Magnetic conduction: The magnetism doesn't just come out on the armature side but spreads into the body and around to meet the opposite pole of the armature. The reason armatures and bodies are made from varnished lamination is to limit losses through generating eddy currents within the metal. All testicle stuff this.
Ahhh, you confused me there. Did you mean "magnetic induction" , and what could possibly have happened to a simple Lucas yoke and pole piece to cause problems in this regard?
So, on the premise that the EMF generated can swing the AVO needle but not necessarily light a test lamp, the thinking now is that it's either a duff pair of armatures, ie unable to generate when rotated in a known good magnetic field, or a duff field coil in that it cannot create sufficient field strength to create enough residual magnetism in the pole?
I think re- doing the test with “independent” current going through the field with a real load on the armature will reveal the probable cause of the problem. Whether or not the field core retains residual magnetism is not relevant for that test, what IS relevant is whether current you THINK is going through the field coil is actually going to earth via a short in the field coil to “earth”, and we don’t know if you have checked for that short yet (unless I missed that). Even a near end of life field coil with 2 ohms resistant will create enough of a magnetic field to make a dynamo light a high watt bulb at 1000 rpm.
I personally think it’s extremely unlikely that the problems you are having have anything to do with residual magnetism in the field core. Given the tests you’ve done it’s more likely in the brush area, I recently found a brush with high enough resistance in the internal wire to carbon joint to prevent “start up” and that was only revealed when pushed I on it when the dynamo was rotating, and charging level increased.
It also might be worth checking the rpm you are testing the dynamo at, they run faster than you think, I’ve never worked it out but you need a relatively fast drill (1500rpm plus?) to spin them fast enough to get to “start up” rpm (eg only using residual magnetism) especially if the brushes etc aren’t 100%.
Someone no doubts knows but I guess the dynamos run well above engine rpm?