"The console typewriter was replaced with a modified Selectric typewriter, which could type at 15 cps — a 50% improvement over the Model I."

Actually, I seem to remember it as 14.7cps -- can someone find out the truth? This relates to the communications channel, bits per char, baud rates, etc... Those were the days! 69.87.204.125 01:46, 17 September 2006 (UTC)[reply]

The figure 15cps came from Joe Crespo, the IBM engineer that designed the Selectric interface for the 1620. However he may have rounded the number to the nearest whole character.

The interface was parallel, one whole character at a time (not serial). On the Model I the typewriter interface took the 6 bit (and 1 parity bit) BCD character code and decoded it in relays to individual signals, one per key solenoid. The Selectric used with the Model II reduced this to 7 solenoids, driving the mechanical decoder internal to the standard Selectric that controlled the ball. Character timing in both the Model I and Model II was controlled by a motor driven cam assembly that switched microswitches on and off.

I have some original documentation on the Model II (but not schematics like I have on the Model I). If I get a chance I'll dig it out and see if it clarifies the detail (e.g., RPM of the timing cam assembly). -- RTC 19:29, 17 September 2006 (UTC)[reply]

OK, I looked it up: IBM Form # A26-5781-0 IBM 1620 Central Processing Unit, Model 2 on page 98 in the section on the Console Typewriter it lists both input and output speeds as 15.5 cps. I'll correct the article. -- RTC 07:48, 21 September 2006 (UTC)[reply]