[comp.sys.atari.st] low cost HARDDISK interface

V61@DHDURZ1.BITNET (Ronald Lamprecht) (08/27/88)

Subject: low cost HARDDISK interface

In his article (9 Aug 88 15:07:47 GMT) icc!jlong@blackbird.afit.af.mil
  (Jeffrey K. Long) asked:
>Could someone please post more info on the article in the German magazine that
>told how to hook up a standard XT drive and controller to the ST!  Several
>people in my user's group are very interested in this prospect.  I understand
> that a circuit board and some driver software are all that is needed.  I also
>heard that with some patches the system could now autoboot also.  So what's the
>catch other than the fact that you don't have access to your DMA port?
>Is this thing slower than using a Berkley board and Adaptec controller?
>What magazine did it appear in (Name and Issue) and would anyone who reads
>German be willing to translate the article to English?  Help!!!!!

The low cost solution mentioned was published in the German magazine "c't"
(Verlag Heinz Heise GmbH & CoKG, D-3000 Hannover, W-Germany)
issue 2/88 p126ff . It uses an OMTI 55xx controller and a small interface
build up of 5 `standard' 74xx ICs. This solution occupies the whole DMA port -
so you have no further access to the DMA port. In issue 5/88 p122ff a ROM
patch was published which allows an autoboot.

Based on the idea of using an OMTI controller I developed another small
interface (9 `standard' 74xx ICs) which does not block the DMA port but
provides drivers for all important DMA port signals. It uses also the OMTI
controller 55xx and I have tested it with two controllers each supporting an
OMTI 5527a with 2 ST277R drives (240 MB in total !). But I haven't tested
the controller with Atari or foreign DMA port devices - in fact I don't
believe that it will work with the Atari laser printer (I have heard that
many other DMA port devices have difficulties with this printer or the
software). But I wanted to keep the port clean for further self build
hardware extentions and not for obscure foreign products.
The controller works with no problems since April 88 - the software (not
identical to that published in c't) consists of a GEMDOS driver,formatter &
utilities, an OS9 driver (real multitasking) & software to access GEMDOS
partitions from OS9 and vice versa. 8 partitions per Harddisk, 2 drives per
controller (up to 8 controllers,...) are supported. The speed is primarily
limited by the operating system (even with Fatspeed) - data transfer rate
with an OMTI 5527a : Interleave 1, 512 Bytes/sector, 26 sectors/track,
60 rpm = 800 kByte/s (There is no hardware to slow down the data transfer
rate !). My favorite solution consits of an OMTI 5527a and 1 or 2 ST277R
(each 65 MB), but I have also used an ST238R (32MB) and an ST251 (40 MB)
with an OMTI 5520a.
So the costs for a 60 MB harddisk are:
  selfbuild controller + cabels   c  50 DM
  OMTI 5527a                      c 170 DM
  Seagate ST277R                  c 850 DM
                                  ---------
                                 c 1070 DM c $580
  + Power supply

If there exits an real interest (several requests) I would try to explain
my own solution in detail and to post the sources (c200 k), but I won't
translate the c't solution (I don't know whether they would agree,...)

In a limited amount I would also be able to produce and send boards for
about 20 DM + postage. If there is a real great request I will perhaps
snake mail a copy of the layout (2 sided) to a few persons who oblige to
distribute this copy to all persons interested in it in their region.




Bitnet:  V61@DHDURZ1                               Ronald Lamprecht
UUCP:    ...!unido!DHDURZ1.bitnet!V61              Theoretische Physik
ARPAnet: V61%DHDURZ1.BITNET@CUNYVM.CUNY.EDU       (Heidelberg, West Germany)

avy@yugas.UUCP (Avy MOISE) (08/30/88)

In article <8808270037.AA23210@ucbvax.Berkeley.EDU>, V61@DHDURZ1.BITNET (Ronald Lamprecht) writes:
> Subject: low cost HARDDISK interface
> 
> In his article (9 Aug 88 15:07:47 GMT) icc!jlong@blackbird.afit.af.mil
>   (Jeffrey K. Long) asked:
> >Could someone please post more info on the article in the German magazine that
> >told how to hook up a standard XT drive and controller to the ST!  Several
> >people in my user's group are very interested in this prospect.  I understand
> > that a circuit board and some driver software are all that is needed.  I also
Tyler Ivanco and I (Avygdor Moise) have designed all the hardware
and software necessary to attach any SCSI hard disk (or tape drive) controller
to the Atari ACSI-DMA port. This adapter comes assembled and tested
or as a bare PCB + one PAL. We also have developed a high performance hard disk
interface and formatting software. We make this board available to all
interested hackers for only $150.00 (assembled and tested) or $99.00 PCB.
Both the assembled and unassembled versions come with a comprehensive
installation and programmer's manual (30 pages) and a very high quality
software kit.

If you are interested, contact
	FutureDOS R&D
	1207-120 Torresdale Ave.,
	North York, Ontario,
	Canada. M2R 3N7

	Tel: (416) 736-0321

They distribute our board.

Avygdor Moise
-- 
Avygdor Moise, York University. (C.R.E.S.S.),  Petrie Bldg. Rm 340,
               4700 Keele St., North York, Ontario, Canada. M3J 1P3
UUCP:   ...!yunexus!yugas!avy                      Tel:1 416 736-5359

V61@DHDURZ1.BITNET (Ronald Lamprecht) (09/19/88)

Sorry for the long pause after my first announcement -- but there was a lot
of work to do and I wanted to collect requests.

Due to the number of requests I decided to post some articles about the Atari
DMA port and my hardware solution, because they may be of general interest.
But the software is too large - I will try to store it (as source and as
executables) at a fileserver.

In this article I want to answer some common requests:

Hardware:
  - The  layout can only be photocopied , there exists no machine readable
    version of it.
  - The assembly is very simple - there is no need of an oscilloscope ,
    a voltmeter is sufficient. Nearly all hardware errors can be detected
    and analysed by the software.
  - General remarks to Harddisks: (My experience, my own opinion)
      - there is no harm in using harddisks with short seek rates (28ms)
        instead of `slow' ones (40ms) (28ms are louder than 40ms). But I
        couldn't recognice any increase in access speed - the operating
        system and it's access to harddisk is more important. So I suggest
        to invest your money in harddisk size and not in speed.
      - Someone asked if it would be possible to install a ST157 (3"1/2)
        instead of the floppy in a ST1040: answer: no :
          - harddisks use other mounting screws
          - there won't be enough space to mount the interface with the
            OMTI controller (especially if you plan or have added an 68020,
            memory expansions or SRAMs instead of (E)PROMs)
          - hardisks need a lot of power and the ATARI power supplies are
            weak (HD-Requirement: c 2A on 12V at power-up, c1A at 5V --
            critical at seek operations)

Software:
  - A MINIX driver will be written immediately after I have received the
    ST version. Please send me a note when you received it in Germany.
  - All programs which use legal GEMDOS and BIOS calls will work with my
    driver - only programs that try to access the harddisks directly won't
    work (besides HD format and partion utilities only a virus would try
    to access the harddisk in this manner)
    Until now we haven't detected a single program that doesn't work with
    our harddisk (besides the ATARI hardisk driver & formater). Even all
    harddisk backup programs like Turtle,... work correctly !
    (some programs have troubles with too many partitions, so we allow you
    to undefine and redefine single partitions)
  - The software was developped for high res. monitors - no restrictions for
    the driver, but our HD utilities will not work with low res. colour
    monitors !!!
  - Now all comments and helps are translated to English !
  - The software can also be used in connection with the c't-solution (one
    short source file must be substituted - no multiple DMA port usage)
  - The boot software exits as a disk bootable version as well as a ROM
    bootable version (ROM patch)
  - Foreign utilities for GEMDOS speed increase and 40 folder bug solution
    are necessary:
    FATSPEED and GEMBOOT recommended
      (Thanks to Ulrich Kuebler &  Konrad A. Hahn)


  I think the main reason for our solution is in fact the software:
  - You can make use of 8 partitions with arbitrary drive IDs (you may use
    K: and O: and keep C: for a RAM disk)
  - You can connect several drives, several interfaces
  - You can create, change, delete partitions at any time (the next version
    will even allow you to move,enlarge and shrink partitions)
  - The drive recognition scheme is `intelligent': you are allowed to
    switch off drives and to reconnect them without rebooting your system
  - The driver needs only about 4k of RAM (!)
  - The utilities make full use of GEM
  - ...

I hope that answers the most important questions. Those of you who had
further special questions I will contact within the next days.



Bitnet:  V61@DHDURZ1                               Ronald Lamprecht
UUCP:    ...!unido!DHDURZ1.bitnet!V61              Theoretische Physik
ARPAnet: V61%DHDURZ1.BITNET@CUNYVM.CUNY.EDU       (Heidelberg, West Germany)

jpdres13@usl-pc.usl.edu (John Joubert) (09/22/88)

--------------
I would like to get in touch with you via US-German mail,
could you post your mailing address please?