[Coco] Re: Supercomm: more details

[John E. Malmberg]

KnudsenMJ at aol.com wrote:
> In a message dated 8/12/04 8:17:29 PM Eastern Daylight Time, james at skwirl.ca 
> writes:
> 
>>Scratch my earlier email, DCD (Coco end) is wired to CTS and RTS (other 
>> end) on my cable. Is that right? I think CTS (other) should be tied to 
>> RTS (coco)..

If you are talking about the bit-banger port, the line labeled DCD is 
treated sometimes like a CTS line by the software, and sometimes like a 
DCD/DSR line.

Otherwise, CTS to RTS, DTR to DSR, and DCD connected to DTR on the same 
side for NULL modem.

> 
> Nope -- it's already correct.  RTS should be wired to the same end's CTS (so 
> each end always gives itself permission to send),

Yes.

> and to the other end's DCD so the other end knows there's data about to come in.

No, that is a bug, and violates CCIT requirements.  DCD on a null modem 
is connected to DTR on the same connector.

DCD is a link level signal, not a flow control signal.  If not present, 
it follows DSR/DTR not CTS.

Some serial interfaces on some host computers implement the handling of 
these signals in hardware or in the serial device drivers, and can not 
support such a miswired cable that ties DCD to a CTS signal if the CTS 
signal is actually being used for handshaking.

A host computer is supposed to treat a loss of DCD or DSR as the 
termination of the serial session.

So is a DCE device such as a modem or a communication multiplexer.

Please do not advocate using a wiring configuration that violates CCIT 
signaling requirements unless it is required to get around a hardware or 
software bug that can not be fixed otherwise.

-John
wb8tyw(a)qsl.net
Personal Opinion Only