Ion> On Tue, 12 Nov 1996, Nathan Bryant wrote:
>> I don't know about the protocol, but the cable is a standard
>> "LapLink" cable. Check the PLIP documentation for info on how to
>> build one.
Ion> I've seen several replies to the original message, and they are all
Ion> rather clueless. I wonder why... :) A cable for an 8 bit parallel
Ion> connection (bidirectional/ECP/EPP/etc) is NOT a regular laplink
Ion> cable, although the latter does indeed work with 8 bit
Ion> bidirectional ports.
Well, the ECP protocoll is described in some M*cr*s*ft memorandum. I didn't
find any hint on M*rc*s*ft's Server ...
The Cable technology is licenced by M*rc*s*ft from Parallel Technology.
Guess why you don't find open documentation?
Ion> Anyway, although I don't have one handy, the idea is to connect
Ion> everything one-to-one, that is data pins (2-9 if I remember
Ion> correctly) to data pins and control pins (10-13 and 15 - wild
Ion> guess) to control pins. Plus ground, of course. :) I think the
Ion> IBM-PC hardware FAQ has a more detailed description of a
Ion> bidirectional parallel cable - you can find an older version
Ion> somewhere on rtfm.mit.edu and it will point you to the site of the
Ion> newer ones.
That won't work, as the controll line are one way. Either input or
output. So you have to connect an output on one end with an input at the
other end. I think, I could figure out some way how to do that in a way it
will work, but being compatible to something existing would be fine.
>> On Tue, 12 Nov 1996, Uwe Bonnes wrote:
>>
>> > Anyway, I am looking for information on the cable needed to connect
>> and the > protocol spoken on that cable with Microsofts Direct Cable
>> Connection.
Ion> See above. As for the protocol... dunno, although it can't be
Ion> something very complex (as it has to be fast). You can try to use
Ion> dosemu to see at least what the Win95 sends when it tries to
Ion> establish a connection (it doesn't sound very encouraging, does
Ion> it?...).
>> > The information to build a ECP DCC Cable could be achieved too if >
>> someone with such a cable would measure with a Ohm-meter which Pins
>> are > connected.
Ion> Right, but it only solves half of the problem; I believe you have
Ion> to initialize the port in a special way to make it talk ECP/EPP. I
Ion> had the patch for an "enhanced" (read: ECP/EPP aware) lp module
Ion> somewhere, if you want it just tell me and I'll look for it.
I have the datasheet for the SMC66[56], so I have information on how to
initialize it. And ECP should work with other chips that way too, as
M*cr*s*ft proposed a standard register layout.
Do you have some module that talks ECP to a printer? I only know of an
enhanced lp module that uses the FIFO found in most modern parallel
ports. It was hacked by me.
Find it on ftp.mathematik.th-darmstadt.de:/pub/linux/bonnes/fifo-0.92.tar.gz
I recently tried to remove some glitch. Please report failure and success.
Uwe Bonnes bon@elektron.ikp.physik.th-darmstadt.de
Institut fuer Kernphysik Schlossgartenstrasse 9 64289 Darmstadt
--------- Tel. 06151 162516 -------- Fax. 06151 164321 ----------