[comp.sys.acorn] High res mono mode 23 with Taxan 795

bdb@cl.cam.ac.uk (Brian Brunswick) (01/30/91)

While messing about one day, I happened to try Mode 23 from my A440 on
my Taxan 795 monitor, (by typing *CONFIGURE MONITORTYPE 2, which makes
all modes into 23) and it successfully synced to it. However, the
result appeared very strange, since I was still using the normal
monitor lead.  (This is because the high resolution mono mode relies
on a special serialisation circuit which converts one 16 colour pixel
into 4 adjacent black and white pixels by splitting out the bits, and
this output only goes to the 2 BNC connectors on a 400 series, not the
normal video connector.) I determined to make up a lead to get a
proper view of this mode, and I will detail here how this was
eventually done.

The first problem was obtaining a 15 pin (three row) D-type connector
(male) to plug into the 795, which took several weeks because it was a
strange thing, and therefore out of stock. Eventually, this arrived,
and I tried making a lead, working from the rather sparse information
in the 795 manual, which gave only a pin out diagram for the
D-connector, and no explanation at all.  From past experience, I
distrusted trying to make up BNC plugs, so I used some BNC to phono
connectors, and some phono plugs, and also some two core separately
shielded cable that I had around.  What I made at first didn't work,
but by circuit testing the standard lead, I eventually found that the
secret was to feed composite sync only to the hsync input of the 795,
and leave the vsync unconnected. Naturally, there was no mention of
this in the 795 manual.

So what I had then, was a fairly blurry picture, on which it was not
really possible to even distinguish the background pattern, which has
a 4x4 pixel cross hatch.

Wiggling wires soon discovered that it was the phono plugs at fault,
and so I then did what I should have at the start, which was to get a
ready made BNC to BNC lead, and cut it in half. The resulting lead,
with the D-connector in a shielding metal case, and stuffed two or
three times through an old ferrite core that I had lying around
produces quite an adequately sharp picture. Individual pixels are not
quite visible, but this is understandable considering it is exceeding
the 795's 50Mhz pixel rate with 64Mhz, and its stated resolution of
1024x768.

Now personally, I don't think I will actually be using mode 23 much.
Its resolution is only 1152x896 in 2 colours, and quite frankly I
prefer the VIDC enhanced 1152x448 in 16 colour mode that is
obtainable, with the same apparant wimp scale. But this may well be
useful to those needing to test stuff in all modes, or emulate X
windows, or whatever. What is really needed is a switch over method,
but I doubt that can be built sufficiently cleanly.

Anyway, the details of connection are as follows:

15 pin D-type connector, looking into monitor, or back of plug
-----------------------------------------
\                 id                     |
 |           GND gnd  B   G   R         /
 \            5   4   3   2   1        /
  |                   +---+---+---- connect all 3 to core of signal BNC lead
  \                                   /
   |               KEY  corresp ground returns
   \            10  9   8   7   6    /
    |            +      +---+---+---- connect all 3 to ground of singal BNC lead
ground of sync BNC lead             |
     \            syncs    Id      /
      |      NC   v   h  NC  GND   |
      \      15  14  13  12  11   /
       |              +           |
      core of sync BNC lead      /
        --------------------------

If anybody tries this, I'd be entertained to hear from you how useful you
found it.
Brian.Brunswick@uk.ac.cam.cl  Disclaimer.  Short sig rules!