Re: 8520/21

From: Gerrit Heitsch <gerrit_at_laosinh.s.bawue.de>
Date: Thu, 05 Mar 2015 19:25:48 +0100
Message-ID: <54F89FAC.8060907@laosinh.s.bawue.de>
On 03/05/2015 01:29 AM, Jim Brain wrote:
> On 3/4/2015 1:47 PM, Gerrit Heitsch wrote:
>>
>> The 8520 had a normal timer instead of the RTC the 6526 had. That's
>> about it, you can use a 8520 in a C64 if the software doesn't use the
>> RTC.
> I just continue to struggle with this.
>
> I am not disputing that they are different, I just cannot understand why
> CSG would have taped out a simpler version of the 6526/8521 when they
> already had a die for the 6526. I can understand doing the work needed
> to move the 6526 into HCMOS II, but why not go straight to the 8521, so
> the resulting part could be used for the 64c/128/1571/1581/etc.

Well, the way it looks with MOS means that the lower last number is the 
chip that was done first, just like the 6520 and the 6522.

So they had that 6526 in NMOS and when the Amiga design came around 
('85), they wanted a CIA for it as well but decided against the TOD/RTC 
from the original design, probably wanting the plain counter instead. 
Which resulted in taking the 6526, taking out the TOD/RTC and doing in 
HMOS-II since that was what everything was moving to. Then, when 
everything else was moved to HMOS-II, they couldn't very well leave the 
6526 in NMOS. Result: Take the 8520 and put the TOD/RTC back in, 
resulting in the 8521.

If you look at the R-Numbers, you will see that it took MOS a bit to get 
the 8520 right, the largest revision I know is R4, but the 8521 worked 
right away (R0) with only a single change done later (R1).

Might not be true but makes the most sense to me.


> There has got to be more to this story.  I don't buy that the TOD BCD
> stuff took up that much die real estate, and so I can't believe that the
> removal would have improved yield considerably.

I doubt it has something to do with yield, more with the way designs 
were moved to HMOS-II.

Here's what I figured out over the years:
(MOS was not 100% consistent...)

6xxx  = NMOS
7xxx  = HMOS-I
8xxx  = HMOS-II
5xxx  = CMOS
4xxx  = ?

x5xx  = CPU/IO/Video/Sound
x3xx  = Video/Sound
x7xx  = Gate Array / support logic / logic gates

xx0x  = CPU
xx1x  = CPU
xx2x  = I/O (parallel)
xx3x  = I/O (parallel and RAM/ROM)
xx4x  = Video controller (simple)
xx5x  = I/O (serial) and RAM
xx6x  = Video controller
xx7x  = Video controller
xx8x  = Sound

Speed grades for old NMOS/HMOS ICs:

xxxx  = 1 MHz
xxxxA = 2 MHz
xxxxB = 3 MHz


On some chips a 2 digit number can be found to the right of the datecode.

First digit:

1 = NMOS
2 = HMOS-II
3 = Low power ROM (24xxx)?
4 = CMOS

Second digit: Revision of the Chip (same as R-number when present)

Known Exception: 6526 with '206A' and '216A'. Those are HMOS-II,
R0 and R1. The meaning of '6A' is unknown.




  Gerrit



       Message was sent through the cbm-hackers mailing list
Received on 2015-03-05 19:00:05

Archive generated by hypermail 2.2.0.