[Aquarius] reading the keymap i/o port in assembly

[nitrofurano]

Hi there!

From the information i'm finding at http://www.vdsteenoven.com/aquarius/keyboard.html and http://www.vdsteenoven.com/aquarius/iomap.html , it looks like the 14bit information from the keymap is all stored in AF 16bit register? i can't get the bits between 8 and 13, and the B register seems to have no value at all

I'm really struggling a lot on understanding it...

What would look like in a z80 assembly routine snippet, that you can read all keys pressed? from the $FF i only can read the colscan (A is a 8bit register, i think), so when i press "w", the same happens when i press "t", "i", "e", "y", etc. - How can i distinguish between "w", "t", "i", "e" and "y"?

Thanks!

[catsfolly]

Hi there!

From the information i'm finding at http://www.vdsteenov...s/keyboard.html and http://www.vdsteenov...rius/iomap.html , it looks like the 14bit information from the keymap is all stored in AF 16bit register? i can't get the bits between 8 and 13, and the B register seems to have no value at all

I'm really struggling a lot on understanding it...

What would look like in a z80 assembly routine snippet, that you can read all keys pressed? from the $FF i only can read the colscan (A is a 8bit register, i think), so when i press "w", the same happens when i press "t", "i", "e", "y", etc. - How can i distinguish between "w", "t", "i", "e" and "y"?

Thanks!

 

You have to set the B register to different values, depending on which column you want to scan.

 

 

The number on the far right of the above table is the number that goes in the B register to scan that column.

 

For example, to scan column 1, you would do:

ld bc,$feff ; 11111110 - scanning column 1
in	 a,(c)


To scan column 8,

ld bc,$7fff ; 01111111 - scanning column 8
 in	 a,(c)

The value returned has a 0 bit set for each button in that column that is being pressed.

 

For example, if the "w" key is pressed, the a register will contain 11111101 when column 8 is scanned, but will return 11111111 when any other column is scanned.

 

If the "t" is pressed, the a register will contain the same 11111101 when column 6 is scanned, but will return 11111111 when any other column is scanned.

 

So the way you can tell a "w" from a "t" is based on which column you are scanning, which is based on what number you put into the b register before doing the "in" instruction.

 

Does this help?

 

If you want to see how basic does it, check out "KEYCHK2" in the disassembly at:

http://archive.konte...es/aqromdis.txt

 

 

Catsfolly

[nitrofurano]

thanks, @catsfolly! in the meanwhile i found that http://archive.kontek.net/aqemu.classicgaming.gamespy.com/Files/aqromdis.txt you're talking about! thanks for confirming! - posted my last experiment from Boriel's ZX-Basic Compiler at http://atariage.com/forums/topic/213965-aquarius-stuff-from-boriels-zx-basic-compiler-first-attempt/page__gopid__2785577?do=findComment&comment=2785577 (attachment there, sources and .rom included! )

[mvdsteenoven]

thanks, @catsfolly! in the meanwhile i found that http://archive.konte...es/aqromdis.txt you're talking about! thanks for confirming! - posted my last experiment from Boriel's ZX-Basic Compiler at http://atariage.com/...77#entry2785577 (attachment there, sources and .rom included! )

 

You can also call KEYCHK2 directly. In this case you won't have to translate the row-scan and col-scan yourself. KEYCHK2 will give the ASCII code of the key pressed.

Better yet is to call the KEYCHK function at $1e80 in ROM.

But these ROM functions use the memory area from $3800 to $38ff, so you need to make sure that these have the correct values before calling KEYCHK.

See http://www.vdsteenov...mempointer.html for the memory area with their default values.

 

The KEYCHK function uses (KWADDR $380b), (SCANCNT $380f), (LASTKEY $380E)

The first thing the KEYCHK function does is an EXX (exchange) of all registers.

If you call the KEYCHK function in ROM you should write something like

 

AQUARIUS_KEYCHK:

XOR A	 ; Set A to zero
LD ($380B), A ; KWADDR
LD ($380C), A
LD ($380E), A ; LASTKEY
LD ($380F), A ; SCANCNT
CALL $1e80 ; Call Keychck
 ; A holds result
RET

 

Regs,

Martin

[nitrofurano]

added aquariuskeychk.bas as library on Boriel's ZX-Basic Compiler - weird that is not working... :S (result would appear in the center of the upper edge of the display) - sad, because i was also curious about how random numbers from bios works as well... :S

aquarius_borielszxbasiccompiler_example01i_201307031201.zip

Edited July 3, 2013 by nitrofurano

[catsfolly]

added aquariuskeychk.bas as library on Boriel's ZX-Basic Compiler - weird that is not working... :S (result would appear in the center of the upper edge of the display) - sad, because i was also curious about how random numbers from bios works as well... :S

I tried it in virtual Aquarius and I could move the ascii-space guy up, down, left and right using the keyboard. What part is not working?

[nitrofurano]

I tried it in virtual Aquarius and I could move the ascii-space guy up, down, left and right using the keyboard.

3 ascii-space guys! you move one with "wsad", other with "tgfh", and other with "ikjl"

 

What part is not working?

far above - there you see an attribute byte (cell) for all key rows (or columns) from the keyboard, a sequence of 8 attribute bytes for each row (or column); and the next one is that part is not working, it should be an ascii character of the key pressed (using that bios call as suggested from @mvdsteenoven ) -

http://img853.images...3/5724/i7fw.png (that one i marked with a red arrow)

Edited July 3, 2013 by nitrofurano

[catsfolly]

3 ascii-space guys! you move one with "wsad", other with "tgfh", and other with "ikjl"

 

 

far above - there you see an attribute byte (cell) for all key rows (or columns) from the keyboard, a sequence of 8 attribute bytes for each row (or column); and the next one is that part is not working, it should be an ascii character of the key pressed (using that bios call as suggested from @mvdsteenoven ) -

http://img853.images...3/5724/i7fw.png (that one i marked with a red arrow)

 

Is that part drawn by this code?

 

 

poke $3412,aquariuskeychk()

poke $3412,$08

 

 

I'm guessing that it should be:

 

 

poke $3012,aquariuskeychk()

poke $3412,$08

 

Otherwise, the second poke writes over the first poke.

 

Catsfolly

[nitrofurano]

Is that part drawn by this code?

 

poke $3412,aquariuskeychk()

poke $3412,$08

 

I'm guessing that it should be:

 

poke $3012,aquariuskeychk()

poke $3412,$08

 

Otherwise, the second poke writes over the first poke.

 

Catsfolly

 

thanks, and sorry about my dyslexia! - btw, i would imagine that would fix, but now it is showing "£" there instead... :S

aquarius_borielszxbasiccompiler_example01i_201307041547.zip

[catsfolly]

thanks, and sorry about my dyslexia! - btw, i would imagine that would fix, but now it is showing "£" there instead... :S

 

The Keychk routine is trying to debounce the keyboard. Basically Keychk is called every frame, and only returns a value if it detects the same key down 4 times in a row. To do this, Keychk saves the value of the pressed key in "LASTKEY", and it keeps a count of the number of times in a row the key was down by incrementing the variable "SCANCOUNT".

 

So, if we clear these variables every time we call Keychk then it will always return 0.

 

AQUARIUS_KEYCHK:

 

XOR A ; Set A to zero

LD ($380B), A ; KWADDR

LD ($380C), A

LD ($380E), A ; LASTKEY

LD ($380F), A ; SCANCNT

CALL $1e80 ; Call Keychck

; A holds result

RET

I guess we have to clear the variables one time as part of some "initialization sequence", and then just call Keychck when we want to check for a key.

 

AQUARIUS_INIT:

 

XOR A ; Set A to zero

LD ($380B), A ; KWADDR

LD ($380C), A

LD ($380E), A ; LASTKEY

LD ($380F), A ; SCANCNT

 

RET

 

 

AQUARIUS_KEYCHK:

CALL $1e80 ; Call Keychck

; A holds result

RET

 

Maybe that will work.

 

Catsfolly