Teach Scorpio to understand OUT (#FD), A. How?
ZXNet echo conference «hardware.zx»
Hello, All
This is such a stupid question :)
The circuit from ZX-Format for disconnecting port 1ffd is known to everyone, but the question is a little
friend, namely: Is it possible to teach Scorpio to understand OUT (#FD),A, without
disabling 1ffd? Those. so you don't have to press the lock button every time
Scorpovsky port.
Maybe I'm saying something stupid, but still... ;)
Hello, Evgeny Muchkin
Evg> Such a stupid question :)
Evg>
Evg> The circuit from ZX-Format for disconnecting port 1ffd is known to everyone, but the question
Evg> a little different, namely: Is it possible to teach Scorpio to understand OUT
Evg> (#FD),A , without disabling 1ffd ? Those. so you don't have to poke
Evg> each time the Scorpov port blocking button.
Evg>
Evg> Perhaps I’m saying something stupid, but still... ;)
On Scorpa, all ports of the type #XXFD are decrypted using A14, A15 after decrypting the
#FD.
00 - 1ffd
01 - 7ffd
10 - bffd
11 - fffd
when outputting out (#fd), a port is determined by the most significant bits of A. This is where it comes from
I need to dance.
Draw the problem in more detail - where is the conflict?
Hello, SAM style
in some old magazine they suggested decoding the command code and, if it is out
(xx),a - chop off 1ffd. you can also check the argument, but the scheme is much more
it will turn out to be a gimor.
Hello skyther
http://atmturbo.nedopc.com/dev_fd.htm - (the very first picture) there, however,
made for the #FDFD port on ATM-turbo 1, but by rearranging the address lines you can
do it under #1FFD - the very principle of recognition on the bus
commands OUT (N),A (and only it - according to the corresponding data bits and signal M1)
and cutting off the corresponding “extra” address lines for the duration of its execution. When
In this case, when executing “full” output commands to a port of type OUT (C),A, these bits are not
are blocked.
Here is a real opportunity to do without a toggle switch.
Hello, Witch Doctor
> Why didn’t Zonov do this right away?
Probably, at first I simply didn’t think about possible problems, and only then when
I was just lazy when designing the green board. ;)
Hello, Maksagor
Maxagor is exactly what you need! Thank you!
But unfortunately, I won’t be able to reinvent this case for Scorpio on my own.
:(
Someone help me please. The case is for a specialist for 2 minutes after all. :)
Thanks in advance.
Hello, Witch Doctor
> Why didn’t Zonov do this right away?
Quote from the site http://scorpion.ru/spectrum/hard/plataSCT.htm
> The only thing that should be done and is not done on the board
> Turbo+ , this is the ability to block port 1FFDh when accessing
> to other system ports at the shortened address FDh. We won't be here
> dwell for a long time on how to correctly access the port
> having the address 7FFDh, and how programs using will work
> shortened addressing, on branded computers ZX Spectrum +2, +2A,
> +3. Let this remain on the conscience of those of our programmers who
> save bytes where there is not the slightest need for it.
> Because, in our opinion, the user should not suffer from this
> computer, then along with calls for programmers to use
> correct addressing, we want to recommend to those who already have
> Scorpion or is just thinking about purchasing it, collect one of the already
> numerous published port blocking schemes> 1FFDh. First of all, we advise you to refer to the ZX Format magazine
> N1 (UZhELAZOF section) Since there is free space for extensions
> improvements and research left on the Scorpion Turbo+ board
> is enough, then this kind of improvement if detailed
> documentation (see below) will not be difficult for those who want them
> do.
Hello, Evgeny Muchkin
Why didn’t Zonov do this right away?
Hello deathsoft
As I understand it, we need to decide what we are fixing, only 1ffd or all ports with
extended addressing. In the first case, through the scheme mentioned by Maxagor
we pass the recording strobe C (9 leg) to the TM9 1ffd port (DD47), but not through
element 2ILI (LL), and through 2I (LI), despite the fact that its second input is connected to 9
leg TM2 (non-inverse output). In the second case, this element is not needed at all and
connect output 9 of TM2 to leg 6 of DD52 ID7, having previously separated it from +5, or
Accordingly, we connect output 8 of TM2 to leg 5 of DD52 ID7, having previously disconnected
it from the ground - whatever is more convenient for you.
As I understand it, you still need the first option, and for 7ffd, bffd, fffd
The possibility of shortened addressing will remain, and for 1ffd - only two-byte.
Hello, Black_Cat
Please check if I understood and drew everything correctly?
Picture: http://i19.photobucket.com/albums/b175/muchast/fd_blk.gif
Hello, Evgeny Muchkin
Evg> Maxagor is exactly what you need! Thank you!
Evg>
Evg> But unfortunately, I won’t be able to rethink this case on my own
Evg> scorpio. :(
Evg>
Evg> Someone help, please. The case is for a specialist for 2 minutes after all. :)
Evg>
Evg> Thanks in advance.
Please, post here a cut-out piece of the scan of the scorpion diagram, where the decoding is indicated
page port scorpion. I want to see exactly what beats are used there.
After this, drawing the modified diagram is a matter of a couple of minutes...
Hello, Maksagor
Here is the complete diagram (I just don’t know what to cut out ;)). On it instead of DD47
(so on green, apparently) TM9 is numbered D49 (in the lower left corner
located.)
File: scorp_fix1.rar http://zx.pk.ru/attachment.php?attachmentid=4429
Hello, Evgeny Muchkin
Evg> did I understand and draw everything correctly?
You understood everything correctly, but the first option won’t work, because... in
Register data is written by edge, not by level. Therefore it will be right
work with a modified second option - solder on top of the ID7 decoder
one (at the same time, solder only the inputs and power, except for pin 6, which
Apparently you are soldered with a wire). This input sends a signal through the circuit
according to the first option (i.e. through LI1) 15 output of the decoder, from which it comes
strobe for writing to register 1ffd - cut off and connect to a similar leg on
decoder from the second floor.
Hello, Black_Cat
Bla> You understood everything correctly, but the first option won’t work,
Bla> because Data is written to the register by edge, not by level. Therefore
Bla> the modified second option will work correctly - solder
Bla> on top of the ID7 decoder is another one (at the same time, solder only the inputs and
Bla> power supply, except for pin 6, which you have, judging by the corrections on
Bla> circuit is soldered with wire). This input sends a signal through the circuit by
Bla> the first option (i.e. via LI1) 15 output of the decoder, from which
Bla> there is a strobe for writing to register 1ffd - cut off and connect to
Bla> similar to the leg on the decoder from the second floor.
Hmm, you completely confused me :) I'm afraid that I won't dare to make such a scheme
get drunk... :(
Hello, Evgeny Muchkin
Evg> Soldering such a circuit...
What kind of circuitry is that... one more MS, and even that one is on the second floor, there’s no room
It occupies and does not create unnecessary wires.
Hello, Black_Cat
Black_Cat, maybe there is an opportunity to catch on to some address signal
(and not behind CS1FFD), like Masagor in the diagram.
PS; Second; on the floor, my snot will absolutely run all over the board ;) I'm afraid so
do.
Hello, Evgeny Muchkin
As I understand it, the main problem is the use of the command in some programs
OUT (#FD),A for addressing port 7ffd which causes problems with all
xxfd port group. As a result, it turns out that 7ffd should be allowed
work on both full and shortened addresses, and 1ffd (and apparently all
the rest too) - only for full ones. This can only be done by dividing the sample
these ports into two branches - separately for 7ffd, separately for all other xxfd,
which is what I offer you. On the scorpion, in the presence of the specified unused
elements on the board, this is less than three additional. small logic cases are not
organize. I don’t see any problems in soldering ID7 on the second floor. In general
Write, if you have a desire and unused elements, then you can do it.
Hello, The Exploited
The> before there were also many people who liked to address AY by 0b101xxxxx #FD
No problem - you can select 1ffd separately, and all the other 7ffd separately,
bffd, fffd or any other combination - it's not based on the number of elements
will have an impact.
Hello, Black_Cat
Bla> separately for 7ffd, separately for all other xxfd
Previously, there were also many people who liked to address AY at 0b101xxxxx #FD. this is necessary
mainly digital sounds