I've had another look at this issue. I'm no longer experiencing the instability problem I described earlier. I think this might have been due to the clock module I was using to generate my system CLK and CLK2 signals, but I'm not sure.
I now have SC725 configured with JP4 set to the main bus clock CLK, JP5 set to C (which JP4 selects as CLK), JP6 set to CTC. In this configuration port A is fixed at 115200 baud from the main system clock CLK, while port B is controlled from CTC channel 1. SCM defaults to setting the CTC channel 1 and SIO port B for 115200.
SCM does not control port A with the CTC as the original SIO+CTC module (SC110) was hardwired to use the CLK signal for SIO port A for compatibility with standard RC2014 firmware that does not know anything about CTCs. Similarly, the standard configurations of RomWBW does not set up the CTC.
When a second SIO module is fitted there is a slight problem. SCM mistakenly assumes that if it detects a CTC at address 0x80 all serial ports, other than the first one, use a CTC. It therefore programs the SIO on the assumption the serial clock has a CTC as a pre-scaler. This results in a baud rate for both serial ports on the second SIO module (which does not have a CTC) being 460800.
Other than that all four serial ports are working with both SCM and RomWBW. Note with RomWBW one of the jumpers needs to be changed.
I've improved the documentation on my website for SC725.
My test systems is:
SC706 - Z80 CPU and 7.3728 MHz clock
SC608 - Memory module with both SCM and RomWBW plus 512kB RAM
SC725 - SIO+CTC module (address 0x80 and 0x88)
SC716 - SIO module (address 0x84)
SC129 - Digital I/O for diagnostic LEDs (address 0x00)
Steve