SuperSpectrums, GMX,...
ZXNet echo conference «zx.spectrum»
From Serge Zonov → To All 22 May 1997
Hello, ==ALL== !
Our company is constantly conducting research and development in order to expand
the capabilities of our computer (when possible, the entire family of Spectrum
compatible computers). In light of the beginning of the next round of conversations (the first was in the fall of 96) about the developments of the Peters company, I want to briefly talk about our vision of the ways of developing the hardware capabilities of Spectrum-compatible machines. To begin with, I would like to express my opinion about the new development of the Peters company. Let me say right away that
I have not seen it in person, and therefore I will not judge purely technical indicators.
I will dwell only on fundamental issues, more related to the economic and
marketing side.
The Peters company has taken the path of developing and mastering a completely new computer.
Its capabilities will apparently be quite good (meaning graphics, speed, etc.). Unless, of course, all this is translated into a finished serial design, which
is very unlikely, based on purely economic indicators. In my opinion, an attempt to create such a new computer today is doomed to absolute failure. And here
for what reasons:
1. The cost of a computer board with such capabilities with properly
organized production will be about $70-100. Accordingly, the price of a finished computer will be at least $150-200. For comparison costScorpio boards are $50-60 - and a finished computer is $160. That is, a ready-made computer is at least $100 more expensive than just a board. Not every Spectrum user today will be able to buy such a miracle. What will he do with the computer he has now? Just throw it away or donate it, because selling
Spectrum is not easy.
2. At such a not very low price, this computer will be absolutely “bare” from the point of view of software support. How many serious programs have been written
by domestic programmers over the past year or two? The fingers of one or two hands are more than enough to count everything. It’s naive to believe that everyone will rush to write for something they don’t have. It is even more naive to believe that everyone will rush to buy something that is not supported in any way by software, and programs from the old Spectruma are practically unsuitable. Hoping that the company itself will write a large number of programs, its own OS, knowing the capabilities of the Petres company is simply ridiculous. They wrote very correctly here that a good operating system, hard drive support, other
important programs, and their fine-tuning take years at our rate. This is the best case scenario,
but usually many people simply don’t make it to the finish line.
3. The Peters company is absolutely unknown to the average user (potential buyer) on the Russian market, and this alone will not allow it to create a mass computer. Fastchanging this situation in today's conditions is simply
impossible and it's not even about money. If earlier advertising on Spectrum topics could be given in dozens of books, several magazines, etc., dedicated to Spectrum, today it is only ZX Review, and even then it comes out irregularly. And here the situation is such that whoever is known is known, and who
is unknown can only sympathize.
All these points, as well as others, perhaps less significant, force me
once again to express the idea that creating a new computer today is a waste
of time, if you treat it as a commercial project that should
make a profit.
We decided to go a different way. When creating the new Scorpio model, it was decided that all users of old computers should have the opportunity to upgrade their computer to the capabilities of the new one. This has always been the case when
new Scorpio boards appeared, and it will be so this time too. Moreover, this time
the opportunity to upgrade the old Scorpio model to new capabilities will appear
even before production of the new board begins. For this purpose, our
company is now developing and preparing for serial production a board under the working
name:
GMX
(Grafic_Memory_eXtension device for Scorpion 256)
I would like to briefly talk about the technical indicators of thisdevices.
1.Memory expansion up to 2MB.
Among the various existing and possible options, the choice fell on the
option that is organized in Profi. That is, it is done in such a way that simply 256kB
are addressed the same way as it was, through the usual Scorpio standard, and everything
higher is addressed according to the Profi standard. DFFDh port. Gradations for memory expansion:
256 kB, 512 kB, 1 MB, 1256 kB, 2 MB - standard Simm modules. Moreover, you can install from one module with a volume of 256 kB (you get Scorpio memory) to 2 modules of
1 MB - the total memory is 2 MB. We deliberately did not invest in a larger capacity due to the fact that, firstly, it is already becoming quite expensive, secondly, few people need it, and thirdly, when using the
Z80, very large memory is not used effectively. Why was the Profi standard chosen? Why come up with new standards if there are already computers with
extended memory, and there are no technical restrictions on using the same
ports to address this expanded memory.
2. Graphic screen 640*200(208) + hardware scroller.
The number of colors is the same as in a regular ZX. For each byte of graphics - one byte of attributes. In order for the image to be in color, it is necessary to have 2 Simms, i.e. at least 512kB. The addressing of the graphic screen starts
from 4000h, completely linear, i.e. first 80 bytes - first linescreen, next
80 bytes - next line, etc. Screen attributes are stored in one of the
extended memory pages. Their addressing is exactly the same. The main purpose of advanced graphics is a standard text screen of 80*25 characters (a character has a size of 8 by 8), necessary for working with text editors and email programs. To speed up work in text mode, hardware
scrolling up and down is used. All other possible applications of advanced graphics are
considered as secondary, since, in our opinion, nothing worthwhile can be created using the
Z80 to process such graphics. Why did the choice fall on a graphic screen, and not on a purely text one? The reason is that
text mode (with a hardware character generator) is much more complicated in
implementation; ideally, the character generator should be loadable, and this further
complicates the design.
3. Professional ROM 040 (512 kByte).
When developing software for current versions, we have already encountered a limitation on the amount of
RAM. Therefore, further development of the embedded software will only take place with support for the GMX expander. It is planned to convert the shadow monitor to
work in 640*200 mode, Russify it, and further expand
the built-in service capabilities of the shadow monitor.
4. Increased compatibility with ZX Spectrum and Pentagon.
To do thisAn additional menu line Standard
Spectrum is added to the shadow monitor menu. In this case, all additional ports are disabled + at the same time, the standard ZX Spectrum 128+ TRDOS ROM is enabled. In addition, the ability to set temporary parameters for Pentagon-128 will be introduced. When switching to these
modes, the computer turns into an absolutely standard ZX Spectrum or
Pentagon, while completely “forgetting” about its additional capabilities, etc.
Exit this mode using the reset button.
5. Turbocharger with an acceleration factor of 1.8-1.9.
The GMX board has a built-in Turbocharger circuit with an increase in speed of
1.8-1.9 when executing programs from RAM. When working with ROM, performance
increases exactly twice. Switching Turbo/Normal as now: both
software and hardware.
6. Design.
How this expander is constructed. This board is 70*110 in size and is inserted into the socket for the Z80 on any Scorpio model. It's better (more reliable)
if instead of the panel you solder a standard connector. In addition, a cable is soldered to 10 points of the old
board, through which the following signals are sent/removed to the GMX board through a 10-pin connector
:
1-R - red
2-G - green
3-B - blue
4-I - intensity
5-Sync - synchronization
6-Magic - Magic button 7-RDR/ - readingROM
8-CSR/ - access to RAM
9- +5v
10- Earth
7.Economic indicators.
The estimated cost of a GMX board with 512kB is $35-40. For comparison today
the cost of turbocharging an old Scorpio board is $10, installing Prof. ROM040 is $20.
That is, all additional features are available to the user for $5-10. If
you want to save a little more, desolder more than half of the IC from the main board.
The list will be attached :-)
All these features will certainly be available on the new Scorpio board. In addition, there will also be a standard installation of a mouse controller and an IBM keyboard, as well as an IDE hard drive controller. These two devices, in our opinion, should be on every computer today. Moreover, it turns out that if we install the first controller, perhaps out of necessity, the production of 58-key ZX Spectrum keyboards has been completely discontinued, then the second is ordered for almost every complete computer purchased. From my own experience, I can say that anyone who has at least once
worked with a HDD in TR DOS, and there is nothing to say about IS DOS, as a rule, immediately
forgets that there is also a disk drive installed on the computer and remembers it
only when it is necessary to rewrite programs from someone or for someone.
I will be glad and grateful for any wishes, critical thoughts
expressed about what was read in this letter
With best regardsbest wishes, Serge!
From Andrew Fedorov → To All 23 May 1997
I hate you, Chaka!
Where did you get the poisoned crepodge? (c)...
SZ>> 1.Memory expansion up to 2MB.
AM> To All: Who sees the use of this memory? Don’t forget that
AM> memory is page-based (hence quite difficult to access) and
AM> there is no multi-tasking system - the memory is exclusively used by one
AM> program.
AM> Well, some of it will go under the monitor, and what next?
So let something hang there. Some kind of commander, etc. Magos needs
to be improved... And let this memory not be cleaned by reset, at the first
start (or after a program that got there) it drives whatever
you want there from the ROM, and then only work with the RAM... You can keep archives in the ROM, a
unpack them into RAM... Although, this is probably nonsense...
Andrey Fedorov AKA Alien
From Andrew Fedorov → To All 23 May 1997
SZ> 1. The cost of a computer board with such capabilities with properly
SZ> organized production will be about $70-100. Accordingly,
SZ> the price of a finished computer will be at least $150-200. For comparison
For that kind of money you can buy three rubles. And just to add - and A1200...
SZ> I will be glad and grateful for any wishes, critical thoughts
SZ> expressed about what I read in this letter
I would like to connect it to a VGA monitor, especially 640x200...
Andrey Fedorov AKA Alien
From Fedin Pavel → To All 24 May 1997
Hello, Serge Zonov! I remember that on May 22, 97 you took it and launched
in ==ALL== with this:
SZ> Among the various existing and possible options, the choice fell
SZ> on the option that is organized in Profi. That is, it is done in such a way that
SZ> simply 256kB are addressed the same way as it was, through the usual Scorpio
SZ> standard, and everything higher is addressed according to the Profi standard. Port DFFDh.
[bad blocks]
SZ> Why was the Profi standard chosen?
SZ> Why come up with new standards if there are already computers with
SZ> extended memory, and there are no technical restrictions on using the same
SZ> ports to address this extended memory.
DAMN!!!!! Well, the world is finally finished!!!!!!!!!!! Finally we are starting to unite again!!! Why didn’t anyone think of this before???!!! Oh, if only everything had been like this from the very beginning... ZONOV - RULEZ!!!!! Scorpion - RULEZZZZ!!!!!!!
SZ> 2. Graphic screen 640*200(208) + hardware scroller.
That's cool!!!!! Why wasn’t a ready-made solution taken from the same Profi?
Perhaps because of the 6x8 character matrix and the attributes that are inappropriate for it?
SZ> All other possible applications
SZ> of advanced graphics are considered as secondary, since
SZ> nothing worthwhile can be created when using the Z80 to process such graphics, in our
SZ> opinion.
WellWhy? I've been experimenting with this on my Profi, and
I can say that using modern algorithms working with a
stack, you can get pretty decent speeds. Moreover, the structure
of this screen is better than that of Profi...
Now please tell me what layout of the #DFFD port did you use?
Are any other bits involved in it, besides those that control memory?
Such information interests me because of the opportunity to upgrade
Profi. Help please!
bye, with all sorts of Wishes, Fedin Pavel!
From Sasha Shost → To All 26 May 1997
Hello Alex!
AL> In Moscow, I often meet people who want to purchase your
AL> products. What is the best way to buy something like that? I felt ashamed of Zonov when he _proudly_ reported that the boards
they are soldered with _acid_ flux
so keep in mind that as designers/developers they are great guys, as
_manufacturers_ - they push
Sasha [Team OS/2] [Team EDSMO]
From Vyacheslav Mednonogov → To All 26 May 1997
Get msg, Andrew!
24 May 97 00:10 -- Serge Zonov told Andrew MOA about SuperSpectrums, GMX,...:
SZ>>> 1. Memory expansion up to 2MB.
AM>> To All: And who sees the use of this memory?
As soon as there is a sophisticated graphic. mode, no megabytes are enough
(if anyone supports this cause ;)
With warm greetings, Slava.
From Vyacheslav Mednonogov → To All 26 May 1997
AL> leaving for St. Petersburg?
AL> + Origin: -+- Spot -+- (2:5020/659.53)
^^^^^^^^^
Alexey Ivanov with /659.51 is the official representative of the company
"Scorpio" in Moscow. Here they are, the paradoxes of big cities :))
Warm greetings, Slava.
From Dmitry Grigoryev → To All 27 May 1997
Hello, Andrew!
Friday May 23, 1997 06:55, Andrew MOA wrote to Serge Zonov:
[missed...]
SZ>> 1. Memory expansion up to 2MB.
AM> To All: And who sees the use of this memory? Don't forget that
AM> memory is paged (and therefore quite difficult to access)
I would argue here ;) It is not much more difficult to access than continuous memory, it is just
different and you can get used to it and adapt. And learn how to effectively
use this feature. The Spectrum screen also has an unusual structure, but no one calls it inconvenient. Still, I wonder what Clive Sinclair was guided by when introducing page-by-page memory? In general,
you will build into the shadow a memory manager, a RAM disk, which you can
address, for example, with a three-byte address. It's slow, but for most database-type applications it doesn't matter. But it will be problematic to make a text editor normal... probably...
AM> and there is no multitasking system - the memory is exclusively used by one
AM> program. Well, part of it will go under the monitor, and then?
Yes, they will find a use, I think;) They are already finding it, loading systems with all
overlays into it. And the development of a multitasking system will be greatly facilitated -
stuffing programs that are in the background into it.
AM> Andrew
Regards, Dmitry(OLDMAN).
From Andrew MOA → To All 28 May 1997
SZ> look. Already today several application programs have been written that support
SZ> 2MB according to the GMX standard.
Several weather programs do not do it, there is no system that supports it (and
as far as one can judge, few people think about it :)
SZ> Well, there seems to be a lot of space in the ROM,
It’s not a matter of whether there is a lot or a little, but that the architecture itself memory does not allow it to be used effectively: the Z80 only has 64K. Actually, not even
the memory architecture, but the processor itself.
SZ> I would like to hope that there are not so many errors left.
I would like...
Andrew
From Andrew MOA → To All 28 May 1997
On the day of May 26, 97 Sasha Shost had the good fortune to write to Alex Letaev:
SS> I felt ashamed of Zonov when he _proudly_ reported that boards
SS> they are soldered with _acid_ flux
SS> so keep in mind that as designers/developers they are great guys, as
SS> _manufacturers_ - push
I always I suspected that something was wrong at the company. Here it is! It turns out to be flux.
Andrew
P.S. Yes, by the way, it would be nice to enlighten Comrade Zonov about the organization of production. (With soap.)
From Andrew MOA → To All 28 May 1997
VM> As soon as there is a sophisticated graphic. mode, no megabytes
VM> will be enough (if anyone supports this matter ;)
A certain amount, apparently, yes, but the fact that
instead of “learning to use the tool perfectly, we
begin to modernize it” (from some kind of echo), and modernizing
is enough thoughtlessly.
Andrew
From Sergei Novikov → To All 28 May 1997
AF> For that kind of money you can buy a three-ruble note.
With a VGA monitor? To break your eyes...
AF> Add a little more - and A1200...
Well - that's what I'm saying. A1200 is more expensive than Scorp!!! Although, in theory, it should be the other way around. Those. It's not all that expensive. ;)
Hasta la vista, Andrew! See you in Eden!
·∙ Novikov Sergei AKA Spirit/CS [Team ZX ASM] [Team Afiga MD] ∙·
From Vyacheslav Mednonogov → To All 29 May 1997
SS> so keep in mind that as designers/developers they are great guys, as
SS> _manufacturers_ are pushing
From Vyacheslav Mednonogov → To All 30 May 1997
Get msg, Andrew!
AM> the memory architecture itself does not allow it to be used effectively: the
AM> Z80 has only 64K. Actually, not even the memory architecture, but the
AM> processor itself.
A bad tradition was inherited from 128x machines - the ability
to place memory pages in only one window :(
Warm greetings, Slava.
From Vyacheslav Mednonogov → To All 30 May 1997
VM>> will be a sophisticated graphic. mode, no megabytes will be enough
AM> A certain amount, apparently, yes, but the
AM> circumstance is somewhat upsetting that instead of “learning to use the tool perfectly
AM>,
That is, there are Scorpio/Pros/etc, but no one uses even their
resources to the fullest? Are you talking about this?
Not to mention the capabilities of their ROM (which each has its own), It was difficult
to use even the top memory, because (see the latest surveys) the number
of users who have more than 256k has only reached
25-40%. It seems that only in the next year (if the Spectrum is still on the go)
256k may become the standard for the minimum amount of memory (for me, for example, it’s already too tight). 128i ;))
AM> we are starting to modernize it"
Sergei Zonov shared his immediate plans for partial modernization. What I mean is that for the last two years you have not stood still (the “green” board, the mouse/kbd controller, SMUK, minitower, midi, etc., etc.), but have been following the path of “step-by-step” improvements.
However, if you remember Enlight'96, they were talking about the _new_ Scorpion board.
That is. you had similar plans, but something got in the way :(( Apparently, the age-old
questions: “1. Who will buy it? 2. Who will support it?”
1. As applied to the Sprinter, its mostIts strong point is that it is quite capable of providing the necessary
impact on an inexperienced user (sinklerist) with colorful and fast graphics, a good music card and the way it _easily_
shows AVI files (by no means a circuit design solution, no matter how golden
it may be). And the magic numbers 28MHz/4MB sound like heavenly trills :))
2. For a long time I thought about making programs in two versions - for 128
and 256. But, no matter how much I thought, I couldn’t find a use for the extra 128k, except for a
RAM disk :(( And until recently it was easy to do it purely at 256k not
profitable.
With the Sprinter the situation is different: it is quite possible to make two variants of games/programs with regular/sophisticated graphics. Under the _main_ condition, of course
that there will be quite a lot of Sprinters :)
AM> and modernize quite thoughtlessly.
It is clear that f. Scorpio has taken the position: “sink the competitor” :((
Unfortunately, f. Peters has a similar one this year. position: “we can completely handle it ourselves” (although at Inlight they were more positive)
Who will suffer from this - as always, the common people, torn between their long-term attachment to Scorpio and the alluring
opportunities of the Sprinter.
The best option (for the common people :), perhaps, would be closer
cooperation between two companies (such as “client inexchange for circuitry"), but...
However, a good option would be if the Scorpio people suddenly felt
the need to improve the graphics and speed (even at the cost of creating a new board). Everything
the rest of Scorpio seems to be in place :) Well, and so that the graphics standard
quite by chance turned out to be in some places similar to Sprinter's (after all,
Peters'256 was compatible with Scorpy :). ask for an emulator. They say you already have a new one:))
From Andrew Fedorov → To All 30 May 1997
SN> Well, that’s what I’m saying. A1200 costs more than Scorpia!!! Although, in theory, it should be
SN> the other way around. Those. Not everything is so expensive. ;)
What did you say?
Andrey Fedorov AKA Alien
From Sergei Novikov → To All 2 June 1997
SZ>>>> 1. The cost of a computer board with such capabilities, if properly
SZ>>>> organized production, will be about $70-100. Accordingly,
SZ>>>> the price of a finished computer will be at least $150-200. For comparison
From Denis Parinov (812/08.16) → To All 6 June 1997
* Message from gate SpbZxNet<=>FidoNet.
The other day on June 04, 97, at 01:20, Vyacheslav Mednonogov wrote to Serge Zonov:
VM>>> in exchange for circuitry"),
SZ>> I didn’t quite understand what you are proposing to exchange for.
VM> Well, I wanted to say that the most an easy solution (naive IMHO) would be:
you
VM> take Peters' share in terms of distribution throughout the country and St. Petersburg, Peters
VM> shares technical developments with you.
I already strongly doubt that after everything that Zonov said here, the Peterses will
cooperate :((, which is a pity (very sorry).
SZ>> Therefore,
SZ>> if If Sprinter really gets some serious
SZ>> distribution, then no one will stop you from changing the ROM firmware (and
SZ>> and therefore the circuit) on the GMX and the new Scorpion-2000 board so that it
SZ>> is as compatible as possible with Sprinter.
VM> Wow - this is our way! that at least now there is a small hope that Scorpion will turn its
face to the Sprinter. But I
don’t understand why there was a need to criticize the car before this!
SZ>> increasing the capabilities that bring the computer closer to IBM or Amiga,
SZ>> since you can only catch up with them in price.
Well, it depends on what goal you set :)
VM> So, where is it - golden? middle:( All onSpectrum has already
cranked up: and
VM> memory, and external devices, and NGZD. All that remains is to make the graphics more juicy. And
make
VM> more juicy - increase the performance of the machine by an order of magnitude. And to increase
performance
VM> - use more expensive m/circuits. Here comes the increase in price:((
Slava, they got the devices right, but they didn’t think about the price. At Sprintere, all the devices
on
boards, they buy everything in bulk, the components are suitable for both PC and Sprinter, etc.
etc. This is where the price of the computer comes from, and hence the additional
opportunities!
VM> However, time will tell :).
From Andrew MOA → To All 8 June 1997
VM> A bad tradition was inherited from the 128x machines - the ability
VM> to place memory pages in only one window :(
Sometimes the seditious thought comes to me that the 128 machine was, in principle, thought out
a little worse (or was a “transitional” option).
“Improvements”, unfortunately, are still not visible...
Andrew
From Andrew MOA → To All 8 June 1997
VM> It’s clear that f. Scorpio took the position: “sink the competitor” :((
No, I’m more “about us”. :)
VM> The best option (for the common people :), perhaps, would be more
VM> close cooperation between two companies (such as “client in exchange for circuitry”),
VM> but...
The question is not for me. :)
VM> However, a good option would be if the Scorpio people suddenly felt
VM> the need to increase graphics and speed (even at the cost of creating a new
VM> board). Everything else with Scorpio seems to be in place :) Well, and so that
VM> the graphics standard quite by chance turned out to be in some places similar to
VM> Sprinter’s (after all, Peters’256 was compatible with Scorpy :)
VM> Unlike Peters’ people, I don’t think you can’t do it. And
VM> there is still plenty of time - until Sprinter gains momentum, a year will pass (IMHO).
About the same amount of time will be spent on “creating” such a device... I will keep silent about its
possible cost. :)
VM> However, I’m not Comrade Zonov and I don’t see very far ahead :( Maybe
VM> is really just a waste of time?
8-)
VM> By the way, I want to ask for an emulator by the fall. They say you already have a new
VM>, a fancy one :))
Yes. Since 1995.
I wonder why toосени.
Andrew
From Andrew MOA → To All 8 June 1997
DG> I would argue here ;) It is not much more difficult to access than continuous, it is
DG> just different and you can get used to it and adapt.
And it is different because the Z80 does not address more than 64K. This is all well known.
But this “not much” makes it unacceptable for some quite important
applications. More precisely, it would do if these applications existed, but since the memory
is inconvenient... well, and so on in a circle.
DG> The Spectrum screen also has an unusual structure, no one
DG> calls it inconvenient.
By the way, it is quite convenient.
DG> In general, you can build a memory manager into the shadow, RAM disk, which
DG> can be accessed, for example, with a three-byte address. It's slow, but
DG> for most database-type applications it doesn't matter. But
DG> it will be problematic to make a text editor normal...
DG> probably...
Most likely, it will be something like EMS (for the IBM PC) and malloc/free for each
bank separately.
DG> Yes, they will find a use, I think ;) They are already finding it, loading systems
DG> with
DG> all into it overlays. And the development of a multitasking system will be greatly
DG> made easier by cramming programs in the background into it.
The Z80 does not have the means to support such a system. :(
Andrew
From Andrew MOA → To All 9 June 1997
On the day 04 Jun 97 Denis Parinov (812/08.16) had the good fortune to write to Serge Zonov:
DP> Or did I not understand something or do you want to say that they have been making the screw controller
DP> for more than 3 years? I'll explain the smiley: on May 11, the Peters showed their computer, everything
DP> was loaded from a 3.5" disk. On May 31, the Peters' everything worked from a screw and they
DP> completed the work in 20 days!!!
If I _now_ connect the HDD to any computer, it will
take me less than 20 days.
For In 20 days you can write a small driver (by the way, the hdd driver for ISdos
was, in my opinion, written overnight, although before that it took a certain amount of time to get acquainted
with the system and understand the problem).
Andrew