Zilog Z8S180 problems - solved

Zilog Z8S180 problems - solved

Some of you may recall that about 3 weeks ago I posted the
following:

-------------------------------------------------------------------
I seem to be unable to access the new registers on the Zilog
Z8S180 CPU, things like the Clock Multiplier register, the
additional DMA register etc. When I try to write to these
registers my writes seem to be ignored and when I read
them back I read FF from most 00 from one. Anyone else
having problems with this chip or know the secret code
to make it work? :) For all the world it's behaving as if it
were a Z180 rather than a S180. I have tried two different
chips, one date coded 0106 and the other 0348 - same
problem with both.
-------------------------------------------------------------------

The solution:
It would appear there are two different mask versions of
the Zilog Z8S180 currently in production - Rev K which
is purely a static CMOS version of the part and otherwise
identical to the Z180 and Rev N which is both a static
CMOS part and has the additional features the datasheet
for the Z8S180 describes. NB. The only way to tell which
part is which is that the Rev K version has "SL1960"
printed on the package whereas Rev N does not; although
the number is not printed as a prefix or suffix to the part
number but on another part of the package in the same
manner as a date code.

Nowhere on Zilog's web site is this information available.
(Although it might be now!) The latest datasheet for the
part (off the website) makes no mention of it. In other
words unless one somehow "knows" this, by telepathy
perhaps, one doesn't have a bloody clue why the damn
chip won't work as the datasheet says it should!

I would have hoped Zilog's tech. support would have
immediately known this was a mask problem and alerted
me to it - however they didn't appear to have a clue.

Interestingly even "Bill A" of Softools (who kindly
responded to my original post) appears to be unawares
of this issue and he's been writing tools for these parts
for years.

In Zilog's defence I will say that once I alerted senior
management to my problem - sparks flew and the issue
was resolved in 24 hours.

Mike Harding

Zilog Z8S180 problems - solved

Some of you may recall that about 3 weeks ago I posted the
following:

-------------------------------------------------------------------
I seem to be unable to access the new registers on the Zilog
Z8S180 CPU, things like the Clock Multiplier register, the
additional DMA register etc. When I try to write to these
registers my writes seem to be ignored and when I read
them back I read FF from most 00 from one. Anyone else
having problems with this chip or know the secret code
to make it work? :) For all the world it's behaving as if it
were a Z180 rather than a S180. I have tried two different
chips, one date coded 0106 and the other 0348 - same
problem with both.
-------------------------------------------------------------------

The solution:
It would appear there are two different mask versions of
the Zilog Z8S180 currently in production - Rev K which
is purely a static CMOS version of the part and otherwise
identical to the Z180 and Rev N which is both a static
CMOS part and has the additional features the datasheet
for the Z8S180 describes. NB. The only way to tell which
part is which is that the Rev K version has "SL1960"
printed on the package whereas Rev N does not; although
the number is not printed as a prefix or suffix to the part
number but on another part of the package in the same
manner as a date code.

Nowhere on Zilog's web site is this information available.
(Although it might be now!) The latest datasheet for the
part (off the website) makes no mention of it. In other
words unless one somehow "knows" this, by telepathy
perhaps, one doesn't have a bloody clue why the damn
chip won't work as the datasheet says it should!

I would have hoped Zilog's tech. support would have
immediately known this was a mask problem and alerted
me to it - however they didn't appear to have a clue.

Interestingly even "Bill A" of Softools (who kindly
responded to my original post) appears to be unawares
of this issue and he's been writing tools for these parts
for years.

In Zilog's defence I will say that once I alerted senior
management to my problem - sparks flew and the issue
was resolved in 24 hours.

Mike Harding

"Mike Harding" <mike_harding@nixspam.fastmail.fm> wrote in message
news:nprgn0p07mhgc0f3a8pngjikvu8c7m60cs@4ax.com...
> Zilog Z8S180 problems - solved
>
> Some of you may recall that about 3 weeks ago I posted the
> following:
>
> -------------------------------------------------------------------
> I seem to be unable to access the new registers on the Zilog
> Z8S180 CPU, things like the Clock Multiplier register, the
> additional DMA register etc. When I try to write to these
> registers my writes seem to be ignored and when I read
> them back I read FF from most 00 from one. Anyone else
> having problems with this chip or know the secret code
> to make it work? :) For all the world it's behaving as if it
> were a Z180 rather than a S180. I have tried two different
> chips, one date coded 0106 and the other 0348 - same
> problem with both.
> -------------------------------------------------------------------
>
> The solution:
> It would appear there are two different mask versions of
> the Zilog Z8S180 currently in production - Rev K which
> is purely a static CMOS version of the part and otherwise
> identical to the Z180 [...]

They aren't indentical in that the static parts did add the clock divider
and power options in the CCR register.

> and Rev N which is both a static
> CMOS part and has the additional features the datasheet
> for the Z8S180 describes.

Rev N also adds problems having to do with the CSIO port which are not in
Rev K.  My understanding is this is why the SL1960 part was retained.  For
some users, N's CSIO just didn't work and they had to have the K part.
These different revisions are available from Digi-Key by part number: [K]
269-3110-ND = Z8S18020VEC1960 & [N] 269-3109-ND = Z8S18020VEC.

I can't find the errata on the N mask for the CSIO.  It was on the site a
few years ago.

> NB. The only way to tell which
> part is which is that the Rev K version has "SL1960"
> printed on the package whereas Rev N does not; although
> the number is not printed as a prefix or suffix to the part
> number but on another part of the package in the same
> manner as a date code.
>
> Nowhere on Zilog's web site is this information available.

There is a small hint in the user's manual that the BRG may be special and
therefore you might suspect it was added, but yes, it's not obvious.  Note
that the clock multiplier register doesn't have an official mnemonic either.

> (Although it might be now!) The latest datasheet for the
> part (off the website) makes no mention of it. In other
> words unless one somehow "knows" this, by telepathy
> perhaps, one doesn't have a bloody clue why the damn
> chip won't work as the datasheet says it should!
>
> I would have hoped Zilog's tech. support would have
> immediately known this was a mask problem and alerted
> me to it - however they didn't appear to have a clue.
>
> Interestingly even "Bill A" of Softools (who kindly
> responded to my original post) appears to be unawares
> of this issue and he's been writing tools for these parts
> for years.

I've known of the differences in the masks and of the "SL1960" part for
several years.  I know of the CSIO problem too in N parts because it has
effected customers, which is how they knew to purchase the SL1960's and why
Digi-Key carries them.  (For a while a customer had one of our "magic"
ICE-Cube ICE's which CSIO worked in a system but his stock processors
didn't - different parts in each, K in the ICE-Cube, N in stock).  It
slipped my mind having been working so long with N parts.  That's no excuse
(and I'm sorry for you I didn't think of it), but this difference should be
easily found on Zilog's WEB site, as should errata notices.

Bill A.
Softools

On Sat, 23 Oct 2004 03:53:11 GMT, "Bill A." <bill_ct@yahoo.com> wrote:

Big snip

>I've known of the differences in the masks and of the "SL1960" part for
>several years.  I know of the CSIO problem too in N parts because it has
>effected customers, which is how they knew to purchase the SL1960's and why
>Digi-Key carries them.  (For a while a customer had one of our "magic"
>ICE-Cube ICE's which CSIO worked in a system but his stock processors
>didn't - different parts in each, K in the ICE-Cube, N in stock).  It
>slipped my mind having been working so long with N parts.  That's no excuse
>(and I'm sorry for you I didn't think of it), but this difference should be
>easily found on Zilog's WEB site, as should errata notices.

Just to clarify Bill; are you saying this info. _is_ available
on the Zilog website or that it _ought_ to be available?

Mike Harding

Hi Mike
just wondering - could idea of "upgrading" to ez80 be interesting to you?
Andrei Kovalev


"Mike Harding" <mike_harding@nixspam.fastmail.fm> wrote in message
news:erpln0hqg27ihh058m5h00pkijg84lqijp@4ax.com...
> On Sat, 23 Oct 2004 03:53:11 GMT, "Bill A." <bill_ct@yahoo.com> wrote:
>
> Big snip
>
> >I've known of the differences in the masks and of the "SL1960" part for
> >several years.  I know of the CSIO problem too in N parts because it has
> >effected customers, which is how they knew to purchase the SL1960's and
why
> >Digi-Key carries them.  (For a while a customer had one of our "magic"
> >ICE-Cube ICE's which CSIO worked in a system but his stock processors
> >didn't - different parts in each, K in the ICE-Cube, N in stock).  It
> >slipped my mind having been working so long with N parts.  That's no
excuse
> >(and I'm sorry for you I didn't think of it), but this difference should
be
> >easily found on Zilog's WEB site, as should errata notices.
>
> Just to clarify Bill; are you saying this info. _is_ available
> on the Zilog website or that it _ought_ to be available?
>
> Mike Harding
>

Mike Harding wrote:
> "Andrei Kovalev" <smithus@servak.de> wrote:
>>
>> just wondering - could idea of "upgrading" to ez80 be interesting
>> to you?
> 
> Thanks for the suggestion Andrei but the Z180 is in a mature
> product with a large code base. I don't think a change to
> the ez80 would be practical.

The Rabbit is the same problem.  If they had just maintained
compatibility with Z80, let alone 64180/z180, code they could have
had a much wider market.

-- 
Chuck F (cbfalconer@yahoo.com) (cbfalconer@worldnet.att.net)
   Available for consulting/temporary embedded and systems.
   <http://cbfalconer.home.att.net>  USE worldnet address!

On Tue, 26 Oct 2004 11:30:53 +0200, "Andrei Kovalev"
<smithus@servak.de> wrote:
>Hi Mike
>just wondering - could idea of "upgrading" to ez80 be interesting to you?

Thanks for the suggestion Andrei but the Z180 is in a mature
product with a large code base. I don't think a change to
the ez80 would be practical.

Mike Harding