Here is a possible explanation:
When the CS02 switches to Program Track mode, there is actually a low-amplitude high-impedance leakage of programming track signal from the CS02 onto the control bus. The leakage is within NMRA S9.1.2 limits and a compliant booster should ignore this signal. However, when using the CS02 with some brands of booster (Digitrax and EasyDCC) a problem does occur and NCE has published a workaround resistor circuit between the CS02 and booster to solve the problem.
To quote from this document:
"Note: With Digitrax or EasyDCC installations you only need the resistors to the tirst booster. You can then ‘daisy chain‘ trom booster to booster. The resistors are only needed to ensure that the mainline shuts down when the programming track is in use. otherwise programming commands will be sent to the boosters programming all locos on the mainline. The resistors can be eliminated if you just want to do a quick hook up for testing purposes."
So the questions we need to ask are:
- Is the system using non-NCE boosters with the Power Pro command station?
- If so, have the appropriate precautions been taken?
- If using NCE boosters, did a fault occur in either the CS02 or the NCE booster to cause the known leakage to be a problem in this instance.
As far as the address change becoming evident after a short, this would be because most decoders require a power interruption for a new address to take effect.