Top Ten Hardware Mods....

[BlackCanyonFX]

Howdy all,
While I realize that the "Top Ten" of anything is always a challenge, I have to wonder what others may think.
So here's a go at my thoughts:

#1 Memory//HRG -
Let's face it, without more the 1-2k memory, the zx really is pretty much an aerodynamic door stop. So I put this at the top of the list. Done right (aka, much like Wilf's desgns) one reaps the benefits of HRG options. With new chip capacities, why not? Nice and neat (and all internal).

#2 Full Size Keyboard-
Not much more to say about this one...........

#3 ULA cooling-
Maybe this a pet peeve of mine. Maybe I'm nuts. I still believe that heat is the big killer of ULA's.
I've ordered a Peltier (sp) cooler from ebay to play with (.5"x1" 2.5vdc 2a). This should be fun...

#4 Monitor Output-
Avoiding the TV out is a deffinate plus and the inverse video mods come into play here as well.

#5 I/O Capability-
The ability to interact with the real world is what makes the ZX so versatile. I have a soft spot for the I2C interface and the Z80 alows for quite a few other choices.

#6 Mass Storage-
A tough one. I'm thinkng of just going with I2C sram (see #5). Bit rate transfers are far better than tape. Of course, others has gone way high tech with flash drives, etc.

#7 Overclocking-
A search of this site and/or the web shows how to replace the standard Z80A (4mHz) with the Z80H (8mHz) and run the full 6.5mHz on your processor (when it 'has the 'time').


#8 Bet ya thought I had all Ten....nope.

I'm going to leave the last three to debate (and all of them for that matter).


Laters....GTB

[siggi]

Some more:

- reset button
- joystick port
- M1-NOT circuit for running MC above 32 K

and instead of overclocking (#7) I recommend Wilf's "NO WAIT"-circuit:
http://www.user.dccnet.com/wrigter/inde ... 81WAIT.htm

Siggi

[BlackCanyonFX]

Good choices, all, Siggi.

I had forgotten about the No Wait mod even though I have already made it. Easy, fast, and simple. A great choice, and another great Wilf design. (Since I can never seem to stand a cold soldering iron, I'm going to go for the 'overclock' as well.)

I think an on/off switch should be in here somewhere too (PWR LED? I like lights...) .

Hmmmm...bet we can think of a few more to add as worthy of Top Ten consideration. Perhaps rated as to usefulness,
ease of implementation, etc.? I would like to link the choices to their forum and web links. Maybe a kind of pseudo
hardware mods FAQ, so to speak. Thoughts, all?

GTB

[siggi]

Ooops, I forgot the most important enhancement:
- GLUE
to avoid ram-pack-wobbeling

Siggi

[BlackCanyonFX]

My God, Siggi!!
GLUE! You have violated the first rule of the universe. Duct tape is the force.
It has a light side and a dark side and holds the universe together. Be careful man.....
GTB

[BlackCanyonFX]

And on a more serious note:
I had to do a redraw of Wilf's 64k mod for Proteus. The sram I chose were a bit different, however the design
remains the same. Wilf's schematic is the reference. The pics on his site displayed weird on my screen.

Here's my rendition:

wilfs64k.jpg

(170.57 KiB) Downloaded 2764 times

The 'socket' (J2,J3) is to piggyback the board into the original RAM socket (thru-pins). J1 contains the 'octopus' lines to
pins soldered to vias on the ZX board. J4 Shadow ROM, J5 write protect, J6 battery.

GTB

[siggi]

Hi GTB

since the /OE and the /CE input of the ram chip are tied together, there might be a timing problem (when using modern and fast ram chips) in Wilf's design. See another thread, where I described some details:

viewtopic.php?t=300#p2620

Siggi

[yerzmyey]

All:


> Let's face it, without more the 1-2k memory, the zx really is pretty much an aerodynamic door stop.
---------
Well, I wouldn't say so harsh, I've seen incredible progs in 1K (including PacMan and Frogger), but You're right I use almost only my ZX81 16K - not less.


> Full Size Keyboard
--------
I love the original ZX81 keyz. But the rubber addition was nice.


> Monitor Output
-----------
Well, yes - that's a problem. I mean, whereas I like the TV output, but year by year all regular TVs will go die.
Then - without a monitor - we will not be able to use our machines.
So the monitor output is unavoidable.


> Mass Storage
----------
There are several, as far as I know. The German ones, as well as Yarek Adamski's ZX81 with C64's diskdrive (works!).


> Overclocking
-------------
Turbo is a great idea as long as it's possible to swith off.


> - reset button
------------
YES.


> - joystick port
------------
Only one joystick would be PERFECT. I don't know what is name of such a device but I can describe it:
You plug the joystick in,
load the game and see: "redefine keys"
You choose this option
and set the "keys" up by moving with joystick, heh. I have such a device for Spectrum.


> I think an on/off switch should be in here somewhere too (PWR LED? I like lights...) .
----------------
Well. Could be. True.


But really: AY chip is THE modification / add-on for me.
or some good emplifier chip on the MIC port, to hear original ZX81 sound loud.

[siggi]

> - joystick port
------------
Only one joystick would be PERFECT. I don't know what is name of such a device but I can describe it:
You plug the joystick in,
load the game and see: "redefine keys"
You choose this option
and set the "keys" up by moving with joystick, heh. I have such a device for Spectrum.

Wilf wrote an article how to build a programmable joystick port for the ZX81. Here it is:

Code: Select all

Newsgroups: comp.sys.sinclair
Von: rig...@cafe.net (wilf rigter)
Datum: 1996/12/28
Betreff: ZX RAMKEY

Seasons Greetings to all Sinclair fans,

I have prepared another "easy to build" hardware project for the ZX81 but
this one can easily be adapted to the SPECTRUM (anyone?).
This is ZX RAMKEY for the ZX81, a 2 chip programmable interface to
translate ATARI joystick contacts or PC printer port ASCII code to ZX81
characters. RAMKEY uses a 6264 RAM chip for storing the joystick keycodes
or a 62256 RAM for storing an ASCII lookup table for an IBM printer port
connection. For preprogrammed applications, a 27C256 or larger EPROM can
be used to provide 8 or more translation tables for joystick emulation
of the most popular ZX81 games as well as an ASCII translation table
for a PC printer port interface.
Unlike other interfaces this one requires NO changes to existing ZX game
software since once programmed, the joystick action is indistinguishable
from the ZX keyboard operation.
In order to assign ZX81 scancodes to joystick positions or ASCII
characters, a translation table must be programmed into the RAMKEY RAM.
The EPROM version must be "burned" with the key scancodes and uses a DIP
switch on the high order address lines to select the required translation
table for each game. A block diagram for RAMKEY is included and the text
explains how it all works. The JSBOOT software is self explanatory:
just load and follow screen instructions.
The ASCIIBOOT software is similarly simple to use but will be covered in
the next installment of this article.

Well, I hope you get a kick out of RAMKEY, if not in the flesh, at least
in the spirit of ZX81 simplicity. Part 2 coming soon!

enjoy wilf

------------------ WARNING SOMEWHAT LONG ARTICLE ------------------------

          ZX RAMKEY - PROGRAMMABLE KEYBOARD INTERFACE
              FOR PC PRINTER PORT OR ATARI JOYSTICK
                    A ZX81 BREADBOARD PROJECT
                          1996 - w.rigter

The RAMKEY port provides a simple method for translating digital inputs
into any combination of ZX81 keys. In the simplest example of a 6264 RAM
version, the 17 possible joystick directions North, East, South, West and
Fire and combinations NE,SE,NW,SW, FN,FE,FS,FW,FNE,FSE,FNW,FSW are
translated by RAMKEY to generate ZX81 keyscan codes on CPU data lines
D0-4 just as if keys were pressed on the keyboard.

RAMKEY operation is similar to the normal ZX81 keyboard hardware.
The normal ZX81 keyboard is a matrix of keys with the A8-15 address lines
scanning the keyboard rows one at a time and each row of 5 keys is read on
the D0-4 lines through the ULA input port at I/O address FE.
RAMKEY can be thought of as multiple keyboards of silicon switches each
consisting of 8 bytes with one or more bits set to zero representing
closed key contacts. Each silicon keyboard is selected by a unique combination
of the 5 or more RAMKEY address lines (ie the joystick contacts) and each
silicon keyboard is scanned one byte at a time with A8-15 in the same
way as the ZX81 keyboard is scanned.

RAMKEY is connected to the rear edge connector but operates in parallel with
the existing keyboard. RAMKEY is scanned at the same I/O address as the
keyboard, so the 74HC32 address decoder enables the RAM CE at address
FE.  Transistor Q1 enables RAM OE when any joystick contact is closed
and turns on an indication LED to indicate RAMKEY is active.
When the write protect switch S1 is closed, the RAM WR line is enabled
at I/O address 1E to program the "silicon keyboards". A second contact
from S1 disables the RAM OE. If the RAMKEY is only used with the PC printer
port the second S1 contact is not needed. The battery backup for RAMKEY is
optional but simply irresistable: A 3V Lithium battery is connected on the
+ side to the +5V line and through a 1N34A germanium diode to VSS (pin 14)
of the RAM chip. A second 1N34A diode is connected between 0V and the VSS
of the RAM chip. Use a 1uf tantalum capacitor from VDD to VSS. Please
note that the diode cathodes (band) are connected to VSS and ofcourse
you must also connect VDD (pin 28) of the RAM chip to +5V. If the whole
scheme seems upside down compared to conventional battery backup designs
just consider this: when the +5V is turned off +5V=0V and all logic outputs,
RAM inputs and pullup resistors are automatically connected to VDD.
(Whose idea was it anyway to connect diodes on the positive side of the
battery for use with negative logic devices?)
By using a 32K byte RAM chip, this design can be also translate 7 bit ASCII
code. When the PC printer port is hooked up to RAMKEY, the seven LPT data
lines together with the A8-A15 CPU address form the RAMKEY address.
The STROBE signal is used to enable the OE line.

Ofcourse pre-programmed EPROMS can be used instead which do not require the
WR line decoding, write protect switches or backup power. In that case just
the EPROM and a 2 diode address decoder can simplify this design to just a
single chip. The anodes go to A0 and IORQ and the cathodes to the EPROM CE.
Finally, a 4.7K pulldown resistor is connected between CE and 0V.

I have included some software here which, although limited to single or
shifted keys, covers 99% of the joystick applications.
The BASIC program steps the user through the programming procedure.
The program prompts the user to enter a shifted or unshifted key and then
to press and hold the desired joystick position corresponding to that key.
Since the RAM powersupply has battery backup the contents are safe as long
as the WRITE protect switch is open. The software includes a two part machine
code routine the first part of which reads the ZX81 key scancode and saves it
in variable A$. The second part transfers A$ to the selected "silicon
keyboard"
while the joystick position is held. If you wish to type in the program, you
have to create a 1 REM line with 39 characters before poking in the MC bytes.

JSBOOT BASIC PROGRAM

1   REM  XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
10  DIM A$(8)
20  PRINT "HOW MANY JOYSTICK POSITIONS?"
30  INPUT A
40  CLS
50  FOR N=1 TO A
60  PRINT AT 0,0;"ENTER KEY ";N
70  IF INKEY$="" THEN GO 70
80  RAND USR 16516
90  IF INKEY$<>"" THEN GOT 90
100 PRINT A$
110 PRINT "HOLD JOYSTICK POSITION FOR 3 SECONDS"
120 PAUSE 180
130 RAND USR 16535
140 GOTO 40

JSBOOT DECIMAL CODE IN 1 REM LINE

16514 - 118,118 ;these make the 1 REM contents invisible
16516 - 1,254,254,42,16,64,17,6,0,25,237,120,119,35,203,0,56,248,201
16535 - 1,246,254,42,16,64,17,6,0,25,126,237,121,35,203,0,56,249,201

        JS SOURCE CODE

SCAN    LD BC,FEFE   ;B=A8-A15 C=KEYBOARD I/O ADDRESS
        LD HL,(4010) ;START OF VARIABLES
        LD DE,06     ;OFFSET TO FIRST ELEMENT OF A$
        ADD HL,DE    ;POINT TO A$
SCAN1   IN A,(C)     ;SCAN KEYBOARD ROW
        LD (HL),A    ;SAVE IN A$
        INC HL       ;NEXT ELEMENT OF A$
        RLC B        ;NEXT SCAN LINE (A8-A15)
        JR C SCAN1   ;LAST ROW?
        RET

WRAM    LD BC,FE1E   ;B=A8-A15 C=KEYBOARD I/O ADDRESS
        LD HL,(4010) ;START OF VARIABLES
        LD DE,06     ;OFFSET TO FIRST ELEMENT OF A$
        ADD HL,DE    ;POINT TO A$
WRAM1   LD A,(HL)    ;GET KEYBOARD ROW DATA FROM A$
        OUT (C),A    ;WRITE TO RAMKEY RAM
        INC HL       ;NEXT ELEMENT OF A$
        RLC B        ;NEXT SCAN LINE (A8-A15)
        JR C WRAM1   ;LAST ROW?
        RET

          RAMKEY JOYSTICK/PC TO ZX KEYBOARD INTERFACE

                  ZX81 EDGE
                  CONNECTOR
    ZX81/TS1000      .                            RAMKEY
                     .
 __________________________      ________________________________
                                             | ADDRESA8-A15   |
                             S
|         |          .     |    |                   |            |
|         /8         .     /8   /7                  /5           /7
|  _______|_______   .     |    |               ____|___     ____|____
| |               |  .     |    |              | ATARI  |   |  D0-7   |
| | ZX81 keyboard |  .     |    |              |Joystick|   |  PC LPT |
| |_______________|  .     |    |              |________|   |_STB_GND_|
|         |          .     |    |                   |          |   |
|         /5         .     |    |                   |__________|  0V
|  _______|_______   .     |    |   +5V                   |
| |               |  .     |    |    |                    |
| | SINCLAIR ULA  |  .     |    |  [10K] x7       +5V     |
| |_______________|  .     |    |    |             |      |
|         |          .     |    |____|            _|_     |
/8        /5         .     |    |           LED  _\ /_    |
|         |  DATA    .   __|____|____ _            |      |
|         |          .  |A0-7 A8-14   |          [1K]   [100]
|         |  ___________|D0-4       OE|____________|c     |
|         | |        .  |     RAM   WR|____  2N3904 \|____|
|         | |        .  |           CE|__  |        /|b   | S1
|         | /5       .  |_____________|  | |       |e      /
|         | |        .    |         |    | |       |      |
|         | |        .    +-{3VBAT}-+    | |      0V     0V
|  _______|/_____    .                   | |
|_|A8-15 D0-4    |   .    ___            | |      +5V
  |            A0|-----10)    \8_________| |       |
  |          IORQ|------9)___ /            |     [10K]     S1
  |              |   .                     |_______|_____/  ___
  |   Z80 CPU    |   .         74HC32                          |
  |              |   .    ___            ___           ___     |
  |            A5|-----13)    \11___/--1)    \3___/--4)    \6__|
  |            A6|-----12)___ /     /--2)___ /    /--5)___ /
  |            A7|__________________|            |
  |            WR|_______________________________|
  |______________|   .

               ---------  END OF ARTICLE ----------

Pins Atari Joystick

Pin Funktion
1 	vor
2 	zurück
3 	links
4 	rechts
5 	NC
6 	Feuer
7 	NC (oder +5V für Elektronik)
8 	GND
9 	NC

You would have to run a small program to define the keys and then start the game using the programmed keys ....

Siggi

[Mustermann]

The RAMKEY of Wilf sounds quite interesting to emulate key presses controlled by a joystick.

The hardware is quite clear to me.(Even if I am planning a different implementation)

Assembler code is also understandable.

But I am confused about the JSBOOT DECIMAL CODE IN 1 REM LINE

Line 16516 is ok.
Line 16535 starts with 1,246,254 which is 01h F6h FEh
To get these binary you need to compile LD BC FEF6h if I am right.
This differs from the source code LD BC FE1Eh

I also do not understand the relative jump in the 16535 line.
56,249 is 38h F9h which will jump back to OUT (C),A not to WRAM1 LD A,(HL) as mentioned in the source code.

Any suggestions?

Max