HDD connection standards
ZXNet echo conference «zx.spectrum»
Hello Oleg! Quoting your msg of Wed 01 Apr 98 18:23 to All:
TO> By the way, here in Saratov we are preparing a screw controller
TO> with support for both TP DOS and the screw itself...
TO> Throw in some info?
Go ahead. You can use soap if no one supports you. Will be the ninth in
my collection. Is there even software?
2All:
I was thinking about the subject.
So that it doesn’t work out like with memory: 60% have more than 128k, and
everyone has different schemes. As a result, the new standard did not work out, and those who remained in the minority are not supported in any way by the software and feel like outcasts. These are Atas256 and MikoBest256 and
ZXNext512 and God knows how many forgotten circuits.
With screws it turns out similar. There is no need for new standards!
Everyone rushed to invent their own interfaces and code their own software. Why? Enough subjects have already been invented. IMHO they appeared in the following order:
Profi
ZXNext (by Slot)
Scorpion
Kay
Sprinter
Plus a few more options from different craftsmen.
Isn’t it too much?
Everything has been there for a long time. Including software (but this is not everywhere).
Profi
In CP/M there is a lot of software, but in TRDOS and ISDOS - complete zero?
Scorpion
The interface is autonomous, all drivers are in ROM, accessible through #3D13
and RST 8. You can work through ports, but it is inconvenient.
Works in TRDOS, ISDOS, CP/M, functionally complete. Works only on Scorpios and compatibles. The most common andIt seems to be well supported by the software. But according to rumors he is the slowest. Although (judging by isdos) it may not be the best.
For those inexperienced in hardware/coding, this is the most convenient and
least labor-intensive option to acquire a screw.
I have exactly this one - and everything seems to be OK.
Kay
Soft is only available in isdos. The port addresses are located crazy
and do not allow you to achieve maximum I/O speed.
Sprinter
The fastest for today. The sector is read with two
INIR commands. But it is complex and difficult to reproduce. (?)
ZXNext (Slot)
TRDOS 5.04H was created specifically for this scheme, which allows you to work with virtual floppy disks on a screw.
The port addresses are located very conveniently, the word is read by two
INI commands. The fastest (after the sprinter).
I have everything for it, including the circuit itself and a complete set of utilities and drivers for isdos.
I will somehow describe the logic of its operation. IMHO this is best!
Let's think: either we should produce standards, or use
long-existing and ready-made software for them.
Otherwise it will turn out like this: a bunch of standards, and for each there is
a little bit of software, but in total there will be no standards, no screws,
no software.
****
PS The funny thing is: I know almost all the details about many schemes, except for the one that I have it! Uncle Zonov kept everything so secret that in my opinion he went too far.
I’ve been wanting to for six months nowIt's hard to hang up the CD as a Slave drive, but HOW? Ports are closed/secret. Damn.
I also want to install a flopof PC controller for 1.44. Also in shock.
Maybe the ports have already been declassified?
Jason.
Powered by ScorpionZS256turbo+/SMUC/HDD'IDE130Mb/FDD'800/
/CMOS/Covox/ATkbd/Mouse/CDOS...
>2All:
> I was thinking about the subject.
> So that it doesn’t work out like with memory: 60% have more than 128k, and
>everyone has different schemes. As a result, the new standard did not work out, and those who remained in the minority are not supported in any way by the software and feel like outcasts. These are Atas256 and MikoBest256 and
>ZXNext512 and God knows how many forgotten circuits.
>
> With screws it turns out similar. No need for new standards!
Denis Dmitriev <Denis.D...@p5.f110.n461.z2.fidonet.org> wrote:
>Great, Tarasov!
Hello, hello... You can just Oleg. By the way, I get here at
two addresses as Oleg Tarasov from Fido and as Oleg Tarasov from i-net
...
> TO> there is, it’s just a set of registers and EVERYTHING!!!
>
>WHAT ELSE IS NEEDED?
>Precisely, access ports to the screw.
Yes, and the software to support these ports, sorry, you will write?
Or, sorry, will you sit under the doss? We found a cool system, I'm just dragging...
>Similarly - the disk drive controller - with its own hardware ports.
>Yes, the trdos controller has a ROM, but it is _part_of_the_operating_system_!
> Moreover, it is irreplaceable, because
>it was made a very long time ago. And it’s stupid to repeat the same idea -
Sorry, but who wrote something better lately?...
>tightly attach your file system to the controller.
>The maximum that can be allowed is a bios, which will
>be able to read/write a sector, etc., that is, perform the functions of the screw itself. And
>it is advisable to stuff it into a free part of an existing
>ROM - there will be a big gain in cost.
>And through the bios (or through a loadable driver, which
>will directly access the ports) any OS can sabotage.
You understand what a thing it is, not only you, sorry, you turned out to beso smart... We have implemented this thing... The ROM contains programs for working with the screw, but everything else is loaded from the screw into our RAM and does not take up a single byte in the main memory...
And if you don’t like something, then you can safely fix it
or even write your own OS...
>You are trying to make your own file system. Why?
Brrrrr, look above^
> TO> Ours is different! It's the controller! And it works both with tr
TO> dos,
TO> and by itself...
>
>Working with trdos is no longer the task of the controller, but of a specially
>processed trdos that will understand it.
But this cannot be done... Let tr dos deal with tr dos, otherwise there will be compatibility zero....
>For example, again, I’ll give a controller by Rst7:
>it has 6 microcircuits, i.e. access to the registers of the screw, and at the same time _FULL_ emulation
>trdos, plus the ability to work with the NTFS file system, use it if you want,
if you want, no: the ports are all in sight, accessible to their drivers.
And we also have access to the registers... We need to sculpt the driver itself and
work... I wonder how complete is achieved there emulation
tr dosa without ROM? And I wonder how many wires you had to hang?
>Why didn’t I do the same for myself? And I would do that!
>But it also has #xxF7 ports, which intersect with many things.
Thissays that all other things simply will not receive proper distribution, and the prospect of their
development raises a complete question...
>That's why I want to finally achieve a single standard for a
controller (that is, ports), so that I can just take and assemble all this for myself. (I still won’t have any opportunity to buy your controller.)
Of course, I understand that freebies are above all!!!
>With the best, Denis.
>
Alas, yes...
For this I say goodbye, Oleg.
PS: What is not clear, write... E-MAIL: st...@fannet.ru To me...:+)
TO> But this cannot be done... Let tr dos deal with tr dos, and
TO> then the compatibility will be zero....
The 3-step method is better:
1. trdos - all in itself, and the screw - with its own BIOS;
2. Trdos+ - understands trdos and also screw.
3. something new, for those who no longer need trdos...
With best wishes - Andrey Khristov (ha...@robin.alt.ru)
[Team Horse Aviation] [Team beer, girls & sofa]
[Team Electric Welding Must Die!] [Team Khasansky border guards]
Greetings, Andrey!
є─────────────────────────── ──────────────────────── AK> Better 3-step method: And it’s even better to install an additional ROM so that with one bit in the required port Ivan, with Big Greetings. ZX-Spectrum - #10 years!
║ On the day
╙────────────────
AK> 1. trdos - all in itself, and the screw - with its own BIOS;
AK> 2. Trdos+ - understands trdos and also the screw.
AK> 3. something new, for those who no longer need trdos...
switch the standard TR-DOS ROM to TR-DOS+, in which everything is done only under
screw. IMHO, connecting the screw is not difficult to add a little ROM.
Hi, Alex!
19 Apr 98, Alex Baskakov writes to Andrey Khristov:
AK>> In general, not everything has been taken from trdos yet.
AK>> What problems are there to solve the letters not only A-D, but also up to Z???
AK>> And, starting with E, map onto the screw, cutting it into pieces of 640
AK>> keel. This is if it's in the forehead. And if you tell trdos that the drive, starting from
AK>> E, can have more than 80 cylinders and more than 2 heads, the capacity of
AK>> partitions will increase. You can actually do a broadcast...
AB> Break it off - A, B, C, D not because you wanted to, but because in
AB> all variables and registers of the VG are allocated two bits for the drive number.
Why the hell do you know about the VG about the screw?? The initiative
can be taken away from her earlier... And the variables too. Let the standard ones live, but no one will forbid you to have a couple of your own...
AK>> Who dug into trdos? What is the terrible secret of 128 files??? And what does the 8th
AK>> bit do??? Why not double the directory???
AB> It's not the 8th bit that is to blame, but the 8th sector - there is system information there, so
AB> you can use something like 144 files.
And there are also places after the 8th sector...
And you can also write a file to disk that will contain a directory...
AB> Don't forget! Write!
I will not forget my mother, the party and the government...
AB> that in all VG variables and registers two bits are allocated for
AB> the drive number.
VG93 does not have such a register. VG doesn't even know what drive
it's working with. A separate chip on the controller board is responsible for selecting a drive, and by adding one bit you can double the number of drives. But you can only work with him yourself. TR-DOS
only knows 4 drives.
Best regards
┬ ┬┌┐┬┐ ┐┬┬
╨ґ╨╙ґ╨╙ ╨ ╨
AK> What problems are there in resolving the letters not only A-D, but also up to Z itself???
The problem is that TR-DOS allocates four bytes for disk parameters and cannot
hold more disk drives. In addition, the program often takes the number of the current
disk, selects two bits and sends them to port #FF as the drive number,
so disk E will end up on disk A.
AK> Moreover, starting with E, map it to a screw, cutting it into pieces of 640 kilos each.
AK> This is, if it’s in the forehead.
AK> And if you tell trdos that a drive starting from E can have more than 80
AK> cylinders and more than 2 heads - the capacity of the partitions will increase. Well
AK> the broadcast can be done...
It’s easier to switch drives to A, B, C, D. You give the TR-DOS command to assign the hard drive file "MY_PROG" to disk A. And there are no problems. It turns out as if a floppy disk with the contents of MY_PROG was inserted into drive A. (This is done on
Sprinter, but so far only with RAM-Disks.)
AK> Who delved into trdos? What is the terrible secret of 128 files??? And what does the 8th bit
AK> do??? Why not double the directory???
8th sector of the directory X sixteen directory elements in the sector = 128 files
And TR-DOS uses the eighth bit of the file number somewhere as an indicator that
the file was not found.
Ivan, with Big Greetings.
Correct, I figured, the VGshka allows this, and information about the number of tracks and sectors IMHO is located on the disk itself. So if you write on a disk that there are more than 80 tracks, and sectors too, then IMHO TRDOS should take this at face value and
drive the head to track 100 without problems, because how do disks with 83
tracks work?
Who do I see - Andrey!
A little nervous with excitement Andrey Khristov 20 Apr 98 at 23:16 with difficulty
told out the following words for his old friend Alex Baskakov:
AB> Break off - A, B, C, D not because I wanted to, but because in
AB> all variables and VG registers are allocated two bits for the number drive.
AK> Why the hell does VG know about the screw?? The initiative
AK> can be taken away from her earlier... And the variables too. Let the standard ones live, but no one will forbid you to have your own
AK> or two...
So why do you need Tr Dos at all? You either insert your ROM with the new OS,
or change TR Dos obscenely... But then all this talk is in vain -
there will be no compatibility.
Greetings, Denis!
є─────────────────────────── ──────────────────────── IM>> And it’s even better to install an additional ROM so that with one DD> and when they agree on a standard, everyone will put their own firmware in trdos+, P.S. In Sprinter, you can have at least a hundred different firmwares at the same time. Ivan, with Big Greetings.
║ On
╙────────────────
IM>> bit in the desired port
IM>> switch the standard TR-DOS ROM to TR-DOS+
DD> this is easy to do, but here, because of the bit number and port address
DD>, the third world war will begin ;)
Well, here Sprinter will definitely be neutral. Which bit of which port wins, the one will connect. :)
DD> and in the end it will be needed like a stork needs a midwife :)
Perhaps just the opposite. Everyone will get tired of struggling with their firmware and they
instead of the second one will turn on the standard one.
AK>> take away the initiative... And the variables too. Let the standard
AK>> live, but no one will forbid you to have a couple of your own...
AB> So why do you need Tr Dos at all? You either insert your ROM with a new OS,
AB> or change TR Dos obscenely... But then all this talk is in vain -
AB> there will be no compatibility.
What about “2 in one” with auto-switching???
Hi, Ivan!
20 Apr 98, Ivan Mak writes to Andrey Khristov:
IM> The problem is that TR-DOS allocates four bytes for disk parameters and
IM> cannot hold more disk drives. In addition, the program often takes
IM> the number of the current disk, selects two bits and sends them to port #FF in
IM> as the drive number, so disk E will end up on disk A.
But you can immediately put a plug in so that work with the screw goes through other
subroutines...
And about putting it in #FF - press. I am for #3D13.
IM> It is easier to switch drives to A, B, C, D. You give the command
IM> TR-DOS to assign the hard drive file "MY_PROG" to disk A. And there are no
IM> problems. It turns out as if a floppy disk with the contents of
IM> MY_PROG was inserted into drive A. (This is done on Sprinter, but so far only with RAM-Disks.)
NOT the best option...
AK>> Who delved into trdos? What is the terrible secret of 128 files??? And the 8th
AK>> 80 cylinders and more than 2 heads - the partition capacity will increase. Well
AK>> the broadcast can be done...
SK> Correct, I figured, the VGshka allows this, and information about the number
SK> of tracks and sectors IMHO is located on the disk itself. So if you write on a
SK> disk that there are more than 80 tracks, and sectors too, then IMHO TRDOS should
SK> take this at face value and drive the head to track 100 without
SK> problems, because how do disks with 83 tracks work?
You shouldn’t puzzle VG 100-track drives. Let her rest...
AK> What about “2 in one” with auto switching???
How are you going to do auto switching? Port?
AK> Okay, Root Entry = 128. After all, no one will prohibit you from entering the extension in the form of
AK> files?
Well then, please... But tr dos will not support - I repeat again,
and all this talk is in vain.
Don’t forget! Write!
Lyokha.
AK>> in the form of files?
AB> Well then, please... But trdos will not support - once again
AB> I repeat, and all this talk is in vain.
The idea is not to corrupt the disks beyond recognition, but that some
new trdos will support all the new features, and old disks will also be supported your own
will do...
This is the only way to make a global transition to a different file system.
You can stupidly port CP/M, MSDOS/FAT, RT-11, something else...
But it’s better to choose some kind of mix, just for a specific architecture.
No, I certainly didn’t start a revolution. Anyone who wants can continue to use 5.25"
and 128 files per disk, no one is forced???
IM>> contents of MY_PROG. (This has been done on Sprinter, but so far only with
IM>> RAM-Disks.)
AK> NOT the best option...
I tried making disks from E to Z. Nothing good came of it.
[about 128 files]
AK> Oh! I remembered. Exactly. :((((
AK> It’s still worth expanding the directory using the data zone. Since this
AK> is only needed by screws, the “flop” part of TRDOS should not suffer...
In Sprinter, this problem is completely solved. The screw is written in the MS-DOS system, with
all possible directories, subdirectories, a bunch of files, etc.
Imho, this the best option from the point of view of portability of files from the PC. I took the screw out of the Sprinter, inserted it into the PC, copied the necessary files, at least forty meters, returned the screw and you can work on the whole operation, if the PC is nearby, in five minutes. And with Sprinter floppy disks. directly understands MS-DOS format.
AK> The idea is not to distort the disks beyond recognition, but that some
AK> new trdos will support all the new features, and the old disks will also pass as
AK>...
When changing the ROM, it is very difficult to achieve full compatibility with old programs -
all sorts of turbo loaders, etc. Are you going to use the free space in the ROM, or write a completely new one, leaving the standard entry points?
Don't forget! Write!
Lyokha.
AK>> that some new trdos will support all the new features, and the old
AK>> disks will also pass for their own...
AB> When changing ROMs, it is very difficult to achieve full compatibility with old
AB> programs - all sorts of turbo loaders, etc.
Let the authors strain for this. An API has been invented for them. If they don’t want to, they don’t need to.
AB> Are you going to use
AB> the free space in the ROM, or write a completely new one, leaving
AB> the standard entry points?
Rather a mixture of both. ;)))
AK>> suffer...
IM> In Sprinter, this problem is completely solved. The screw is written in the
IM> MS-DOS system, with all possible directories, subdirectories, a bunch of files, and
IM> etc. Imho, this option is the best in terms of file portability from
IM> PC. I took the screw out of the Sprinter, inserted it into the PC, copied the necessary files,
IM> at least forty meters, put the screw back and you can work. The whole operation,
IM> if the PC is nearby, in five minutes. And with floppy disks it’s even easier. Sprinter directly
IM> understands the MS-DOS format.
In principle it’s great, but you have to touch it to understand.
IM>> regular Spectrum commanders and many
IM>> other programs will not work with it.
AK> Commanders are easy to rewrite if there is a serious reason.
Great.
And for commanders there is no difference how to connect to the new device, either
*"E" or /rmd e
The last command in Sprinter will lead to connecting a RAM-Disk to the current drive. Do you need to return to the drive? Instead of *"A", type /fdd and that's it.
What's the difference for the program? In essence - none. And in TR-DOS this approach is
easier to implement and less buggy with old software.
Ivan, with Big Greetings.
IM> Great.
IM> And for the commanders there is no difference how to connect to the new
IM> device, either *"E" or /rmd e The last command in Sprinter
IM> will lead to connection to the current RAM-Disk drive. Do I need to return
IM> to the drive? Instead of *"A", type /fdd and that's it. What's the difference for the
IM> program? In essence - none. And in TR-DOS this approach is easier
IM> to implement and less buggy with old software.
In principle, I stand for the same thing, but maybe in other ways...