On 11/20/24 08:19, Chris Kmiec wrote:
Hmm, I don’t see that feature in the specs for REX#. It has 1mb of Flash
memory for storing ROM images and can backup and restore the ram, but I
don’t see and RAM on REX that would allow me to max out Tandy’s onboard RAM.
There's a few different things to explain here...
The machine can only address up to 32k at a time, so in one sense there
is no ram expansion beyond filling all 4 possible 8k sockets inside. You
have a few different options for that. GGLabs sells a module called
M10-RAM that you can buy. They have a site for the info but the actual
buy link is on ebay or tindie etc. Otherwise you can build your own by
following https://github.com/bkw777/Model_T_RAM
(Apologies but this mail will end up having a few links to things from
me on github since I've been playing with several forms of ram & rom
addons for the 100-200 the last few years)
There have been several forms of ram expansion that connect to the
system bus and work by containing multiple banks of 32k and you have a
software method to switch from one bank to another. You can only access
the contents of a single bank at any one time. It's like unplugging all
your chips and putting other chips in. I think the PG Design ones driver
software might have some feature to be able to move a file from one bank
to another, so there was some just barely form of access to the other
banks. No one makes those currently but years ago Steve Adolf made
something called QUAD which replicated the PG Design, and more recently
I replicated & updated QUAD to make reQUAD
https://github.com/bkw777/reQUAD
No one makes it for sale but you can buy all the parts and build one
yourself or find someone to do it for you.
All 3 types of REX, (REX Classic, REX#, REXCPM) include a feature to do
full backups & restore of the 32k internal ram. It's not quite the
one-button instant swap like with the ram expansions, but just about.
It's only a few presses to go into REXMGR and do a backup & then a
restore. In fact I think he might have even streamlined that operation
in the latest version for REX#. I know he did add some overwrite
protection to avoid accidental deletes or overwrites. So in the end,
it's functionally almost the same as having a ram expansion. It's just
that underneath what's really happening is you are dumping the current
ram out to a chunk of flash and then restoring some other chunk of flash
over the ram. The actual ram expansions only gave you essentially the
same outward functionality, just a little faster.
REXCPM is a special case in that it also actually replaces your internal
ram with it's own 32k. So if you have less than 32k, a REXCPM would give
you 32k.
If Steve can't sell you a REX# or REXCPM at the moment, there is an
option to build your own REX Classic.
https://github.com/bkw777/REX_Classic
Not ram but probably the more practical way to have more room to work is
disk space. The most practical answer for that is a Backpack.
https://www.soigeneris.com/universal-backpack-drive
I have something you can build called PDDuino which actually predates
backpack, but backpack is far more feature rich and polished. (From what
I can tell. I have not used one because since I want to work on the
fully open source option that does the same thing, I don't want to ever
be accused of even seeing the firmware in the backpack.)
https://github.com/bkw777/PDDuino
Basically the idea is 99% of the hardware is just an off-the-shelf
Adafruit Feather or Teensy dev board and you just plug it in to a little
serial port adapter board. It has a rechargeable lithium cell you
recharge by usb, or you can just run on usb alone, or you can even run
on power tapped from the BCR port.
I've also made a version of a kind of ram disk that seems like at one
time a few different people all made versions of essentially the same
thing. It's a very neat little circuit where basically it's nothing but
an address latch, a binary counter, and some ram, and that ends up
emulating a spinning disk. The device only supports 2 functions, select
a block address, and read/write a byte. Each time you select a block
address, the binary counter is also reset to 0. Each time your read or
write a byte, the binary counter is also advanced by 1. A block has 1024
bytes. So You select say block 40, then just read 1024 times to read the
entire block. Like picking a track number and reading the track.
https://github.com/bkw777/NODE_DATAPAC?tab=readme-ov-file#minindp
But this needs to connect to the system bus and that's inconvenient on a
Model 100 because it would need to be on a cable extending out to the
back of the machine. You can actually use the Model 100 adapter portion
of a DVI (Disk/Video Interface) cable for this.
But really it's only good on a 102 or 200 where it just plugs directly
onto the back of the machine.
And what you get is disk space not ram. You could get disk space in a
generally more convenient and flexible form from Backpack or PDDuino
because those use the TPDD protocol, for which there are several bits of
compatible software and hardware. IE, you have a choice of a few
different kinds of TPDD driver software to run on the 100, and then the
drive can also be used by other machines. the TANDY WP-2 has built-in
tpdd support in it's firmware. You can connect a real TPDD drive or a
backpack/pdduino to a PC and there are several softwares to access the
drive or emulate a drive in any os, since the connection is just rs-232
and the protocol is known.
While the RAMPAC-alikes there is no pc software to access the drive, and
no hardware adapter to somehow connect it to a pc anyway. There are a
few utils that made it a little more convenient to move files between
the rampac and a tpdd, so to move files between rampac and pc you would
use a 102/200 in the middle with the pc running a tpdd emulator.
Still it has some little advantages. Small and simple. The driver
software is smaller than most TPDD dos's, much smaller than TS-DOS but
larger than Teeny. And it holds 512k.
And the low level hardware interface is so simple and easy to operate
directly even from basic, that you can actually use the raw space almost
like ram if you are writing your own program.
I have a 1 Meg version that actually has the entire hardware interface
described right on the silkscreen on the back. because it's basically
just 3 instructions with a tiny bit of explanation for the possible
values for each. The normal RAMDSK software doesn't know how to access
the extra space beyond banks 0 & 1, so the only way to use that is to
write your own program that accesses banks 2 & 3 directly and uses it as
raw space.
TPDD is not convenient to access at the low level. It's only convenient
at the user level using TS-DOS. Although TS-DOS does provide some hooks
to provide high level file access functions in basic or from machine
language.
--
bkw
Chris
On Tue, Nov 19, 2024 at 7:44 PM Bert Put <[email protected]
<mailto:[email protected]>> wrote:
Hi Chris,
Not the screen so much, but your REX will already give you oodles of
RAM
in banks of 32K. So you can swap out and in, as many RAM banks as your
REX has space for.
I just discovered that recently while playing with my (original) REX,
M-100 and T-102.
Cheers, Bert
On 11/19/24 17:59, Chris Kmiec wrote:
> I have to admit, I have not been following the latest
developments for
> the 100/102. I have the backpack, REX# and REXCPM, but they have
been
> sitting lonely as for the past year or so my hobby time has been
taken
> over by the HP 71b.
>
> Anyway, I want to start playing with the 100-series again, but
have a
> question - are there any modern upgrades for the LCD screen and
increase
> in RAM beyond the 32Kb?
--
bkw
