When soldering you also have to hold the wire, terminal, and solder blob absolutely still as the temperature drops through the temperature band where some of the metals of the solder mixture are liquid and some are solid. If you move it even microscopically while it's at this temperature, the solid and liquid components separate and you get the pasty form of a cold joint. This will also test "connected" in ordinary ohmmeter mode, but will also be pulled apart into non-connectivity by mechanical stress, and have micro-cracks that can be infiltrated by
Solders tend to have a composition near - but usually not quite at - the "eutectic" mixture, where the melting point of the higher melting temperature component is brought down to (eutectic), or near (near the eutectic mix, or with the additional error from the foreign metal from the wire or terminal), that of the lower melting temperature component by quantum mechanical interactions. The closer to the eutectic mix, the narrower the temperature gap, and the less time the new joint spends in the vulnerable temperature region. But when you have a heavy piece of metal you're soldering to, such as a bolt that will be your terminal, it cools slowly. So even with the narrow vulnerable temperature band it spends a lot of time in the bad temperature range. So you need to use a jig to hold it, rather than holding it by hand, so it will stay still until the joint has solidified.
With delta all six wire ends were connected to terminals. It looks like at least three of the joints are "cold" - or all three windings got pulled apart by potting compound shrinkage.
If I had this rotor to salvage, the first thing I'd do is dig the potting away from the terminals and the first half-inch of wire. Then I'd check whether the coils were intact. If so I'd make new solder joints to new terminals, make SURE they're good joints, then re-pot the terminals. (Or just bring out the wires and put a connector on them, rather than potting a nut to be the terminal an thus burying the joint in the potting again.)
