Lol, I though you were going to write a commentary on TI's near-monopoly on educational graphing calculators.

Good luck on your endeavor.  You should probably start with the TI-83 plus family, as it is onfe of the most popular platforms.

Sorry for the semi-necropost.  What exactly is this for?

This code works perfectly in Wabbitemu, but it fails on actual hardware.  By looking at the calculator's display, it seems that the interrupt isn't even triggered at all.  What am I doing wrong?

#include "ti83plus.inc"
.org $9d93
.db t2ByteTok, tAsmCmp

#define dwait in a,($10) \ rla \ jr c,$-3

start:
bcall(_delres)
di

dwait
ld a, 1
out ($10), a

dwait
ld a, 5
out ($10), a
dwait
ld a, $20
out ($10), a
dwait
ld a, $80
out ($10), a


im 2
ld a, $8B
ld i, a

ld hl, $8B00
ld (hl), $8A
ld de, $8B01
ld bc, $FF
ldir

ld hl, intRoutine
ld de, $8A8A
ld bc, intRoutine_End - intRoutine
ldir

ld a, %00001000
out ($03), a

ld a, $44
out ($30), a
ld a, 3
out ($31), a
ld a, 1
out ($32), a
ei

loop:
ld a, $FE
out ($01), a
ld hl, (soundNum) ;do something
nop
nop
in a, ($01)
bit 0, a
jr z, keyDownPress
bit 3, a
jr z, keyUpPress

ld a, $FD
out ($01), a
nop
nop
nop ;how much delay are you supposed to have?
nop
nop
nop
nop
in a, ($01)
bit 6, a
jr z, keyClearPress

jr loop

keyDownPress:
inc hl ;counterintuitive.  I know.
ld (soundNum), hl
jr loop

keyUpPress:
dec hl
ld (soundNum), hl
jr loop

keyClearPress:
di
xor a
out ($31), a
out ($30), a

im 1
ld a, %00001011
out ($03), a
ei
ret ;bye bye


intSound:
ld a, $44
out ($30), a
ld a, 3
out ($31), a
ld a, 1
out ($32), a
call fillScreen

push bc
push hl

.db $21 ;ld hl,**
soundStore:
.dw $0100
dec h
jr nz, intSound_ret
.db $01 ;ld bc,**
soundNum:
.dw $1600
add hl, bc
.db $3E ;ld a,*
intSound_aSave:
.db $00
xor %00000011
ld (intSound_aSave), a
out ($00), a
intSound_ret:
ld (soundStore), hl
pop hl
pop bc
pop af
ei
ret

intRoutine:
push af
;int handling
;jump if port 30 interrupt
jp intSound
intRoutine_End:

fillScreen:
push af
push hl
ld hl, fillScreen_rowCount
inc (hl)
ld a, $C0
cp (hl)
jr nz, fillScreen_Draw
ld a, $80
ld (hl), a
out ($11), a
dwait
fillScreen_Draw
ld a, r
cpl
rlca
ld r, a
out ($11), a
pop hl
pop af
ret
fillScreen_rowCount:
.db 80

Thanks for your help!

---

EDIT1: I need to read port ($04) to acknowledge the interrupt timer.  One would normally do this in a program, so most would not encounter the problem.  it just seems like an odd hardware quirk to me.

---

EDIT2: Okay.  I have now idea what is going on.  I thought the above was a solution, but stuff is still acting weird.

I enjoyed playing along, but now I am disappointed that DJ won't make a Justin Bieber cover album. 

I don't think anybody's said this explicitly yet: The TI-84+ series ASICs do not support the 512 K flash chip TI used for the TI-83+. Specifically, the (easily disabled) circuits that are supposed to censor the cryptographic certificate and prevent the boot code from being modified only support mapping 1, 2, 4, and 8 MB flash chips.

They must have done the same thing as they did when they switched the 84+ series to 48kb of RAM.

I have always been a huge fan of Justin Bieber.  A power metal cover album would be a perfect tribute to the most brilliant, talented pop star of all time.

I can imagine it now.  A beautiful guitar solo to the tune of "Boyfriend".  A dream come true.

Have you tried the calculator with ASM programs or custom apps?  I have a bad feeling that it might ship with a new OS that disables assembly programs.  We all see how TI has gone with the nSpire. 

This is technique is similar to the one I posted in Heylia's thread, but I used an 8.16 counted with the 560Hz interrupt.

I'm not sure is the multiple interrupt routine would work, though.  560, 248, 170, and 118 are all modes of the same hardware interrupt timer.  I don't think you can run then simultaneously. 
However, You can ask Runer112 to implement control of the second hardware timer, which runs up to 1120Hz.