On 2017-Jun-14, at 7:44 AM, Noel Chiappa via cctalk wrote: >> From: Brent Hilpert > >>> The trimpot on the board says to me that the clock is most likely a >>> simple RC affair. > >> There's a 7493 (4-bit counter) on the board as well, which looks to >> have connections to the dip switches beside it, in all likelihood the >> baud rate divider and rate selection > > Yes; but the trim pot is also there because one has to set the basic clock to > one of two values, depending on whether one wants the 150/.../2400 > selections, or the 110/etc selections; one has to set it to 26 usec for the > former, and 35.5 usec for the latter. > > So drift would have been an issue, but not initial component values...
Certainly calibration was required, I don't expect anyone was making boards like this expecting to get the target timing from fixed/off-the-shelf component values (although in theory one might if using a sine-wave rather than switching oscillator). There are two trimpots on the board, they could be one for each of the rate series and switch-selected, so one wouldn't have to recalibrate between rate selections.
