taylor (05/17/82)
The bi-directional modem is a really neat idea. However, after spending a bit of time reproducing the efforts already reported, I decided to drop back to the beginning and try to produce specs for how a modem 'should' behave to allow for this capability. (Obviously, the ven-tel does not.) My conclusion is that this is not possible - i. e., the eia leads do not provide sufficient info. The rub comes in that when in answer mode, the d'con expects carrier to act in the 'normal' mode - as a control signal. In originate mode, tho, carrier must be up before the number can be passed, so obviously it must be up before a real carrier is established. I was able to discover a couple of points which might be of help in finding a solution, especially if some of you with more ee experience than i have can help. A key here is the fact that if the ven-tel option for ring indicator is set for the 'stay on' option, ri acts very much like carrier detect, and drops when the carrier drops, if and only if the modem answered. If the modem originated the call, ri stays low. proposed circuit: DTE SLIM Ven-Tel 7------------------------------------------7 2------------------------------------------2 3------------------------------------------3 8-------- | | \ <--------------------------------22 | | 20------- Where the /<-- construct is used to represent a 'logocal relay' such that if pin 22 is negative (off), 8 is connected to 20, but if 22 is positive, 22 is connected to 8. An additional circuit element would probably have to added as a delay timer if pin 22 goes from positive to negative. This would ensure that if 22 drops (carrier loss), then 8 drops (and conse- quently 20 drops at the switch) before 8 stays high due to 20 being high. (I see the possibility for a loop here.) Do some of you with more ee experience than I know how this element would work? My gut feeling is that not much more than a transistor (and possibly a SLOW one) is necessary. Steve Taylor NC Educational Computing Service (919) 549-0671