Notice most signals are passed directly between the modem and the user port, but two signals are transistor-shunted to ground. First, I realize now this configuration could have blown my CIAs by feeding 12v directly to the chip. Secondly, I am not certain which signals are shunted to ground, but I believe they are DTR, DCD, and RTS. I cannot recall the details on the build, but I remember this being a minor modification to a circuit I found on a BBS. In any case, it worked perfectly with my Commodore 64 with several 1200 and 2400 bps modems, as well as my Commodore 128 then 128D with a 9600 bps modem.