65802 anyone?

[BitWise]

Time for some experimentation.

[BitWise]

I have found a source for WDC W65C802P-6 processors and have placed an order. I've ordered quite a few so there will be some spares.

[Windfall]

BitWise wrote:

I have found a source for WDC W65C802P-6 processors and have placed an order. I've ordered quite a few so there will be some spares.

Wouldn't mind two of those. Any pointers ?

[Tor]

I would be interested as well.

[rwiker]

Tor wrote:

I would be interested as well.

Me too!

[BitWise]

rwiker wrote:

Tor wrote:

I would be interested as well.

Me too!

Seems my new source over estimated their available stock. I was told they had thousands but in fact its only 14 -- I'd ordered 30.

[Windfall]

BitWise wrote:

Seems my new source over estimated their available stock. I was told they had thousands but in fact its only 14

Yeah ... 1000 ... 14 ... it's all the same, really.

[BitWise]

These arrived today. 16 WDC 65C802s. I'll test them out on my three chip board when I've sorted out the new firmware.

[handyandy]

Nice score Andrew!

Those are a little faster than mine (4 mhz) that run at 3.6 mhz in an apple II clone (laser 128 ex).

Cheers!
Andy

[BigDumbDinosaur]

BitWise wrote:

These arrived today. 16 WDC 65C802s. I'll test them out on my three chip board when I've sorted out the new firmware.

IMG_3106[1].JPG

I wonder if they exhibit the same behavior as the 65C02 when doing indexed write operations.

[barrym95838]

BigDumbDinosaur wrote:

I wonder if they exhibit the same behavior as the 65C02 when doing indexed write operations.

I seem to remember reading somewhere that the 65c802 is more "hardware-compatible" as a plug-in replacement for the NMOS '02 than even the 65c02, which implies that their electronic behaviors are a bit different. My experiments long ago with an '802 in my 64K Apple ][+ running DOS 3.3 went without a hitch, but I didn't try to run any programs with undocumented NMOS op-codes ...

Mike B.

[handyandy]

I did an extensive exploration of the capabilities of the native characteristics of the 65802 back in the 1990s with a custom interpreter for the HyperC development system written in native mode. So, those undocumented NMOS op-codes are documented 65816 opcodes restricted to a 64k address space. The computer I was using was an Apple IIe clone Laser 128ex with 65802 installed. The only issues I had was trying to do any calls to ROM code or ProDOS operating system in native mode with 8 bit registers and zero page direct register and stack in page 1. I would have to always enter in emulation mode saving the native mode state and restoring upon return. My belief is that the stack got corrupted in native mode when opcodes such as TSX and TXS were performed in the ROM code. In emulation mode it was fine.

In older 65802/816 data sheets there was a caveat regarding Apple II systems regarding addressing. Earlier in this thread is an old datasheet from 1991 that says for Apple II systems VDA and VPA should not be used to qualify addresses for disk operations. The 65802 doesn't have these signals available anyway...

Cheers,
Andy

[BigDumbDinosaur]

barrym95838 wrote:

My experiments long ago with an '802 in my 64K Apple ][+ running DOS 3.3 went without a hitch, but I didn't try to run any programs with undocumented NMOS op-codes ...

There aren't any undocumented opcodes in either the 65C802 or 65C816.

[Dr Jefyll]

barrym95838 wrote:

BigDumbDinosaur wrote:

I wonder if they exhibit the same behavior as the 65C02 when doing indexed write operations.

I seem to remember reading somewhere that the 65c802 is more "hardware-compatible" as a plug-in replacement for the NMOS '02 than even the 65c02, which implies that their electronic behaviors are a bit different.

Yes, different. The 'C02 incorporates a mechanism to fix the NMOS problem of a partially formed address appearing on the bus during indexing. (More detail on the problem, the fix and a related issue NOT fixed here.)

AIUI, the fix was removed on the '816 and '802, thus putting them back in the NMOS '02 camp WRT Partially Formed (aka "invalid") Addresses. The '816 has some new signals (VDA and VPA, but mainly VDA) that can prevent trouble. NMOS '02 and the '802 lack the new signals but they and the '816 can instead benefit from some simple decisions about where data and I/O get mapped. There's just one easy rule regarding partially formed addresses. Try to fix things so there's never anything sensitive $100 below the target of an indexed write. (It is the simplest method. But on an existing system you might not have sufficient freedom -- decisions such as where I/O is mapped may not be under your control. Also the rule protects against PFA's only. For a workaround regarding the less common "related issue NOT fixed" see the linked thread.)

[eightbitdoh]

BigDumbDinosaur wrote:

... My experiments long ago with an '802 in my 64K Apple ][+ running DOS 3.3 went without a hitch, but I didn't try to run any programs with undocumented NMOS op-codes ...

Mike B.

Mike,
Interested to try an '802 in my ][+. I was wondering whether these cmos cpus were compatible with the TTL motherboard and whether the speed rating of the cpu would affect success when running at 1MHz.

If it's likely to work, then I'd be interested in an '802 if anyone has one to spare.

Raymond.