Zilog Z84C0020PEG in my CPC not all works.

[McArti0]

Done it.   Zilog Z84C0020PEG in CPC.

Basically it consumes no energy. It is cold. 
According to the documentation, at 4MHz it consumes 20mA or 0.1W.

BASIC works.
keyboard works. 
GOTEK works.
CPM works.
Bomb Jack works.
Sound works

some undocumented instructions Z80 works.
I checked LD A,LY
v%=0:CALL &BFF8,&C912,0,&7DDD,@v%:PRINT v%
246
(LD A,LY : NOP: NOP: LD (DE),A : RET)

Pinball Dream without synchronization and bad noise screen
AY3 only produces noise.

Batman Forever many scenes don't work.
AY3 only produces noise.
The relay to the tape recorder is ticking.

That's all for now.

[Prodatron]

Thanks for reporting!
Very interesting, I would like to hear about more testings.
I wonder if the issues with BG-stuff are undocumented instructions or timing.
(btw I really like your Basic hacks!)

[roudoudou]

most of problems are due to (i guess) OUT (C),0 undocumented opcode which was fixed (or reassigned) in later versions

[Prodatron]

OUT (C),0 is used in all optimized AY routines, so yes, that makes sense with the corrupted sound.

[McArti0]

OUT (c),0 working   but like OUT (c),255  

OUT &BC00,15: OUT &BD00,15: rem CURSOR REGISTER
?INP(&BFFF)
15
CALL &BFFC, &e9dd ,&c9 ,  &71ED,     &BD00, &100
REM             JP(IX)   RET   OUT(c),0    LD BC,&BD00
? INP(&BFFF)
255

Dear Zilog, does 1 CY really save you? 

But SymbOS working 

[andycadley]

Isn't this a reasonably well known issue with CMOS Z80s ( I think I've seen speculation that, since the instruction was unintended, it's just a result of how internal signals value when floating differs between NMOS/CMOS)

There's also a slight difference in LD A,R/LD A,I behaviour according to @arnoldemu 

https://www.cpcwiki.eu/forum/amstrad-cpc-hardware/z80-cpu-nmos-or-cmos/

[McArti0]

correct.
But I think nobody uses cmos.

Realy? 

[Prodatron]

There's also a slight difference in LD A,R/LD A,I behaviour according to @arnoldemu

But SymbOS working 

Thanks, yery good to know!

SymbOS is using LD A,I/LD I,A a lot, so this part seems to be 100% the same for both Z80 versions.

[McArti0]

I will be making a hardware patch in an inelegant version, i.e. shorting D0-7 from high to low. If anyone knows whether the CMOS output can withstand such frequent treatment, please write something.

[andycadley]

There's also a slight difference in LD A,R/LD A,I behaviour according to @arnoldemu

But SymbOS working 

Thanks, yery good to know!

SymbOS is using LD A,I/LD I,A a lot, so this part seems to be 100% the same for both Z80 versions.

According  to that thread it's a bug on the NMOS version only where the P/V flag gets cleared if an interrupt occurs while processing one of those instructions. I imagine caring about that is such an incredibly niche situation that nothing would be affected unless specifically testing for it.

[VincentGR]

Can you also test Relentless? (sound)

[McArti0]

https://www.cpc-power.com/index.php?page=detail&onglet=dumps&num=10072

This?

[VincentGR]

Yes.

https://www.cpc-power.com/index.php?page=detail&onglet=dumps&num=10072

This?

Yes.

[McArti0]

Can you also test Relentless? (sound)

GamePlay work. Sound noise (Intro and Game).

[VincentGR]

Thank you.