> On Apr 3, 2024, at 11:16 AM, Paul Koning via cctalk <[email protected]>
> wrote:
>
>
>
>> On Apr 3, 2024, at 11:01 AM, Guy Fedorkow via cctalk <[email protected]>
>> wrote:
>>
>> Vintage computer enthusiasts might want to keep track of where to find
>> CRT-based analog oscilloscopes, for use as output devices.
>> The early MIT and Lincoln Labs computers used D/A converters to steer and
>> activate the beam on analog scopes to draw vector images.
>> Working on Whirlwind simulation, we've been able to get this technique to
>> work with "real" oscilloscopes, e.g., Tek 475, but we have not yet found a
>> single DSO that has X/Y _and_ Z inputs (let alone the required phosphor
>> fade).
>
> So did a whole range of DEC computers, of course. And the famous CDC
> mainframe console (DD60) though it did vectors only for text (graphics was
> dot-mode only since it wasn't a major use case for that device).
The DD60 and its associated controller in the mainframe (6612 or 6602) was an
interesting beast. The interface between controller and display is a hybrid,
with the positioning information delivered as 9 bits each of X and Y, but the
character vectors are generated in the controller and sent to the display as
analog waveforms, X and Y on differential pairs.
Another oddity is the character waveform generation: that uses two pairs of A/D
converters, and the converters are essentially base one -- 6 equally weighted
inputs to produce output values 0..6. And since ROMs were hard to come by in
1964, at least ones with 100 ns cycle time, the digital inputs for the waveform
generators are an amazingly large pile of gates.
paul
