GP9s and EMDs of the era use a load regulator to control excitation. The diesel engine and main generator have different power curves, so the load regulator balances the electrical load with the diesel engine output, so neither is overloading or underloading. The load regulator is a large commutator connected to a series of resistors. A vane motor driven by the governor oil moves a brush to increase or decrease the output and resulting excitation signal.

In the normal or "road" position, the Overriding Solenoid (ORS) is energized in the governor which causes the load regulator to assume the minimum field position. Once the throttle is opened, the ORS is de-energized allowing the load regulator to gradually increase excitation until it balances.

Switch engines in the switching mode did not energize the ORS so when the throttle opened they were already in the maximum field position and then balanced from there. This was later changed to a relay that provided an initial amount of excitation until the load regulator caught up from minimum field.

On GP7 and GP9 units, in switching mode the various throttle positions would energize a series of relays which would provide the basic excitation in parallel with the load regulator. In higher throttles, the system would switch over to the load regulator.

Not sure what SP did on the commute GP9s, but they might have used a variation of one or more of the switching systems.

I assume the F40PH units you are referring to have the HEP alternator driven by the prime mover like Amtrak used. (EMD and M-K built some with separate engine driven HEP plants for some commuter agencies.) On the Amtrak version, the 16-645 at 900 rpm in Run 8 drives a 2:1 gearbox to turn a Delco HEP alternator at 1800 rpm which results in 60 hertz AC. On later EMD units, each throttle position creates a basic voltage for the excitation through a throttle response system. This basic signal is further modified to keep the amperage and voltage of the AR10 output within limits, stay within adhesion limits, and reduced by the wheel slip system. There is still a load regulator, now just a simple rheostat, that further modifies the signal. The rate of change in the excitation signal is controlled by capacitors and discharge resistors. This signal is used to control firing SCRs off the 3 phase AC system which actually provides the DC for excitation.

An EMD with throttle response and static excitation systems can be easily wired to provide an excitation signal lower than the actual diesel engine speed. An SW1000/SW1500/MP15 series use a similar system in their "kicking" circuit where the locomotive is in Run 3 but the but the actual excitation level is determined by the throttle position in idle, Run 1, 2 or 3[ so there isn't any delay waiting for the diesel engine speed to increase to the required level. I believe this option was also available on other EMD models. GE did something similar with the U33C with just a few (3?) diesel engine speeds.

Back to traction motor connections - I think the ATSF/BNSF hump SD slug sets are wired as the mother is 2S3P (2 traction motors in series, then three strings is parallel) which is then connected in series to the slug with the same arrangement, making the equivalent of 3 strings of 4 motors in series. Slugs bring in other issues, so as speed in creases the slug traction motors have to be either disconnected, or if in series with active motors some short across to bypass them.