I am trying to convert "words per minute" into "bits per second."
Bits per second, in turn, is APPROXIMATELY equal to baud, a common
measure of modem (or other means of data transmission) speed. I need to
quantify one factor: How many letters are in a "word?" If we assume that
there are 5 (five) letters to a word, my calculations look like this:


WPM = 50
LPM = WPM * 5 # letters per minute
BPM = LPM * 8 # bits per minute
BPS = BPM / 60 # bits per second
BPS = 33.33


I have assumed 8 bits to the byte, which is quite generous
considering that Morse cannot encode an 8 bit character set or, for that
matter, the full ASCII character set, which is only 7 bit.
Can anyone see any obvious errors? Is 50 words per minute really
equal to about 33 baud?

--
Klystron

Here is how Morse speed is usually calculated:

If the text is typical plain-language English, the test word "PARIS"
is used. WPM is the number of times PARIS can be sent in 1 minute,
using proper spacing between dits and dahs, letters, and words.

It turns out that the word PARIS and one word space equals exactly 50
"dit times", with a dit time being the length of time the key is
closed for a dit. (A dah is three dit times, the spaces between dits
and dahs inside a character are one dit time, the spaces between
letters are three dit times and the spaces between words are seven dit
times.)

So if the word PARIS is sent 50 times in 1 minute, that minute is
divided into 2500 dit times. Which is 41.66 bps.

The reason for the difference is that there are so many different
timing issues in Morse Code. The elements, characters and spaces are
all different lengths, with the most-common characters (like the
letter E) being the shortest.

73 de Jim, N2EY

wrote:
Here is how Morse speed is usually calculated:

If the text is typical plain-language English, the test word "PARIS"
is used. WPM is the number of times PARIS can be sent in 1 minute,
using proper spacing between dits and dahs, letters, and words.

It turns out that the word PARIS and one word space equals exactly 50
"dit times", with a dit time being the length of time the key is
closed for a dit. (A dah is three dit times, the spaces between dits
and dahs inside a character are one dit time, the spaces between
letters are three dit times and the spaces between words are seven dit
times.)

So if the word PARIS is sent 50 times in 1 minute, that minute is
divided into 2500 dit times. Which is 41.66 bps.

The reason for the difference is that there are so many different
timing issues in Morse Code. The elements, characters and spaces are
all different lengths, with the most-common characters (like the
letter E) being the shortest.

73 de Jim, N2EY


Jim,

I'm sure your explanation is correct, but it leaves me confused: I know
bps baud, but they're close, and the Model 15 Teletype I used to own
operated at 45 baud. It seems illogical that Morse would be so high in
the bps count.

73, Bill

--
Bill Horne, W1AC

(Remove QRM from my address for direct replies.)

On Mar 18, 3:30 pm, Bill Horne wrote:
wrote:
Here is how Morse speed is usually calculated:

It turns out that the word PARIS and one word space equals exactly 50
"dit times", with a dit time being the length of time the key is
closed for a dit.

So if the word PARIS is sent 50 times in 1 minute, that minute is
divided into 2500 dit times. Which is 41.66 bps.

I'm sure your explanation is correct, but it leaves me confused: I know
bps baud, but they're close, and the Model 15 Teletype I used to own
operated at 45 baud. It seems illogical that Morse would be so high in
the bps count.

The difference has to do with how the coding is done. The following is
all from memory:

60 wpm Morse works out to 3000 bits per minute or 50 bits per second
using the "PARIS" formula.

Your 45 baud Model 15 Teletype was in all probability what hams called
a "60 wpm 5-level Baudot" machine. We had similar machines at the
University. (In this post I use the term "Baudot" to mean the 5-level
TTY code US hams used for many years until FCC allowed us other codes
like ASCII in the early 1980s)

"Baudot" takes 7 bits to send a character: one start bit, five data
bits, one stop bit. A space between words is a character, so to send
the word "PARIS" would take six characters including the space
character. That's only 42 bits, rather than the 50 bits that Morse
requires. Thus the difference - the Baudot machine uses 16% less bits
to send the same message. The speed difference works out to about 10%
because the Baudot stop bit was longer than the others in the machines
US hams typically used.

So you don't get the full 16% advantage that you'd expect from the raw
numbers. But since only six of the 42 bits are stop bits, the
difference is small.

To make it even more of a sporting course, the above WPM advantage of
the Baudot machine is message-dependent, same as for Morse. In Morse,
the message-dependency comes from the different characters being of
different length; a five-letter word like "TENET" takes a lot less
time to send than one like "JUICY", while in Baudot they both take the
same time to send.

But in the Baudot code the numbers and some other characters are sent
by shifting from "LTRS" to "FIGS", (letters to figures), so sending
mixed groups could take a lot of extra characters that Morse does not
require.

For example, in Morse you could just send the group "6A8G7" as 5
characters, but to send it on a Baudot machine you had to send
"figs6ltrsAfigs8ltrsGfigs7", which is 10 characters.

So the WPM are really approximations, and the BPS/baud measures took
over.

73 de Jim, N2EY

In article ,
Bill Horne wrote:
I'm sure your explanation is correct, but it leaves me confused: I know
bps baud, but they're close, and the Model 15 Teletype I used to own
operated at 45 baud. It seems illogical that Morse would be so high in
the bps count.


Your Model 15 Teletype at the nominal 60 wpm speed, which is actually
368 chars/minute and 45.45 baud works out like this. The character
length is 7.42 bits long (for ancient, interesting reasons I won't go
into right now) and the bit duration is 22 milliseconds. The character
duration is therefore 7.42 * 22 = 163.24 milliseconds, and that works out
to 6.12595 characters/sec = 367.55 characters/minute. To convert that
to words you have to figure 6 characters per word because the space
between words is also a character. So the speed is actually 61.26
words/minute.

Teletype speed is sometimes confusing because there are a couple of
other speeds out there. Western Union liked to use a 7.00 unit
character rather than 7.42. With 45.45 baud, or 22 ms pulses, this
gives 154 milliseconds/character, or 6.49 characters/second, 389.6
character/min and hence 65.9 words/minute. This is completely
compatible with 7.42 unit code because the baud rate is 45.45 for
both. But then there is European 50 baud Telex using a 7.5 unit
code. This is a 20 millisecond bit for a character length of 150
milliseconds, 6.67 characters/second, 402 chars/minute, 67 words
per minute. This is not compatible with the other two codes because
the baud rate is different; but if you say something like "66 wpm"
you could be talking about either scheme.

Now when you get to ASCII, the old Teletype machines transmitted 8
data bits per character and used an 11.0 unit code. This makes 100
wpm work out to 110 baud. Electronic terminals don't need 11 unit
code; they can do just fine with 10. Thus the words-per-minute is
numerically equal to the baud rate. 100 baud - 10 ms/bit -
100 ms/char - 10 chars/sec - 600 chars/min - 100 wpm.

Morse has already been explained. A Morse dot is actually two bits,
since there is the dot followed by the space that makes it distinguishable
from what comes next. A Morse dash is four bits, counting the space,
and the word space is three dot times or 6 bit times. Then the
word PARIS contains 50 bit times counting the space. So one word
per second is 50 bits per second and 60 wpm. As an aside, the
military sends a lot of encrypted 5-letter code groups, so instead
of PARIS the Signal Corps uses CODEZ as a test word more statistically
correct for their kind of traffic. And CODEZ contains 60 bits.

Jim Haynes posted on Tue, 18 Mar 2008 19:47:44 EDT:

Bill Horne wrote:

I'm sure your explanation is correct, but it leaves me confused: I know
bps baud, but they're close, and the Model 15 Teletype I used to own
operated at 45 baud. It seems illogical that Morse would be so high in
the bps count.

Your Model 15 Teletype at the nominal 60 wpm speed, which is actually
368 chars/minute and 45.45 baud works out like this. The character
length is 7.42 bits long (for ancient, interesting reasons I won't go
into right now) and the bit duration is 22 milliseconds. The character
duration is therefore 7.42 * 22 = 163.24 milliseconds, and that works out
to 6.12595 characters/sec = 367.55 characters/minute. To convert that
to words you have to figure 6 characters per word because the space
between words is also a character. So the speed is actually 61.26
words/minute.

For what it is worth, my paper reference on TTYs is NAVSHIPS
0967-255-0010 "Principles of Telegraphy (Teletypewriter)" from
Department of the Navy Electronic Systems Command. I bought it
from the US Government Printing Office back in the early 1970s as
a reference. [I am an Army veteran, not Navy] The first chapter of
Part A in that TM has a nice historical record of 'telegraphy'
(which includes teletypewriting). It says only "60 WPM" but mentions
other Baud rates.

As far as we in Army communications of the mid-1950s were concerned,
all the teletypewriters that the Army used were called "60 WPM" and
only the teletypewriter maintenance people (and a few carrier systems
types) cared about many numbers. We did have Distortion meters used
to determine irregularities in a circuit.

BTW, the Army and the rest of the military used Teletype Corporation
Model 15s through 19s, variation being only the paper tape punch
and transmitting distributor (P-tape reader).


Now when you get to ASCII, the old Teletype machines transmitted 8
data bits per character and used an 11.0 unit code. This makes 100
wpm work out to 110 baud. Electronic terminals don't need 11 unit
code; they can do just fine with 10. Thus the words-per-minute is
numerically equal to the baud rate. 100 baud - 10 ms/bit -
100 ms/char - 10 chars/sec - 600 chars/min - 100 wpm.

OK on that. Teletype Corporation Model 28s (explained in intimate
detail in the NAVSHIPS TM I referenced) would easily do 100 WPM
equivalent 24/7 as long as supplied with paper, ribbon, and oil. :-)


...word PARIS contains 50 bit times counting the space. So one word
per second is 50 bits per second and 60 wpm. As an aside, the
military sends a lot of encrypted 5-letter code groups, so instead
of PARIS the Signal Corps uses CODEZ as a test word more statistically
correct for their kind of traffic. And CODEZ contains 60 bits.

I never encountered any test word 'CODEZ' 1953 to 1956, nor
elsewhere in the Signal Corps or in DoD contracts after that. In
the mid-50s we simply used a continuous 'R-Y' generator (from
Teletype Corporation) for circuit checks with the old 60 WPM
equivalent machines. Teletype Corp. also made a 'fox test'
generator consisting of a half dozen cams operating as many
switches to generate "The quick brown fox jumped..." sentence
(with Tx station ID at the end) for radio circuit checks.

Electromechanical teletypewriters are now rather passe' in the
military and government (I use a French word to replace Obsolete
which so many have trouble with). It is all electronic and, for
permanent installations, over the DSN (Digital Switched Network)
anywhere...including interfaces with the regular civilian telephone
infrastructure. The DSN allows encryption on-line as per protocol.
For field radios, the electronic data protocols are compatible
with hard-wired ones and also allow encryption on-line. It was so
in the first Gulf War (1990-1991) which 'battle tested' the whole
military communications network DX to no-DX via TDRS (Tracking
and Data Relay Satellites) and other military commsats from/to
Florida to/from the Middle East.

Not having any access to the DSN or intimate details of military
cryptographic equipment now, I have no exact knowledge of what is
used for a test word, sentence, or whatever. For the OLD electro-
mechanical teleprinters, I'd say the specifications for a
specific TTY Distortion Meter would tell the exact story on
timing for both polar and non-polar TTY circuits and equipment.

73, Len AF6AY

wrote:

[...]
So if the word PARIS is sent 50 times in 1 minute, that minute is
divided into 2500 dit times. Which is 41.66 bps.
[...]


It still seems like an awfully slow data rate. I have seen people
throw 14400 baud modems in the garbage because they considered them to
be so slow as to be worthless. A data rate of 42 bps is about 3 orders
of magnitude slower than that. It just seems inconsistent with the way
that so many hams have fought tooth and nail to hold onto Morse and to
hinder the move toward digital modes.

--
Klystron

On Sat, 22 Mar 2008 14:04:16 EDT, Klystron wrote:

It just seems inconsistent with the way
that so many hams have fought tooth and nail to hold onto Morse and to
hinder the move toward digital modes.

The joy of Morse is not the speed at which data is transferred but the
means of transferring. A good Morseist (mot me....) doesn't need a
computer or software to decode it.

And I know several Morseists who not only use "high speed data modes'
in addition to using Morse, but hold advanced degrees in development
of those modes.

Morse is for fun.
--

73 de K2ASP - Phil Kane

From a Clearing in the Silicon Forest
Beaverton (Washington County) Oregon

e-mail: k2asp [at] arrl [dot] net

Phil Kane wrote:

Morse is for fun.

Indeed, this says it all.

73, Steve KB9X

On Mar 22, 1:04�pm, Klystron wrote:
wrote:
So if the word PARIS is sent 50 times in 1 minute, that minute is
divided into 2500 dit times. Which is 41.66 bps.

� �It still seems like an awfully slow data rate.

Compared to what? And for what application, in what bandwidth?

If you have a pile of data to send, or a picture, etc., 41.66 bps is
quite slow.

But for a real-time conversation, 41.66 bps isn't all that slow. The
average person doesn't talk or type at a sustained speed much faster
than 100 wpm. 50 wpm isn't that much slower.

I have seen people
throw 14400 baud modems in the garbage because they
considered them to
be so slow as to be worthless.

11 years ago, when I first went online, it was with a 56k modem. I
gave up on dialup modems several years ago and went broadband. I don't
think anybody who has a choice is still using dialup.

But that's because the options exist, with no significant downsides. A
14400 modem uses the same phone line as a 56K modem. DSL can be run on
the same phone line and not tie it up for telephone calls.

Operating on the limited bandwidth amd high variability of the HF
amateur bands is a completely different thing.

A data rate of 42 bps is about 3 orders
of magnitude slower than that. It just seems inconsistent
with the way
that so many hams have fought tooth and nail to hold
onto Morse and to
hinder the move toward digital modes.

A lot of hams like Morse Code and use it on the air. It has a lot
of advantages. Why should they give it up?

And how has "the move toward digital modes" been hindered by hams?

73 de Jim, N2EY