If the set should happen to come with EPR-14869B on the (currently unemulated) I/O board, that needs a redump as well (by ordinary means). The bad bytes are few enough in number that they could be hand-patched by comparing them with other versions, but that's no substitute for a definitive dump.
Also of interest to me is the asynchronous serial protocol implemented by the 315-5338A custom I/O controller to transfer data between the I/O board and dual-port RAM on the main board. There is a serious lack of hardware documentation for both this device and its CRX/ST-V successor (315-5649).