actually those parameters have very different purposes.
The first parameter, maxTargetStepSize, should guarantee that 2 hits are not separated by more than 3mm, however, I believe, this is only valid for electrons. Do you get a step larger than 3mm, for electrons?
The second parameter, subEventTimeDelay is a parameter that allows to split a single Geant4 generated event into independent subEvents that are finally treated as an event itself.
For example, if I simulate an event decay from a full chain, I will simulate all the decays in that chain. However, between event and event there will be a period of time, that might be even years.
If the time between two physical interactions in Geant4 is more/higher than subEventTimeDelay, then the hits/deposits will be considered to be in independent subevents.
A reasonable value to introduce as subEventTimeDelay is the total drift time of the TPC.
This will split, for example, a full chain decay simulation, into independent subEvents. It might happen then that you simulate 1000 events, and end up with a larger number of events in the final file because few events were split into subEvents.