Wanted: informations about RS-422
Laurent Marguet
Has anyone experience of RS-422 ?
Can anybody send me informations about ... ?
a) PIN-OUT diagram for standard 25-pin connector (IBM compatible).
b) Pin names.
c) Is it possible to connect RS-232 with Rs-422 without adapter.
if it's possible, then connections diagram.
Thanks
Laurent Marguet
Henry Spencer
> a) PIN-OUT diagram for standard 25-pin connector (IBM compatible).
There is no standard 25-pin RS422 connector. The standard RS422 connector,
which almost nobody pays any attention to, is 37 pins (ugh!) with an
optional 9-pin auxiliary connector (double ugh!). You can't just use the
RS232 connector for RS422, because 422 needs two pins per signal.
> c) Is it possible to connect RS-232 with Rs-422 without adapter.
No. The signalling conventions are thoroughly incompatible, apart from
the fact that 422 is typically used for signalling at speeds too high for
standard-conforming 232 hardware. A few 422 systems can sort of talk to
232, if the cables are short and other circumstances are right, but this
is by no means guaranteed.
--
Programming graphics in X is like | Henry Spencer @ U of Toronto Zoology
finding sqrt(pi) using Roman numerals. | he...@zoo.toronto.edu utzoo!henry
Jim Cathey
>There is no standard 25-pin RS422 connector. The standard RS422 connector,
>which almost nobody pays any attention to, is 37 pins (ugh!) with an
>optional 9-pin auxiliary connector (double ugh!). You can't just use the
>RS232 connector for RS422, because 422 needs two pins per signal.
Actually, these are two different specs. RS-422 does not specify anything
other than the electrical characteristics of a single digital signal
travelling on a wire pair. No signaling convention or connector is
specified. The RS-232 equivalent using RS-422 signaling is a
different number, which I forget (as has, apparently, the rest
of the world! :-)
RS-485 is the successor to RS-422, and should be used in preference to it.
It has more robust drive capablilities, and is officially capable of
being bussed (though we got away with bussed RS-422 for years on our
keyboard lines, which were good to some 3000+ feet).
>> c) Is it possible to connect RS-232 with Rs-422 without adapter.
>
>No. The signalling conventions are thoroughly incompatible, apart from
>the fact that 422 is typically used for signalling at speeds too high for
>standard-conforming 232 hardware. A few 422 systems can sort of talk to
>232, if the cables are short and other circumstances are right, but this
>is by no means guaranteed.
It is possible, under good conditions. All Apple Macintoshes, for
example, use RS-422 for their serial ports. These apparently have no
problems talking to most RS-232 devices, so long as the cables are
short. You just ignore the inverted transmit line, and ground the
unused receiver line. Works best if you are using +/- 5V supplies
on your RS-422 end. Of course, you're giving up all the advantages
of using RS-422 in the first place, and are essentially reverting
to RS-423 (single-ended RS-422) by doing this.
+----------------+
! II CCCCCC ! Jim Cathey
! II SSSSCC ! ISC-Bunker Ramo
! II CC ! TAF-C8; Spokane, WA 99220
! IISSSS CC ! UUCP: uunet!isc-br!jimc (ji...@isc-br.isc-br.com)
! II CCCCCC ! (509) 927-5757
+----------------+
"PC's --- the junk bonds of the computer industry"
John Whitmore
[replying to the question]
>>> c) Is it possible to connect RS-232 with Rs-422 without adapter.
>>
>>No. The signalling conventions are thoroughly incompatible, apart from
>>the fact that 422 is typically used for signalling at speeds too high for
>>standard-conforming 232 hardware. A few 422 systems can sort of talk to
>>232, if the cables are short and other circumstances are right, but this
>>is by no means guaranteed.
>
>It is possible, under good conditions. All Apple Macintoshes, for
>example, use RS-422 for their serial ports. These apparently have no
>problems talking to most RS-232 devices, so long as the cables are
>short. You just ignore the inverted transmit line, and ground the
>unused receiver line. Works best if you are using +/- 5V supplies
>on your RS-422 end.
Macintosh ports are RS-423, not RS-422 (RS-423 has signalling
levels that go above and below GND, and gives you your choice of
RS-232 or differential receivers, though the RS-423 receiver will
be differential). The drivers and receivers (Am26LS30 and Am26LS32)
can drive serial signals a kilometer of so with no difficulty,
but the RS-232 end might have some difficulties.
>Of course, you're giving up all the advantages
>of using RS-422 in the first place, and are essentially reverting
>to RS-423 (single-ended RS-422) by doing this.
I think this is not right; the single-ended version of RS-422
is RS-232. RS-422 is a version of differentially driven serial that
allows 0-5V signal range (and which will not require any inconvenient
+/- 12V supplies like most RS-232, or +/- 5V supplies like the
Mac implementation of RS-423).
John Whitmore
Henry Spencer
> Macintosh ports are RS-423, not RS-422 (RS-423 has signalling
>levels that go above and below GND, and gives you your choice of
>RS-232 or differential receivers, though the RS-423 receiver will
I fear this is getting pretty confused. Back to basics for a moment:
RS232 is, roughly speaking, +-12V (the actual range of permissible
voltages is pretty broad), one wire per signal.
RS423 is sexed-up RS232. :-) It is +-5V, still one wire per signal
(one ground line per signal is recommended but nobody does it),
with somewhat different characteristics. It is compatible
with RS232 given favorable conditions.
RS422 is +5 only, two wires per signal, differential (that is, one wire
is at circa 0 and the other at circa +5).
The Mac ports are 422. Not 423. See the Apple documentation. You
can sort of make them talk to 423 and 232 by grounding one side of the
422 inputs, keeping speeds modest and cables short, and relying on
some pretty tolerant receiver/transmitter chips. (422's differential
receivers are not designed for negative voltages but tend to be able
to tolerate them, while 232/423 receivers are not designed for voltages
that go down only to 0 but will usually cope properly.) All bets are
off, noise immunity is poor, and it's not something to rely on if
you're getting paid for the job, but it often does sort of work.
>The drivers and receivers (Am26LS30 and Am26LS32)
>can drive serial signals a kilometer of so with no difficulty,
This is true only for 422. Note that those chips talk either 422 or
423 depending on how they are connected; on the Mac they're wired for 422.
>...(and which will not require any inconvenient
>+/- 12V supplies like most RS-232, or +/- 5V supplies like [423]
Actually, this objection to 232 is pretty much dead now, because modern
232 transmitter chips generate the +-12 themselves from +5. 422's big
advantage is not power supply but much higher speed, which is why Apple
used it.
Wayne Diener
>
> [replying to the question]
>>>> c) Is it possible to connect RS-232 with Rs-422 without adapter.
>>>
(stuff deleted)
>
> John Whitmore
Just to clear up these conventions a little, here they are:
(Extracted from the Texas Instruments Interface Circuits Data Book)
Specification RS-232C RS-423A RS-422A
Mode of operation single-ended single-ended differential
Max cable length 50 ft 4000 feet 4000 feet
Max data rate 20 kb/s 100 kb/s 10Mb/s
Max voltage applied
to driver output +/- 25V +/- 6V -0.25V to 6V
Driver output signal
Loaded +/- 5V +/- 3.6V +/- 2V
Unloaded +/- 15V +/- 6V +/- 5V
Receiver input
voltage range +/- 15V +/- 12V +/- 7V
Receiver input
sensitivity +/- 3V +/- 200 mV +/- 200 mV
Don't get confused by the max data rate. i.e. the 10 Mb/s for
RS-422A is defined at a cable length of 40 feet. At 4000 feet
the max data rate drops to 100 Kb/s. Similar apply to the other
two also.
--
|---------------------------------------------------------------|
| // Wayne D. Diener |
| // Spokane, WA |
| \\ // E-mail reply to: |
| \X/ To: isc-br!hawk!wddami!way...@uunet.uu.net |
|---------------------------------------------------------------|
John Whitmore
>RS232 is, roughly speaking, +-12V (the actual range of permissible
> voltages is pretty broad), one wire per signal.
>RS423 is sexed-up RS232. :-) It is +-5V, still one wire per signal
> with somewhat different characteristics. It is compatible
> with RS232 given favorable conditions.
It's allowed to be +- (4 to 6) Volts, actually, and the
compatibility with RS232 is guaranteed for wiring resistance
less than 900 Ohms (trust me, I just looked up the specs).
>RS422 is +5 only, two wires per signal, differential (that is, one wire
> is at circa 0 and the other at circa +5).
So far, so good.
>The Mac ports are 422. Not 423. See the Apple documentation.
I have seen it. It's 423. The receivers ARE RS-422 compatible,
but if you connect the output of the Mac serial port to a compliant
RS-422 receiver, it could burn up (RS-422 minimum input voltage spec
is -0.25V, and the Mac DOES deliver -5V to the serial output).
The Mac serial ports are a differential version of RS-423
and are RS-423 compliant, with an extra (noninverted) serial output
wire. Once you ground RXD+, the RS-423 spec is completely
satisfied by the Mac serial ports; driving RXD+ instead of grounding
it gives extra capabilities (like AppleTalk uses), so it makes
a lot of sense to leave RXD+ available on the output connector.
>>The drivers and receivers (Am26LS30 and Am26LS32)
>>can drive serial signals a kilometer of so with no difficulty,
>
>This is true only for 422.
Let me expand on that; the drivers and receivers are
capable of anything RS-422 can do, BUT you have to have the
right circuitry at the other end to get this performance (like,
another Macintosh). If the gizmo on the other end of the wire
is a minimum RS-423 instead of a Macintosh implementation, long
serial connections could fail.
The biggest advance of RS-423 over RS-232 is in the
ability of an output to drive multiple receivers (minimum of 10).
For RS-232-D (Nov '86) that minimum is 1. The Macintosh
ports are an advance over RS-423 (because they employ differential
drivers/receivers), but are RS-423 compliant in all respects.
John Whitmore
Henry Spencer
>>RS423 is sexed-up RS232. :-) It is +-5V, still one wire per signal
>> with somewhat different characteristics. It is compatible
>> with RS232 given favorable conditions.
>
> It's allowed to be +- (4 to 6) Volts, actually, and the
>compatibility with RS232 is guaranteed for wiring resistance
>less than 900 Ohms (trust me, I just looked up the specs).
I trust you on this :-), but in practice the specs don't tell the whole
story. Much the biggest concern for 423-232 compatibility is whether
your 232 hardware really meets 232. There is commercially-sold hardware
(in the PC market, of course -- where else?) whose "RS232" interface
generates signals bearing only a passing resemblance to the 232 specs.
In fact, very little "RS232" hardware really meets the *entire* spec
(including fine points like slew-rate limiting), but the violations
I speak of are pretty gross, and can definitely prevent interoperation
with 423. People have run into this.
>>The Mac ports are 422. Not 423. See the Apple documentation.
> I have seen it. It's 423...
Which Apple docs are you speaking of? "Guide to the Macintosh Family
Hardware", 2nd edition, chapter 10, "Serial I/O Ports", says "422".
(I looked earlier this evening.) Not being fond of overpriced computers
with undersized screens, I have not checked this out personally...
John Whitmore
>In article <1991Sep27....@milton.u.washington.edu> wh...@milton.u.washington.edu (John Whitmore) writes:
>>>The Mac ports are 422. Not 423. See the Apple documentation.
>> I have seen it. It's 423...
>Which Apple docs are you speaking of? "Guide to the Macintosh Family
>Hardware", 2nd edition, chapter 10, "Serial I/O Ports", says "422".
To my surprise, it does (and it's wrong, according to the
other info and the RS-422 spec). The first edition added to that
the statement that omitting the TXD+ connection and grounding RXD+
made the port RS-423, and THAT statement wasn't kept into the next
edition.
The outputs are rated (_Guide to the Macintosh Family
Hardware_, 2nd edition) to provide 7.2V differential into a load,
and the RS-422 spec states that the differential output into
a load must be between 4V and 6V. The outputs also go below -0.25V
(which is an RS-422 unspecified-behavior region). For
these reasons, I don't really like to think of it as a RS-422
implementation.
I DO rather like the implementation, however; with a resistor
and a zener, the TXD+ can be connected to a TTL serial input,
and I ran a Z8/BASIC micro directly off the serial port of my
Mac 128. A couple of resistors to bias RXD- to 1.5V made the
TTL serial output compatible with RXD+. The serial port provided
+5V power and the entire attached computer was a single chip
and clock crystal, three resistors, one diode, all on a protoboard.
Try THAT with EIA-232!
John Whitmore
j chapman flack
> The biggest advance of RS-423 over RS-232 is in the
>ability of an output to drive multiple receivers (minimum of 10).
>For RS-232-D (Nov '86) that minimum is 1. The Macintosh
>ports are an advance over RS-423 (because they employ differential
>drivers/receivers), but are RS-423 compliant in all respects.
Well, I know I'm being picky now, but it's been several years since the EIA
stopped naming their standards RS-*. They're all EIA-* now, and the change
must have been at least 5 years ago, because the published D spec is titled
EIA-232-D, not RS-232-D. Btw, I'm told EIA-232-E is out now.
It's a minor point, but I think we do look a bit silly as a profession when
we're five years behind on knowing the names of our own more important
standards. If we all make a point of using the proper names in conversation,
we should be able to speed up that process.
--
Chap Flack Their tanks will rust. Our songs will last.
ch...@art-sy.detroit.mi.us -MIKHS 0EODWPAKHS
Nothing I say represents Appropriate Roles for Technology unless I say it does.
Henry Spencer
>Well, I know I'm being picky now, but it's been several years since the EIA
>
>It's a minor point, but I think we do look a bit silly as a profession when
>we're five years behind on knowing the names of our own more important
I think we look just as silly when we keep changing the names of important
standards for no good reason.