YAMOD.ATBUS - fast IDE multiplexer
3 messages · 2004-07-01 → 2004-07-06 · Yahoo Group era · View archive on archive.org
Participants: Jarek Adamski, aralbrec, Jack Boatwright
Preserved from the Timex/Sinclair 2068 Yahoo Group (2001–2019), which is no longer online. Text reproduced from the archive.org archive; email addresses masked.
Messages
1. YAMOD.ATBUS - fast IDE multiplexer
Jarek Adamski · Thu, 01 Jul 2004 10:27
Hello,
I would like to ask you, if you were interested in a board
with fast IDE multiplexer with real time clock as an option
(details below).
I'm going to order the YAMOD.ATBUS boards soon and I would
like to take into account your needs in that matter.
So, read the description below and wonder, if you want to
get such devics. If yes, please contact me.
In case of daubt please ask me.
Best regards,
Jarek Adamski.
BTW If you want to get from me informations about new projects
directly to your mailbox, just check this in the questionnaire
at the link below:
http://zxq.yarek.pl/
--------------------------------------------------------------
YAMOD.ATBUS
It is a fast IDE multiplexer that allows 8-bits computers to
be connected to 16-bits IDE devices (hard disks, CD readers).
For the Z80 CPU there exists dedicated low level driver called
DSKHNDLR.ATBUS. Since #03 version includes also code for fast
sector reading (INI repeated 512 times).
The fastest reading is possible with a sequence of 512 INI
commands. With 3.5MHz clock this gives over 200kB/s (i.e. 4kB
per ZX Spectrum screen frame), what allows to watch movies of
25 frames per second. (Or monochromatic of 2/3 screen at
50 FPS.)
While normal disk usage (a loop with four INI and a DJNZ)
tipical transfer is 180kB/s for ZX Spectrum, 200kB/s for
Timex FDD, 250kB/s for Sam Coupe - both while reading and
writing.
YAMOD.ATBUS can operate with DMA chip, giving transfer speeds
several times faster.
YAMOD.ATBUS is as standard supported in ZXVGS P3M (interface
PL3MEM or ZX Spectrum +3 upgrade) and optionally can be in
versions: ASP (ZXATASP), MB2 (MB-02), PEN (Pentagon or ZX
512kB with Beta 128 interface). Direct hard disk connection
support (independent from cable server) is in plan for
ZXVGS TMX (interface TI-of-TTL) and UPB (Timex Computer 2068
and Unipolbrit Komputer 2086) versions. Thre's also support
in CPM22QED - versions SAM (Sam Coupe), UPB i YTF (YABUS
upgrade for Timex FDD). CPM22QED UPB can work on Timex
Sinclair 2068 when there's extra 64kB RAM in DOCK (or EXROM).
Please contact me if you would like to use CPM22QED.
Beside IDE multiplexer, YAMOD.ATBUS can also contain real
time clock (exactly like in PC AT) with battery backup. The
RTC has also extra 50 or 114 bytes of non-violate memory
and option of periodic maskable interrupts generation (range
2 to 8192 per second).
Periodic interrupts are required to run on Timex FDD a system
like UZI (with process switching and TCP/IP stack). Those
interrupts allow also to generate sampled sound of maximal
frequency 4kHz (telephone quality). Probably it will be
possible to generate sampled sound while playing movies from
HDD. Periodic interrupts at 512Hz can be used for reading
movements of amiga mouse connected to Kempston Joystick port
(such driver is written for ZXVGS).
The IDE device is available in ports #B0..#B7 (basic ports)
and #BE, #BF (extra ports). RTC is available in ports #B8
(address) and #B9 (data). Direct connection to ZX Spectrum
requires correctly working port disabling with /IORQGE line
(especially in ZX128 and TC2048) and YABUS like interface.
(For decrease costs or limit place YAMOD.ATBUS can be
connected with wires without the DIN 41612 slots.)
The TS2068 doesn't need port disabling, so only e.g. a twister
board and an YABUS.ZX interface are required.
http://www.zx.yarek.pl/interfaces/yabus.zx/
Basic IDE ports can be easy remapped to other addresses,
what can result in compatibilty to IDE interface connected
to MB-02 and the DivIDE (usually this no more requires port
disabling with /IORQGE). Compatibiliyt with interface designed
by Pera Putnik - simulated in RealSpectrum as 48K IDE and
128K IDE - requires also some changes on the board, to reverse
bytes order while writing .
YAMOD.ATBUS is a board of dimensions 100mm × 80mm and with
a slot to YABUS on the short side. All tracks are only at
bottom side, at elements side there are 29 wire connections.
All chips (perhaps except RTC, that can be get from an old
PC AT motherboars) are available at electronics shops. Unlike
most multiplexers projects available from Internet, there's
no need to program GALs. It is not recommended to use sockets,
as the device must work correctly while writing to disk.
YAMOD.ATBUS is available in several assembly versions - from
a ready-to-use device with RTC, through do-it-yourself kit
with minimal set of elements to pure board. Assembly
documentation can be printed from Eagle software. The price
depends from assembly version and is from 5 euro (board) to
15 euro.
The prototipe view (zoomable), schematics, printed circuit
board project and description (however, only in Polish at
this moment) are available at:
http://www.zx.yarek.pl/interfaces/yamod.atbus/
2. Re: YAMOD.ATBUS - fast IDE multiplexer
aralbrec · Sat, 03 Jul 2004 21:03
--- In [email], "Jarek Adamski" <zx@y...> wrote:
Hi Jarek,
I have been following your projects page closely,
particularly the TC2144 upgrade to make the TS2068
compatible with the 128k Spectrum.
In a large project, the one thing I would like
to do is remove the cpu from processing anything
by making all peripherals interrupt driven and
all data movements done by DMA. But not the
Z80-DMA chip which halts the Z80, but custom
DMA channels implemented in a CPLD or FPGA
that can do data movements in the background
without affecting the z80's performance. If
such a project contains a substantial amount
of memory, we have the opportunity to do that,
operating the extended memory much faster than
for a 3.5MHz z80 and assigning access slots
to the z80 so that it can access that RAM
unimpeded, with the rest of the time allowing
DMA channels access to the RAM. This would also
allow the possibility of accessing an IDE device
in something other than PIO mode 0.
Such a project is ambitious but it is something
that would make it challenging (and therefore
fun).
Alvin
3. Nazir Pashtoon
Jack Boatwright · Tue, 6 Jul 2004 13:59:
Hi,
Does anyone know how to contact Nazir? The last email address we
have for him no is longer valid.
Thanks,
Jack
Indexed under
Pico / modern interfaces (UnoDos, etc.) · Hardware projects & new boards