RS-232 baud rate vs bps
Lockheed WsmrSmts
When I first told someone here that you could look at a serial port
RS-232 line and calculate the baud rate based on the bit period, it
sparked a "spirited" discussion. On fellow suggested that since I quoted
from software oriented books that use "baud" and "bits-per-second" as
synonyms that I was "ignorant".
Well Paul, here are some "hardware" references for you:
Motorola MVMV-162 Hardware Reference Guide page 1-10:
"The MVME162 hardware supports asynchronous serial baud rates of 110b/s to
38.4Kb/s.
(note the use of baud and b/s as synonyms)
Signetics SCN68562 DUSCC Product Specification:
Several examples, but I'll quote one "16 fixed rates 50 to 38.4k baud"
(I use this chip extensively, and the 1/baud test is true every time)
Zilog SCC Hardware Users Manual:
Refers to baud and bits-per-second as synonyms numerous places in the
description of the Z80X30 and Z85X30 SCC.
There. Since your so damn fixated on "hardware people" maybe you'll
believe those references.
Each and every time I mention looking at a serial port to verify this
you begin talking about the analog stage of a modem. STOP.... you may
notice that I did NOT mention a modem ONCE in my statements, quotes, or
examples, and in each and every one of these cases the bit period DOES
equal 1/BAUD. Now you may argue with the above manufacturers if you wish,
or accept that you actually can transmit data WITHOUT a modem. In any
case if you're still stuck on your strict narrow-minded definition of
baud, then I pity you.
______________________________________________
| Lockheed Martin Engineering and Sciences |
| Software Engineering Group |
| P.O. Box 189, Bldg 1554 |
| White Sands Missle Range, NM 88002 |
|____________________________________________|
Alexander J Russell
In article <4qbvv7$3...@liberator.concentric.net>, Les...@cris.com says...
still doesn't make it correct.
Baud means the rate of change in the signal. BPS means bits per second
transmited with that signal. They are OFTEN not identical with modern signaling
hardware.
This disctinction is often ignored, but is still the correct definition.
If you choose to use BAUD as BPS, it doesn't mean you are ignorant, just a bit
sloppy with terminology.
--
The AnArChIsT! Anarchy! Not Chaos!
aka
Alex Russell
ale...@iceonline.com
Thomas Krug
Lockheed WsmrSmts wrote:
>
> When I first told someone here that you could look at a serial port
> RS-232 line and calculate the baud rate based on the bit period, it
> sparked a "spirited" discussion. On fellow suggested that since I quoted
> from software oriented books that use "baud" and "bits-per-second" as
> synonyms that I was "ignorant".
>
> Well Paul, here are some "hardware" references for you:
Hm... is that discussion THAT hard??? :)
I think def. of baud is bits-per second (so 300 bauds are 300 bps).
So baud is a count of bits being sent in a certain time.
Another thing is bit-period. That should be the TIME you use to transmit
(or receive :) ) a bit.
So if you use 300 baud, then this will be 300 bits per second (bps) or
a bit-period of 1 bit sent per 1/300 second.
I never did think about that topic to be a religion-like ;)
Thomas Krug.
----------------------------------------------------------------------
tom...@cip1.e-technik.uni-erlangen.de
Thomas Krug
...so I am "just a bit sloppy with terminology", too. Sorry.
Thomas
----------------------------------------------------------------------
tom...@cip1.e-technik.uni-erlangen.de
Lockheed WsmrSmts
Alexander J Russell (ale...@icebox.iconline.com) wrote:
: Yes, its true that the MIS-USE of the term BAUD to mean BPS is common, but that
: Baud means the rate of change in the signal. BPS means bits per second
: transmited with that signal. They are OFTEN not identical with modern signaling
: hardware.
And if you read my post, you would know at the point I used this it IS
true, and is NOT a misuse of the terms. In both this thread and the
original, I was stating that at the RS-232 stage (EIA standard, bistate,
bipolar data) the width of a bit = 1/baud. This is true (accept it or
not). Then the discussion went off on a tangent about the analog encoding
of data by modems, and other statements not related to the original
thread.
: This disctinction is often ignored, but is still the correct definition.
: If you choose to use BAUD as BPS, it doesn't mean you are ignorant, just a bit
: sloppy with terminology.
(see above)
Paul Schlyter
In article <4qbvv7$3...@liberator.concentric.net>,
Lockheed WsmrSmts <Les...@cris.com> wrote:
> When I first told someone here that you could look at a serial port
> RS-232 line and calculate the baud rate based on the bit period, it
> sparked a "spirited" discussion. On fellow suggested that since I quoted
> from software oriented books that use "baud" and "bits-per-second" as
> synonyms that I was "ignorant".
>
> Well Paul, here are some "hardware" references for you:
>
> "The MVME162 hardware supports asynchronous serial baud rates of 110b/s to
> 38.4Kb/s.
become "bit/s/s". Talking about "baud rate" is like talking about
"speed rate" - which really is acceleration, not speed.
> (note the use of baud and b/s as synonyms)
>
> Signetics SCN68562 DUSCC Product Specification:
> Several examples, but I'll quote one "16 fixed rates 50 to 38.4k baud"
>
> (I use this chip extensively, and the 1/baud test is true every time)
>
> Zilog SCC Hardware Users Manual:
> Refers to baud and bits-per-second as synonyms numerous places in the
> description of the Z80X30 and Z85X30 SCC.
>
> There. Since your so damn fixated on "hardware people" maybe you'll
> believe those references.
don't they? Which means that IN THESE PARTICULAR SITUATIONS one
baud is indeed one bit/s.
> Each and every time I mention looking at a serial port to verify this
> you begin talking about the analog stage of a modem. STOP.... you may
> notice that I did NOT mention a modem ONCE in my statements, quotes, or
> examples, and in each and every one of these cases the bit period DOES
> equal 1/BAUD.
from "bit/s", and then you claim that "baud = bit/s" is universally
valid. THAT's your fallacy. If you want to claim that something's
universally valid, then you must investigale ALL cases, not just those
cases that happen to agree with your pet theory.
> STOP.... you may notice that I did NOT mention a modem ONCE in my
> statements, quotes, or
the situation in the signal over a phone line from a modem? Are
"baud" always equal to "bit/s" in all such situations too? Don't
escape this situation once more, instead answer the question!!
> Now you may argue with the above manufacturers if you wish,
> or accept that you actually can transmit data WITHOUT a modem. In any
> case if you're still stuck on your strict narrow-minded definition of
> baud, then I pity you.
beliving one baud always is always one bit/s. As "proof" you point
to situations where one baud indeed is one bit/s, while you're
carefully avoiding situations where one baud is not one bit/s, and
then you're claiming that "one baud = one bit/s" is universally
valid. It isn't.
--
----------------------------------------------------------------
Paul Schlyter, Swedish Amateur Astronomer's Society (SAAF)
Grev Turegatan 40, S-114 38 Stockholm, SWEDEN
e-mail: pau...@saaf.se p...@home.ausys.se
Paul Schlyter
In article <4qe9r7$1...@liberator.concentric.net>,
Lockheed WsmrSmts <Les...@cris.com> wrote:
> Alexander J Russell (ale...@icebox.iconline.com) wrote:
>: Yes, its true that the MIS-USE of the term BAUD to mean BPS is common,
>: but that still doesn't make it correct.
>: Baud means the rate of change in the signal. BPS means bits per second
>: transmited with that signal. They are OFTEN not identical with modern
>: signaling hardware.
>
> And if you read my post, you would know at the point I used this it IS
> true,
> and is NOT a misuse of the terms.
valid. It isn't - it's valid only in some special situations.
> In both this thread and the original, I was stating that at the
> RS-232 stage (EIA standard, bistate, bipolar data) the width of a
> bit = 1/baud.
And there you were wrong, as several people has pointed out to you
a number of times.
> This is true (accept it or not). Then the discussion went off on
> a tangent about the analog encoding of data by modems, and other
> statements not related to the original thread.
valid.
Lockheed WsmrSmts
Paul Schlyter (pau...@electra.saaf.se) wrote:
:
: Of course, since you selected a situation where 1 baud = 1 bit/s.
:
Yes, and that's why I made the statement.
:
: Your misuse of the terms is your claim that baud = bit/s is universally
: valid. It isn't - it's valid only in some special situations.
Show me in ANY of my posts where I said this is "universally
valid" ...!
: > In both this thread and the original, I was stating that at the
: > RS-232 stage (EIA standard, bistate, bipolar data) the width of a
: > bit = 1/baud.
:
: No you didn't -- you stated that the width of a bit always is 1/baud.
: And there you were wrong, as several people has pointed out to you
: a number of times.
(Show me where I said "always")
:
Wrong. Once again....the original thread was not about modems,
but about sending data over an EIA RS-232C handshake line called CTS.
THAT is where I said a bit period = 1/Baud.
--
Lockheed WsmrSmts
Paul Schlyter (pau...@electra.saaf.se) wrote:
:
: Of course, since you selected a situation where 1 baud = 1 bit/s.
:
where I made that claim. YOU brought up the analog modem discussion, and
I merely stated that it is widely accepted terminology, NOT a universal
truism.
Marc
Les...@cris.com (Lockheed WsmrSmts) wrote:
> When I first told someone here that you could look at a serial port
>RS-232 line and calculate the baud rate based on the bit period, it
>sparked a "spirited" discussion. On fellow suggested that since I quoted
>from software oriented books that use "baud" and "bits-per-second" as
>synonyms that I was "ignorant".
>
> Well Paul, here are some "hardware" references for you:
>
>Motorola MVMV-162 Hardware Reference Guide page 1-10:
>"The MVME162 hardware supports asynchronous serial baud rates of 110b/s to
>38.4Kb/s.
>
>
>
I was always under the impression that "baud rate" is number of state
changes, where as bits/sec is, well, "bits per second". These two values
can be vastly different when you take into account that a modem on a
phone line may have 8 states for a given bits/sec. The result is that
the actual "baud" rate is less than the "bits/sec" for a given throughput
of data. Confusing, huh?
If you want, e-mail me and I can point you toward some references.
Although these terms are often accepted as synonyms, it is not
technically correct when you look at actual data transmission.
Paul Schlyter
In article <4qh5js$1...@liberator.concentric.net>,
Lockheed WsmrSmts <Les...@cris.com> wrote:
> Paul Schlyter (pau...@electra.saaf.se) wrote:
>:
>: Of course, since you selected a situation where 1 baud = 1 bit/s.
>
> Finally! YOU used the two as synonyms.
they don't mean the same thing in just some special situations.
One example: for water at +4 deg C, 1 liter = 1 kg -- does this mean
that liter and kg are synonyms? Of course not! They just happen to
be the same in this particular situation - usually they are different.
> That is the ONLY situation where I made that claim. YOU brought up
> the analog modem discussion, and I merely stated that it is widely
> accepted terminology, NOT a universal truism.
to claim that bauds could be different from bit/s, and you also
claimed that the definition "baud = signal changes per second" was
"obsolete" and that the "modern" definition was "baud = bit/s".
Paul Schlyter
In article <4qh58b$1...@liberator.concentric.net>,
Lockheed WsmrSmts <Les...@cris.com> wrote:
> Paul Schlyter (pau...@electra.saaf.se) wrote:
>:
>: Of course, since you selected a situation where 1 baud = 1 bit/s.
>
>
>: Your misuse of the terms is your claim that baud = bit/s is universally
>: valid. It isn't - it's valid only in some special situations.
>
> Show me in ANY of my posts where I said this is "universally valid" ...!
"bit/s" are synonyms (which means the should always be the same, not just
in some particular cases). In addition you claimed that the original
definition of "baud" as "signal changes/s" was obsolete, and that the
"modern" definition was "baud = bit/s" -- this too only makes sense if
the relation "baud = bit/s" would be universally valid.
>: > In both this thread and the original, I was stating that at the
>: > RS-232 stage (EIA standard, bistate, bipolar data) the width of a
>: > bit = 1/baud.
>:
>: No you didn't -- you stated that the width of a bit always is 1/baud.
>: And there you were wrong, as several people has pointed out to you
>: a number of times.
>
> (Show me where I said "always")
>
> Wrong. Once again....the original thread was not about modems,
> but about sending data over an EIA RS-232C handshake line called CTS.
> THAT is where I said a bit period = 1/Baud.
"bit/s" -- they are synonyms only if the're _always_ the same. No,
you didn't use the word "always" but what you said would imply them
always being the same.
Now that you seem to have dropped the claim that "baud" and "bit/s"
are always the same, you admit that they're sometimes different,
right? Perhaps you could also drop your earlier claim that the
original definition of "baud" is "obsolete"? The original definition
of "baud" would work quite well in the case you describe above, if
you encode bits as high/low voltage transmitted in uniform time
intervals: then one baud would be one bit/s .......
..... at least if you send your data synchronously (i.e. data bits
all the time, no start/stop/parity bits). However if, as is
customary in asynchronous serial communication, you also send start
and stop bits, an 8-bit byte could contain between 10 and 12 bits on
the serial line:
10 bits/byte: 1 start bit, 8 data bits, 1 stop bit
11 bits/byte: 1 start bit, 8 data bits, 2 stop bits
11 bits/byte: 1 start bit, 8 data bits, 1 parity bit, 1 stop bit
12 bits/byte: 1 start bit, 8 data bits, 1 parity bit, 2 stop bits
Thus if you send signals on your CTS at a rate of 1200 baud, and you
add 1 start bit, 1 parity bit and 2 stop bits to each byte, then
you'll only acheive a throughput of 800 bps on that line, since 400
bps will be overhead with no information content. If you send 1
start and 1 stop bit, and no parity, you'll acheive a throughput of
960 bps using 1200 baud.
There are still other possibilities: each byte in the asynchronous
communications line could contain not 8 but 5, 6 or 7 data bits (I
hope this sentence won't open a new flame war here on the subject
"Can the number of bits in a byte be different from 8?"). Or you
could use another encoding scheme, for instance morse code (which
isn't binary but rather "trinary" since any one moment could contain
a dot, a dash, or a pause -- the pauses contain information too).
This is a quite valid application -- the CTS line of the port could
for instance be connected to control a ham rig. In all these cases 1
baud would be different from 1 bit/s. The definition of "baud" in
Morse Code has some problems of its own due to the fact that
different signalling elements have different lengths - the standard
way to define it is to use the word "PARIS", including its final
pause).
Tim Osler
> Marc <ma...@mail.snet.net> wrote in article
<4qgrc2$1g...@CT1.SNET.Net>...
> Les...@cris.com (Lockheed WsmrSmts) wrote:
> > When I first told someone here that you could look at a serial port
> >RS-232 line and calculate the baud rate based on the bit period, it
> >sparked a "spirited" discussion. On fellow suggested that since I
quoted
> >from software oriented books that use "baud" and "bits-per-second" as
> >synonyms that I was "ignorant".
> >
> > Well Paul, here are some "hardware" references for you:
> >
> >Motorola MVMV-162 Hardware Reference Guide page 1-10:
> >"The MVME162 hardware supports asynchronous serial baud rates of 110b/s
to
> >38.4Kb/s.
> >
> >(note the use of baud and b/s as synonyms)
This is referring to the actual baud rates of the modem. The hardware is
truly capable of 38.4Kbaud (state changes). This same modem is also (via
v34/v42bis) capable of transferring data at 115Kb/s which is not its
actual baud rate. (A 28.8 modem will transfer at 115kb/s via v35 -- don't
know if a 38.4 will do 153k)
>
> I was always under the impression that "baud rate" is number of state
> changes, where as bits/sec is, well, "bits per second". These two values
> can be vastly different when you take into account that a modem on a
> phone line may have 8 states for a given bits/sec. The result is that
> the actual "baud" rate is less than the "bits/sec" for a given
throughput
> of data. Confusing, huh?
>
Add wood to the fire:
It may be worth noting that when the RS232 standard was written, baud and
bps were interchangable. (equal if not equivalent) ie. a 2400 baud
modem/port transferred information at 2400 bps. With the wonders of data
compression, a 2400 baud could transmit info at 9600 bps. 'Course when
you looked at the modem box, it indicated a 9600 baud modem. A 2400 baud
modem used the high and low signals whereas a '9600' used the high and low
signals plus the high->low and low->high state changes. Thankfully
manufacturers(vendors?) have gotten away from this practice and advertise
the actual baud capabilities of the device. (Due to IEEE standards
compliance)
Process this useless information carefully >:-)
Tim -no sig
Paul Schlyter
In article <01bb6251.83fba240$7c8280d0@hairball>,
Tim Osler <hair...@avicom.net> wrote:
> It may be worth noting that when the RS232 standard was written, baud and
> bps were interchangable. (equal if not equivalent) ie. a 2400 baud
> modem/port transferred information at 2400 bps.
bauds, as did the 1200 bps modems.
The only modems where baud = bps were the old slow 110 bps and 300 bps
modems. 600 bps modems too could use 1 bit per baud, however 600 bps
never became a common type of modem.
The 75/1200 bps modems also used 1 bit per baud: the 1200 bps forward
channel used almost all of the bandwidth on the phone line, and the
75 bps backchannel used a small portion of the bandwidth. The idea
was that a terminal user typically typed much more slowly than he
received information on the screen, thus data from the terminal could
be sent at 75 bps while the terminal received data at 1200 bps. This
was the fastest modem available using simple FM of the carrier.
Many 75/1200 bps modems also required a 75/1200 "split speed" UART.
PC's had then begun to appear, however their 8250 UART had no
hardware split-speed capability. Of course there were software
solutions to that problem: one way was to, for each byte that was
about to be sent at 75 bps, send 16 equivalent bytes at 1200 bps,
approximating the 75 bps character (thus if you wanted to send say 5A
at 75 bps, instead you sent something like 00 00 FF FF 00 00 FF FF FF
FF 00 00 FF FF 00 00 at 1200 bps). Of course the start/stop bits
could not be controlled, but by applying a suitable capacitor much of
them could be filtered out... A better method was to pulse the BREAK
signal on and off, to make it appear as if bytes were sent at 75 bps.
The 1200/1200 bps modems that appeared later used more sophisticated
modulation methods, but had to revert to 600 bauds because only half
of the bandwidth of the phone line could be used for each direction.
Only with the more modern 9600+ bps modems, with digital signal
processors and echo cancellation circuits, could each direction use
the full bandwidth of the phone line - simultaneously.
John Bejjani
> From: Thomas Krug <tom...@cip1.e-technik.uni-erlangen.de>
> Newsgroups: comp.os.msdos.programmer
> Date: Fri, 21 Jun 1996 08:02:23 +0200
> Organization: Regionales Rechenzentrum Erlangen, Germany
>
> Lockheed WsmrSmts wrote:
> >
> > When I first told someone here that you could look at a serial port
> > RS-232 line and calculate the baud rate based on the bit period, it
> > sparked a "spirited" discussion. On fellow suggested that since I quoted
> > from software oriented books that use "baud" and "bits-per-second" as
> > synonyms that I was "ignorant".
> >
> > Well Paul, here are some "hardware" references for you:
> > < (cut)
>
> Hm... is that discussion THAT hard??? :)
>
> I think def. of baud is bits-per second (so 300 bauds are 300 bps).
> So baud is a count of bits being sent in a certain time.
> Another thing is bit-period. That should be the TIME you use to transmit
> (or receive :) ) a bit.
>
> So if you use 300 baud, then this will be 300 bits per second (bps) or
> a bit-period of 1 bit sent per 1/300 second.
>
> I never did think about that topic to be a religion-like ;)
>
> Thomas Krug.
>
> ----------------------------------------------------------------------
> tom...@cip1.e-technik.uni-erlangen.de
Woa there Nelly, there's a BIG difference between bps and baud.
Baud is the number of encoded "pieces" that are transferred at one time.
For example, for an analog transmission, if you have 4 different frequencies
used to encode the patterns 00, 01, 10, 11, then the baud rate is the number
of bps * 4.
For digital transmission, you can only encode 0, and 1, so in this case the
baud rate IS equal to the bps (bps * 1).
Just thought I'd clarify this...
--
+--------------------+---------------------------------+
| John Bejjani | Email: jbej...@newbrdige.com |
| Software Engineer | Phone: (613) 599-3600 Ext. 6562 |
| Newbridge Networks | Fax: (613) 591-3680 |
| Corporation | |
+--------------------+---------------------------------+
Paul Schlyter
John Bejjani <jbej...@jafar.newbridge> wrote:
> Woa there Nelly, there's a BIG difference between bps and baud.
> Baud is the number of encoded "pieces" that are transferred at one time.
> For example, for an analog transmission, if you have 4 different frequencies
> used to encode the patterns 00, 01, 10, 11, then the baud rate is the number
> of bps * 4.
First if one baud encodes two bits, the baud will be SMALLER than the
bit rate -- in this case baud = bps/2
Second, the factor is 2, not 4. Yes, there are 4 combinations, but
only two bits are encoded in each baud.
> For digital transmission, you can only encode 0, and 1, so in this case the
> baud rate IS equal to the bps (bps * 1).
should have arrived at baud = bps * 2 (since 0 and 1 are 2 possible
combinations), but that would of course be wrong...
Ben North
In article <4qp2ir$b...@electra.saaf.se>,
Paul Schlyter <pau...@electra.saaf.se> wrote:
>Tim Osler <hair...@avicom.net> wrote:
>
>> [... something ...]
>
>[...]
>
>The only modems where baud = bps were the old slow 110 bps and 300 bps
>
>The 75/1200 bps modems also used 1 bit per baud: [...]
>
>[... discussion of modulation and duplexing methods ...]
Without wishing to appear /too/ much of a smartarse, or wishing
to thrash out an almost dead topic, just one small point, Paul:
In one of your other posts you (quite correctly) pointed out to
someone that using the phrase `baud rate' was dimensionally not
correct --- `baud' == `line changes per second', so `baud rate'
== `line changes per second per second' (assuming the time unit
is the second). Here you talk about `bits per baud'; does this
not commit the same crime? Shouldn't it be `this modem encodes
42 bits per baud-second' (since baud-second == line change)? I
would set followups to alt.syntax.tactical, but then I wouldn't
get to see them. I would also insert a smiley, but I am trying
to give them up.
--
Ben North
Computer Officer, New College, Oxford.