Spectrofon #14 23 августа 1995

Железо

System - an article about compatibility issues and modification of domestic clones ZX Spectrum'a.

           SYSTEM


 Kirill Gromov (CHUNG Software).


   It so happened that my
in fact, "draft" version
article was published in
"ZX-REVIEW" N3 1995, causing maybe, just comments like: "What 
idiot wrote this?". Around the cause of my laziness, because

that in fact complete,
tested, and the corrected version of the article was ready
05/06/1995. Well done! Atonement of guilt oposlya ...



    FORUM ON THE ETERNAL QUESTIONS

  COMPATIBILITY, AND COMPUTERS

      That we choose.


   For this subject led me to
undertake numerous schemes
Port # FF and INTov which to
still treated to the readers / readers in the pages of 
"ZX-REVIEW." 


   Digression: get by just flipping through the recent "Review" 
and stumbled on the section devoted to "Computer Compatibility

ATM TURBO, Alexei (FFC) said: "Why the compatibility?
Was-would be far more appropriate to print: 'Oh noncompliant 
computers ATM TURBO !!!'". 



   From the date of purchase
I have all the filings "ZX-REVIEW," I
watched with undisguised interest
for sections in which raised this issue. For some time I 
noticed that instead of exhaustive answers to the already bored 
the issues raised wave of muddy mess. This

prompted me to take the burden
pervozakryvatelya. All I
write checked me in practice
in three models of computers -
"PROFI +", "PENTAGON 128", "DELTA-C" (old "Moscow"). So:


            PORT # FF

Axiom: When reading from ANY
unencumbered port read ATTRIBUTES alternately
255 (A detailed description
his work is in the old editions REVIEW).

   Should immediately provide that
This applies to the ZX-SPECTRUM 48k;
as I was informed that
SINCLAIR SPECTRUM 128 firms
RESEARCH LTD (!) Port # FF too
there, but work with it several
is different from what I saw and expounded. It turns out that 
since SPECCY produced not only

SRL, but it AMSTRAD and circuitry of them were slightly 
different solutions. A production of a firm's computer can be 
read at m / scheme ULA. These data told me ALEXEY

(FFC COMP.), As well as myself
did not see it, it can only
repeat it. One thing is clear, that probably because of this and
rumors of a "random values ​​with a predominantly attribute the 
current familiarity" to read, this is actually not port.


   I aimed to ensure that I have a good
were using this program
port for its synchronization
Plus it to be the most
easy to implement and more
similar to the original source - ZXSPECTRUM 48k.

The right decision: Almost every
scheme performed on a computer
discrete logic to be found in
site video controller register
charge of the issuance of the attributes
raster on the output stage multiplexers video controller.
This DD56, DD63 in "DELTA"; DD37,
DD40 in the "Pentagon 128", in "Profiles +" DD59 v4.01, DD63 
v5.03. In DELTA dostatotchno bridge the

2,7,10,15 outputs above
chip with a data bus through
Resistors 820 - 1.5k and forget about
all the problems.

   This is slightly more accurate than to say in the pages of 
the implementation of Port REVIEW # FF in computers separate 
fields of memory. But more sensitive (RG ATTR) in question, 
their input OE connected to ground, thus

constantly giving information on
multiplexers. In schemes with a normal vremyadiagrammoy 
("PROS") do this: cut the OE RG ATTR from the "land", and the 
plant it the signal POLE ATTR (neFLD2 in

"PROFI", C5 at the Pentagon 128 ").
After which it exits soedenyaem
(I1-I8 in "Profiles") with the data bus via resistors 820th. 
Sometimes it happens that after that the screen may appear 
undesirable effects, such as: a thin strip of flashing on the 
left screen in the "Pentagon 128".




   Then it is better to put in such
schemes, an additional register for
above mentioned rules of inputs
which is connected to the inputs of RG ATTR, OE to POLE ATTR, 
and outputs through resistors to the bus

data. As you can see a fairly universal approach, and in fact 
most accurately copying approach to the branded car. Exactly

as well as on the company's
"SPECCY" go ARCANOID 1
F.I.R.E. etc.

Encountered "buggy" implementation:

1. The port is designed as a port with address decoding # FF + 
+ neRD neIORQ.



   a) First of all attributes will be
not read by ANYONE unencumbered address.

   b) If there is no clock
from the signal POLE ATTR, then ARCANOID a no go.


   c) From the displacement vremyadiagrammy work programs will 
be flawed (drygane and clicks in

ARCANOID 1).

2. Attributes (right?) Have taken from (eg from register
Buffer RAM).


   May be subject to
all other rules of the port will be
work, but themselves, then attribute
is like as not, but it is already
not as a firm Spectrum,
but it's just like in the book Rodionova - "random values" ...
(Where he got them all? "Do not
understand ...)



   And one more precondition: the existence of resistors 10k
"Pull-up" the data bus to
+5 V and (possibly) correct position and the duration of INTa.
This is, perhaps, would not have hurt any computer, even
without the port # FF.

                          R1

         AT0 D0 RG Q0 D0

         AT1 D1 Q1 D1

         AT2 D2 Q2 D2

         AT3 D3 Q3 D3

         AT4 D4 Q4 D4

         AT5 D5 Q5 D5

         AT6 D6 Q6 D6

         AT7 D7 Q7 D7

                          R8
 CLC RG ATTR / C


   POLE ATTR oOE


RG - 555/1533 IR23 R1-R8 820th

   I think the comments on the scheme
not needed. I note only that all
"Salt" in the "choice" m / s circuit signal POLE ATTR and mix it
outputs with the data bus with
"Installation or". It gives something
that m / scheme is not "clog" the bus
data in a loop pottverzhdeniya
interrupts and does not create a conflict during the call to 
another computer ports.


The scheme will certainly be useful to owners' PENTAGON 128 ". 
Additional register is conveniently soldered on top of the D40, 
of course not napaivaya outputs Q1-Q7 and 1 foot

(Input OE).



   I hope that the "finding"
Register of attributes and controlling
signals, and application tips on setting up the port as a
scheme "profit" or the same setup
additional registers (depending on the scheme of your computer
at the Pentagon 128 "need extra. register) will not make you 
too much difficulty, otherwise better leave this to the 
honorable and right thing to someone else. 



           SIGNAL INT

Axiom: The signal INT is generated on the trailing edge of 
personnel clock and has dlitelmost 9mks.




ATTENTION! Measurement results
held in the company of Spectrum 48k "showed that the duration 
INTa 9mks. and the situation - in early human clock.



   Circuit implementations have
usually two approaches: the formation of
pulse using a rigid logic, the formation by differentiating RC 
chain. Both approach is well covered in the pages of reviews. 
Only You can add that in the first case, you can use the m / 
scheme AG3, although for some reason, put a trigger signal from 
taktuya his "personnel sync" and "clipping" signal through the 
input pulses with R period 9mks who take on

Counter / dividers in the node clock.


   In the second case, the duration
roughly selected capacitor,
and just "brought" resistor.
In both cases, the signal INT may be
attaching a capacitor to shift a little with one foot to ground,
and the other to the input of INT processor. The important 
point is that fact that for some reason, the main indicator of 
"correctness" INTa consider the situation curb strips in some 
games like: VENOM STRICKES BACK, ACADEMY,

MONTECARLO CASINO, etc. Here
should be noted that the "ideal" situation stripes will be 
observed only on machines with separate fields of memory,

reasons, the circuit realization
display controller. In my
practice only on a computer
Delta-C "to" straightened "
INTom were perfectly all of the above programs.

   Adapting the same signal at the INT
other computers with a common field
memory leads to the result
"Nose pulled out - the tail stuck", then
is, if spurred on the strip
top, bottom stripes moved,
and vice-versa. And in "Profiles" creators of the latest boards 
have chosen to maximize performance by reducing almost to a 
minimum WAITov from ULA to the processor. Yes

what to say, even on company
128th behavior above mentioned
different programs (which,
By the way, was written in the pages
REVIEW). Therefore, I discover a criterion of "correctness" 
INTa think no jitter sprites

ARCANOID 2, and the lack of "disappearance" of letters in the 
CHASE HQ 1 and SHOCK MEGADEMO, as well as the same steady work 
a computer keyboard. 

   In this case, INT "PROS"
should be considered quite acceptable.


   In computers, where implemented
mode "TURBO" should be "put out"
INT by signals M1 and IORQ
so because of too long duration source INTa not happen "again" 
his capture in turbo mode. I do not know whether this is good 
or bad, but a "dynamic" INT looks on an oscilloscope is very 
ozadachivayusche ... 


   To my great regret,
Today this topic
can not be considered closed, but you can be considered 
sufficiently elucidated. 



P.S. Those who do not give easy
sleep compatibility (or more correctly, incompatibility) of its
komputera may zvyaknut me, and
I downloaded the suffering whole disk
"Tricky" programs, which
will be judged on the extent
(In) compatibility of the computer.


   Now you can move smoothly to the
Review

COMPUTERS, we choose


   Here perhaps the situation is even starker than ever.
Abundance patterns, lots of peripherals, so
and the sea "bugs" in each of them.
There is a very interesting approach
who showed me
ALEXEI (FFC COMP.):



If you do not have the opportunity to acquire proprietary 
computer, Lift LENINGRAD, redo

the whole scheme of decryption ports
connect the 128kb RAM, assemble and connect the controller
BETA-DISK without PLL and other gadgets with TR-DOS version 
5.03. Oh yeah, and straighten INT postavte

"Rodionovskaya" port # FF. Then
amidst a sea of ​​wires will be something
however, we must say, very
exactly like the brand
SPECTRUM 128. But ARCANOID 1 on
It all did not go. In other
cases, look at how much
you will have the memory (128 - 1024
kb), what additional bells and whistles you will give 
generously to our circuitry in the form of a super-cell size 
and ekstraservismonitorov. 



   But do not forget to pay attention to one small detail:
whether your computer has a sort of tiny buttons, which
it all off. If not,
cool you're out of luck: due to
terrible reluctance of domestic programmers completely 
specified address # 7FFD when working with page of RAM / ROM, 
your computer will give you a lot of "surprises" from the 
persistent refusal work STS2.6, and ending samoformatirovaniem 
discs. Huge thanks to the firm "Condor" for our

happy button! It is
the proud name of the ON / OFF scheme "PROS" and can run on it
99% of hacker intrusions and loaders. Perhaps in the "Scorpion"
It also appears, and then his
owners will not be painful
hurt for aimlessly spent
money.

   A good example of an almost has identified
world company "slots", almost, because
What's next printed circuit board
Album schemes it is perhaps not
matter. To date, most
"Dodelannym" computer with a working periphery and decent
view, is "PROS + v4.01"
Company Condor ". Although it
there is room for improvements in a series of "small but nice" 
somehow the correct port # FF, and other

little things that are so worried
sinkleristov - amateur soldering.
So if you wish, you can
in the process of steadily increasing
compatibility of your computer
at the same time learn how to repair radios, televisions and 
other household appliances, which will undoubtedly

very useful to you in the future.



   Once on the screen "spectrophone" flashed asked one uncle to 
tell him the story 

 ON THE "PROS" and "Condor"

ATTENTION! All of the above / below, this applies to the "PROS 
v4.01", buying a 5 th version of "PROS" I was

just horrified by this nightmarish creations! How can such a
soon make a muck of good
scheme? While I do not correct the "bugs" of this flawed 
creation, Please do not call names indecent

words!


   I myself being a frequent guest of this office, I can tell 
many interesting things. "Profiles", in general, a good 
machine. Y It has many advantages.




   Normal INT, 512-1024KB.
RAM, Extended videorastr, heap all kinds of peripherals, 
"snapped" hardware port extensions (very important for 
"Scorpio" schikov), and much more. Above it all is up and 
running as opposed to promises such as "would-be ..."

for the "Scorpion" and uncollected
board zalipami and peretravami
supercomputer ZEHA NESCHT.


   But, like all post-scoop, and there is nemyslemye idiocy. 
Strange people in the "Condor" evidently set out to do as a 
result of the "PROS" the most inconsistent even with its

previous models of the computer.



And on my timid attempts to at least
The Appendix describes the improvements
computer for a more complete compatibility with the Spectrum I
a loud voice say: "Why
us all "Sinclair"? We
steep orientation on CP / M!!! ".
Even though no brainer that the "PROS" is purchased, both good 
and Powerful Spectrum, but not as wretched "PI-SI scale, as 
would be desirable employees "Condor". A poor

software for such opportunities -
this is indeed blasphemy! Prices
although, in general, and not above the Scorpion-native, but I 
do not always understood, such as, for example, stripping 
20,000 for "turbo mode 8MHz, "consisting of 1 and KP11

two small Conder.



   In general, I will not shit in your
garden, and'll get better advice unfortunate "users"
regarding improvements and maintenance of their vehicles rather 
good, so-as advice on the improvements

related to the Spectrum mode
company generally does not hold.


   If only a small fraction of programmers holders "PROS"
became not find fault miserable sipiemnye
Toys himself CP / M, and would
write COMPUTER PROGRAMS (not
debility purzhischu on BASICe or
PASCALe), then I think, "Prof"
would be beyond any competition.

P.S. Only recently learned how to
cool in the "Scorpion" is made
Port # FF! CDOS-modem with him
gets.



   That just added fuel to
the fire of our eternal disputes
ALEXEY (FFC), and caused a new
explosion of abuse. Apparently this will be
continue for a long time, and yet
if you want you to have everything working, follow my advice.


RESULTS MATTER
RELATED turbirovanie SH-93


   After spending a lot of time searching for information and 
practical studies on this topic, hasten

share the results.


   Let it be known that the clock frequency of 2 MHz is
for SH with the KMT in diameter
203mm. And if while working with the KMT diameter of 133mm at
read operations all sort of Ok,
then the RECORD case badly.

   Without going into detail details say that it is possible to
principle codes change delays
in the firmware TR-DOS, picking up their
a clock frequency of the SH
2MHz, it will do minimal alteration of the controller, but if 
the drive does not "pull" step with increased frequency, you'll 
have all put in place, including the old firmware. On my 
experience is not very happy as' turbirovanie "brake" or as it 
is called "Fifth Armenian" drive, he

Basically briskly humming, but
This little bit wrong ...


   Next time: hardware
refinement. The basic layout of turbirovanie has a very strong 
disadvantage, as described in "S" N12. 


Touch on it in detail: frequency
"Jumps" in time with a gate entry WSTB, causing distortion of 
official information on the disk. Introduction trigger scheme 
described in "S" N12 gives a half-result: there is still a 
possibility of distortion of information in the first front 
gate entry. This "probability" has inoperable during the hours

intensive two experiments
discs. A signal WF / DE in practice generally does not play a
role, check for yourself. Exiting
This situation is very simple
(Suggested by me MIKHAIL Khokhlova, circuitry of "Condor"), you 
must switch on the SH "Normal" frequency up to gate

recording, not during it.



   In this case, the gate
perfectly perform signal DRQ
(38 output VG93) and which should be connected to the 
synchronization input of TM2 instead WSTB. That's when you and

will have no problems with the discs.


   Actually, I'm interested in
this problem by reading the magazine
"SPECTROFON". At the beginning of collected
the first schema - spoiled discs
then the second one - drives become less spoiled, but as we know
better to do nothing than to do badly. Himself had engaged in 
studying this issue. And then it turned out that the switch 
VGshku with increased frequency in the normal need NOT IN TIME 
gate record, as do the authors of the letters, but DO GATE 
RECORDING. 


   Why I opted for
signal DRQ: signal is generated by querying the data bytes for
record for the first time, then VGshka
calculates the service information
after which the signal DRQ is generated a second time. It was 
only after this outstanding record on the gate

drive. Although in my scheme of switching will have to
first signal DRQ. I tried
make sure that does not alter much, "old" scheme turbirovaniya, 
and succeeded in it. Holders of "old" scheme turbirovanie 
trigger-TM2 only need to submit at its input C is not signal 
WSTB, and the signal DRQ. Here then we can confidently 
guarantee the normal operation of the scheme. Owners (?) 
ZX-NEXT, I think that should be taken in its controller is the 
same thing. Below bring an amended scheme:



2MHz
1MHz
____
STEP oS T 1.0 MS Z

              Q 1.1

         D 2.0 a CLC

                   2.1 SH
DRQ / C 3.0 2

              Q 3.1
WF / DE 0 oR 4.0 3

                   4.1
+5 DD1 4

        R1 SE


                  o EZ


                       DD2



DD1 - 555/1533TM2
DD2 - 555/1533KP11A
DRQ - 38n VG93
R1 - 10K
WF / DE - 33n VG93
CLC - 24H VG93
____
STEP - taken from diskovodnogo
computer connector.

By the way, instead of KP11A quite possible to put another 
multiplexer, eg KP12, naturally taking into account its 
connectivity. 



   Dare to say that instead of TM2 is quite possible to do LE1 
organized in It asynchronous RS flip-flop with

direct control inputs and
priority in the scheme on the input from
signal DRQ:


 STEP
(15N VG93) 1 1

                 3

           2 O



           April 1 "1" - 2 MHz
 DRQ 6 "0" - 1 MHz
(38n VG93) 5 O 0>

                  to the switch



P.S. Advise the owners of the "Scorpion" and "ZX-NEXT" s (and 
is there they?) rework their schemes

turbirovanie VG93, throwing
only one wire to WSTB to DRQ.


   And I hasten to say that the scheme with
LE1 not tested, but I think that
it is efficient.


   Very funny to read lines of
how, with the jumpers in
drive "MS-5313" people are making "interoperability" of their
computers kladovskimi elites and STSom. In such cases,
book opens with a technical
description of disk drive, where it receives information about 
the jumpers and what they do. Advertisement floppy drive 
controller in our country for some reason "sygnorili"

signal "READY TO DRIVE" inferred to the interface connector
drive, resulting in and
surprising then absurd behavior drive. That's right: as
I'm sorry, he will write or
read, if not yet ready, but
VGshka does not know about?

   Thus, only through the choice of switching signal
"NGT" and "HEAD" in Paris, you can
"Make friends" flop and programs. Yes
and general description of the "mysteries" of the drive and 
razgadok - entire epic, consisting of 50%

of mental rotation Russian
of national folklore. One of the
such sweet mysteries: the ability to
"Suppress" information on the sealed
disks. The news almost on the verge
Fiction: The controller write to the sealed
DRIVE! In a well-glued
"Serebryashkoy" ... Although not exactly
records, and spoils.


   It turns out that this is not nonsense inflammation of the 
brain, and objectively reality.




   Prompt, how to do it:
standing on one track (for rapid effect) is not paying 
attention to the signal PROTECTION FROM RECORDS "stoically" 
pull " signal "WSTB" more often, the better.



   After some time of your
CD will tell you, "TD-TD" (if
fate has awarded you excellent
floppy drive and a brilliant
controller). Angry boys are already
certainly have rushed to write new
"Megademo with spoiling a sealed
disc simultaneously with the withdrawal of useless greetings. 
And those Who does not want to fall victim to the machinations 
of inquisitive lomschikov carefully read further. 



   Explains the science: it turns out
regardless of the input signal PROTECTION OF RECORDS VG93 yields
Gate entry to the drive. Last dernetsya, the benefit of the 
scheme does not ideal even TEACa, and give

protection signal to the entire RECORD tighter VGshku. However, 
all still suppressed the weak-and-a-Of special and short pulse 
erasure. 


   Now imagine that some wise guy starts to pull
Strobe record on one track (a
even worse in the same sector), with some frequency and 
regularity. Information sooner or late fall. The same way

"Fly" recording auodiokassetah
when you play them on iPods production obscure firms
Taiwan.



   But as tapes are out of luck,
how lucky us! We need only
one element of "I" and the skill with a soldering iron:


                     (Break)
WSTB 30n VG93 0 X 0



                      '
____
WPRT 36n VG93


              DD1 - 555/1533 LI1

   Only after you have collected this scheme, you can
sleep peacefully, without fear that
some bad boy
viciously abused by your
sealed magnitonositelem,
using the cost of circuitry drives.



      COMPUTER POEM


      "Edification # FDetyam"


     Little kids!

     Not for the world,

     In the port # FD is not necessary

     number of send,

     From just such chips

     little kids

     Many computers

     like hang!

     What do you really hard?

     Just a couple of buttons,

     In excess byte address

     full indicate

     After all, the same good word,

     and not a dirty mat

     Your name people

     will be remembered!


   Full address of the port configuration of your computer 
ZX-SPECTRUM 128 # 7FFD, # 7FFD, # 7FFD, # 7FFD!!!!



   I sincerely hope that my
landmark work will be left without
attention.


   And another: "Well really need first-hand information on
AY-3-8910 (12) or analogues
(YM2149F)! And it should
information to be more circuitry character.


          CHUNG Software, Moscow
Kirill Gromov (C) CHUNG Software

                  tel. 397-65-08