On 2013 Mar 13, at 2:04 PM, Tony Duell wrote:
 
 Idle curiousity, does anyone happen to know whether the TMS4100 ROM
 was a standard off-the-shelf character generator from TI, or was it a
 generic mask-ROM-to-customer-order, or otherwise have data on it such
 as it's native size. 
 The 1976 'Memroy and Microprocessor Databook' (Texas Instruemnts)
 lists a
 TMS4102 Cheracter Gerneator. Like the one in the HP9830, it reads
 out a
 column at a time, and has the same pinout. I will give the pinout fro
 mthe databook in a  minute. The difference seems to be that the
 TMS4103
 is specified as having 'Inputs and Outputs Fully TTL compatible) while
 the inputs o ntne TMS4100 seem to be 12V levels. The HP9830 Dispaly
 Driver PCB (09830-66542) drives them with high voltabge open-collector
 buffers ('145 and '07), pulled up to 12V via a 1k2 resistor. 
 
<snip>
(I had the pinout)
  Unforutnately there is only 1 page of data on it in
the book and it
 doesn't give the paterns for the characters. The data sheet is
 dated May
 1975 (after hte design of the HP9830, of course).
 Full id:
    TMS4100
    ZA 5705
    T-04163
    7308   <-- date code
 It's the character generator in the HP9830, arranged for character
 column scanning order rather than row scanning (7-bits columns at the
 outputs rather than 5-bit rows).
 I see three possibilities:
    - the TI TMS4100 is a character generator ROM
    - it's a TI TMS 4100 mask ROM, programmed and sold by TI as a
 character gen
    - it's a TI TMS 4100 mask ROM, programmed to order for HP 
 My _guess_ is that the TMS4100 is a standrd character generator ROM
 and the
 TSM4103 is an imporved version with TTL-compatible inputs.
 The ratehr odd address inpout (1-of-5 colum select, for example)
 seems to
 suggst it's not a standard ROM device (I would epxpect fully bianry
 coded
 address inputso n such a device).
 .. 
Another list member (Rob) found the ref in a 1970/1 TI book.
Apparently, it's part of a character generator series, the 4103 is
the off-the-shelf standard ASCII version in the series.
As Rob pointed out, the 9830 one must be special order to HP though,
to obtain the lazy-T cursor, unless there's some alternative tortured
way they're injecting the T into the display scanning that I'm not
seeing.
Yes, the-1-of-5 select did suggest it was an intended character
generator (although in principle it could be a straight ROM depending
upon how much ROM space one is willing to waste, although (back on
the other hand) being that wasteful would have been pushing it for
ROM sizes of the period, and the 9830 has the binary encoding right
there anyways).
The earlier data only indicates MOS levels (for TMS 4103 JC/NC).