My first run was using a 28C256 on a SOIC-28 chip. Same thing you did, I
used a resistor to hold /WE so that I could pull it with the burner when
I wanted to erase/re-write the image. But, also like you said, those
chips are $10 each now and hold a single image. I built the carrier
(actually, designed in KiCAD and sent to GoldPhoenix for fab), soldered
the SOIC on and burned it using a DIP-28 to "Molex" adapter that I built
out of Cat5E wire and a DIP header, with 1" long "fingers" on the other
size made out of gold-plated spring wire used for craft jewelry. The
carrier clips into that 28-pin finger arrange, which has all the pins
for a 28C256 DIP-28 re-mapped to the Molex pinout and the burner just
sees it as a DIP-28. Works pretty good, but it's a pain, and it's expensive.
So the PROM version uses a AT27C020 in PLCC-32 format on a different (of
course) carrier, and via a suite of programs I wrote in Perl, builds a
single ROM image out of 8 32k images, which I burn in the TL866-II. Then
I solder the chip to the carrier, and the DIP switches select which
32-image is available at any given time. Simple enough, and it works
very nicely. I have several of my prototypes in my T102 and several
T200s. No complaints. The chip is $4 at Digikey, and as long as I get
the image right, it's one-and-done.
This is a working project. It's done, except for the actual ROM images.
I have the boards fabricated, the chips on hand, and have a half-dozen
working prototypes on my bench. I know the REX is a great device, but
sometimes K.I.S.S. wins out in my head. Sometimes I just need a ROM in
that slot all the time, and since I have a good number of Ts that are
all used in various ways, I can't justify putting a REX in each.
--Justin
On 7/8/21 3:19 PM, Brian K. White wrote:
On 7/8/21 10:06 AM, Justin Poirier wrote:
It appears that Club100 on bitchin100 only have a handful of ROM
images. Where do I go to find SuperROM, Disk+ and those others? They
have generic enough names that Google has been of very little help.
I have been working on an inexpensive carrier solution (for myself,
mostly) that will hold (8) ROM images that are selected with a group
of DIP switches on the carrier itself. Nothing fancy, but if I like
the results, I could probably crank them out, burned and ready, in
the $20-$25 range. Not committing to anything at all, since I’m still
in prototyping, but with (8) ROM slots, I’m not sure what to put in
them. So far, I made one that has TS-DOS, Ultimate, Cleusseau and
TS-Random. And since I have twice as much space as all that, it
includes those titles in both the M100/102 versions as well as the
M200 version. That seems wasteful. Maybe I’m wrong!
How are you connecting up to burn them? Through the edge connectors
with a reverse pinout adapter? DIP-28 test clip on the outside edges
with the wires arranged into a reverse pinout adapter? Or are you just
burning before soldering and no re-writing after that?
I made this single-rom carrier that, since it's an SOIC package, and I
have a resistor rather than a trace for /WE, is easy to just connect
normally with a soic test clip to program.
http://tandy.wiki/Teeprom
But that 28C256 is now OVER $10 just for a single 32k, and I'd like to
try to ditch the requirement for the test clip if possible, and
definitely don't want to require an actual Molex socket. (I have molex
sockets, but I'm trying to make a design anyone can use, not just
something for myself) So I tried this
https://github.com/bkw777/Teeprom/blob/master/Teeprom2.md
4 or 8 roms (that's just a 4-rom version but 8 would be a
straightforward progression from there) and no special parts needed,
and it's even both cheaper and more convenient than the soic-28 test
clip, and the flash part is both more readily available (multiple
manufacturers still) and just over $1 instead of over $10.
Which *almost* works as envisioned. The programming adapter is built
out of all normal off the shelf parts, and the connection between the
programming adapter and the carrier is made by dint of having the
holes on the carrier be offset staggered so they work against each
other. I think I just need a different stagger pattern, and slightly
more offset to the stagger. I was able to get it to work by tediously
testing each pin for connection to identify a handful that didn't
connect, and bending the pins manually until they all worked.
Not practical.
But that was only the first proof of concept, no iteration yet, so
maybe with a little dialing-in the idea would work out.
But then again maybe that many little pins in that kind of arrangement
is just never going to be reliable.
So I was thinking of next either using pogo pins, because those are
actually cheap now, or using long wire-wrapping pins to make something
that can act like a DIP test clip that can just contact all the edge
contact pins on the outside like a normal socket does. That would make
the carrier a LOT simpler!
I really thought those pins were slick ;) If they would just all
actually make contact, it does pretty much work as expected, meaning
it wasn't too hard to build and that carrier pops right onto the
programmer simply and even with polarity enforcement.
I've gathered links to all the roms I've seen here:
http://tandy.wiki/ModelT_roms
Most you can get all in one spot from Steve's REX docs on bitchin100
(link in there), but there are a few other oddballs.
Documentation has not been gathered into one convenient spot that I
know of, other than club100 which is a good start but incomplete, and
not going to get any better, it's a static site now just being
presereved.
You get info from searching through the M100SIG
https://archive.org/details/M100SIG
, club100, scanned magazines on archive.org, and general google for
info on other vintage computer sites. I don't have a link farm handy
to list those.
