[GreenKeys] Gears and Baud rates

Don Robert House [email protected]
Sat, 15 Feb 2003 16:49:52 -0800


MANY THANKS BEN!!
Don

>...I checked the speed and pulse information from a chart I used in 
>the applications engineering department of Teletype in the 1960's.
>
>The receiving shaft always runs faster than the transmitting shaft. 
>This is because the receiving selector comes to a screeching halt 
>before the end of the stop pulse.  It restarts at the leading edge 
>of the next start pulse, which is the trick to keeping the receiving 
>selector in synch with the transmitted signal.  Because of this 
>trick, a receiving machine will copy any signal regardless of unit 
>code (i.e., the length of the stop pulse) as long as the baud rate 
>is the same.  This way, any 5-bit machines can communicate with any 
>other as long as they are 45.5 baud, which works out to a 22.0 
>millisecond bit length.  This is the standard for so-called 60 words 
>per minute operation.
>
>Various services used different length stop pulses.  The military 
>often used 7.00 unit code, the Bell System used 7.42 unit code due 
>to a historic requirement to make competing Bell Labs and Teletype 
>designs compatible, and Europe (e.g., Telex) used 7.50 unit code. 
>The shorter stop pulse signals obviously ran at a slightly faster 
>WPM rate than those with longer stop pulses.  However, if the 
>receiver got out of synch due to a line hit, the slower machines 
>with longer stop pulses would re-synch more quickly.
>
>So what this comes down to is that the receiving shaft on a 60 wpm 
>45.5 baud machine always was geared at 420 rpm.  It didn't care how 
>long the stop pulse was, because it always stopped before the end of 
>the stop pulse.
>
>The different gears are for the transmitting shaft, which ran at the 
>full character rate.  The correct numbers for 45.5 baud are:
>
>        7.00 UC: 390 rpm, 65 words per minute
>        7.42 UC: 368 rpm, 61.3 words per minute
>        7.50 UC: 364 rpm, 60.7 words per minute
>
>These are all generically referred to as "60 words per minute," and 
>they are all compatible as long as they run 45.5 baud.
>
>Now back to your e-mail, your 368 rpm gears are the old Bell System 
>standard, and the transmitting cam is ground for 7.42 Unit Code. 
>Your "387.7 oddball" is probably working on a transmitting cam set 
>at 7.00 Unit Code, which would make it completely compatible with 
>any other 45.5. baud machine. 
>
>Remember that these machines do not enjoy the precision of our 
>carefully clocked electronic circuits of today.  The cams were 
>ground on machines which were not perfect, and even if they were 
>perfect, wear would in time make them less than perfect.  We used 
>3-digit figures (not four as you used) and figured they were close 
>enough.  In the real world, the receiving selector was adjusted a 
>bit this way or that to make up for variations in the transmitted 
>signal.  To adjust the range finder of the receiving selector, tune 
>in a good test pattern, and then turn the range finder each 
>direction to locate the points where it goes from clear text to 
>garbage, and then set it halfway between those two extremes.  Lock 
>it in place, oil it occasionally, and it should give you clear copy 
>forever.=
>
>Ben Stephens K9KOM=


-- 
--------------------------------------------------------
Don Robert House
North American Data Communications Museum
URL: http://www.nadcomm.org
Computer Museum of America
URL: http://www.computer-museum.org