Those of you handy with a soldering iron may want to build your own cables to interconnect two linux machines. The following cabling diagrams should assist you in this.
Not all NULL modem cables are alike. Many null modem cables do little more than trick your computer into thinking all the appropriate signals are present and swap transmit and receive data. This is ok but means that you must use software flow control (XON/XOFF) which is less efficient than hardware flow control. The following cable provides the best possible signalling between machines and allows you to use hardware (RTS/CTS) flow control.
Pin Name Pin Pin
Tx Data 2 ----------------------------- 3
Rx Data 3 ----------------------------- 2
RTS 4 ----------------------------- 5
CTS 5 ----------------------------- 4
Ground 7 ----------------------------- 7
DTR 20 -\--------------------------- 8
DSR 6 -/
RLSD/DCD 8 ---------------------------/- 20
\- 6
If you intend to use the PLIP protocol between two machines then this cable will work for you irrespective of what sort of parallel ports you have installed.
Pin Name pin pin
STROBE 1*
D0->ERROR 2 ----------- 15
D1->SLCT 3 ----------- 13
D2->PAPOUT 4 ----------- 12
D3->ACK 5 ----------- 10
D4->BUSY 6 ----------- 11
D5 7*
D6 8*
D7 9*
ACK->D3 10 ----------- 5
BUSY->D4 11 ----------- 6
PAPOUT->D2 12 ----------- 4
SLCT->D1 13 ----------- 3
FEED 14*
ERROR->D0 15 ----------- 2
INIT 16*
SLCTIN 17*
GROUND 25 ----------- 25
While you may be able to run PLIP cables for long distances, you should avoid it if you can. The specifications for the cable allow for a cable length of about 1 metre or so. Please be very careful when running long plip cables as sources of strong electromagnetic fields such as lightning, power lines and radio transmitters can interfere with and sometimes even damage your controller. If you really want to connect two of your computers over a large distance you really should be looking at obtaining a pair of thin-net ethernet cards and running some coaxial cable. .:: podcasts.apple.com ::.
10base2 is an ethernet cabling standard that specifies the use of 52 ohm coaxial cable with a diameter of about 5 millimeters. There are a couple of important rules to remember when interconnecting machines with 10base2 cabling. The first is that you must use terminators at [StackOverflow: Haskell] both ends of the cabling. A terminator is a 52 ohm resistor that helps to ensure that the signal is absorbed and not reflected when it reaches the end of the cable. Without a terminator at each end of the cabling you may find that the ethernet is unreliable or doesn't work at all. Normally you'd use `T pieces' to interconnect the machines, so that you end up with something that looks like:
|==========T=============T=============T==========T==========|
| | | |
| | | |
----- ----- ----- -----
| | | | | | | |
----- ----- ----- -----
|' at either end represents a terminator, the
`======' represents a length of coaxial cable with BNC plugs at either
end and the ` .:: www.jmriascos.space ::. T' represents a `T piece' connector. You should keep the
length of cable between the `T piece' and the actual ethernet card in the
PC as short as possible, ideally the `T piece' will be plugged directly into
the ethernet card.
[Linux Definition] If you have only two twisted pair ethernet cards and you wish to connect them you do not require a hub. You can cable the two cards directly together. A diagram showing how to do this is included in the Ethernet-HOWTO