128K upgrade for 800XE
Frank Schuster
I like to upgrade my 800XE to 128K as the 130XE.
Since they both have the same board layout, my question is:
Just installing only the missing RAM`s and Freddy-chip is suficient
in order to get it done?
Regards
Frank
--
Posted via Mailgate.ORG Server - http://www.Mailgate.ORG
Andreas Magenheimer
the 800XE already has the Freddy chip. what you need is the Gate Array
chip (GAL), which can only be found on (broken/defective) 130XE machines
or at BEST electronics... -Andreas
Frank Schuster schrieb:
Rick D.
Frank Schuster wrote:
> Hi folks,
>
> I like to upgrade my 800XE to 128K as the 130XE.
> Since they both have the same board layout, my question is:
I found schematics for the 256K upgrade done by Wes Newell at
www.atariworld.com/newell.html. I am sure you can use it to build your
own.
Rick D.
Matthias Reichl
On Fri, 08 Mar 2002 11:08:55 -0600, Rick D. wrote:
> I found schematics for the 256K upgrade done by Wes Newell at
> www.atariworld.com/newell.html. I am sure you can use it to build your
> own.
Unfortunately atariworld.com has closed and you are currently
redirected to 'Holmes Atari 8bit Games'.
I've been able to locate the (old) html file in google's cache, but
the image of the schematics (megxlscm.gif) is missing (of course).
Does someone have a local copy of the schematics he/she could
upload/email/whatever?
so long & thanx,
Hias
machf
I believe I have it, let me look for it among my Atari stuff...
--
__________ ____---____ Marco Antonio Checa Funcke
\_________D /-/---_----' Santiago de Surco, Lima, Peru
_H__/_/ http://machf.tripod.com
'-_____|( http://www.GeoCities.com/Hollywood/2645
remove the "no_me_j." and "sons.of." parts before replying
TXG
To upgrade it's easy when you have the 4 chip version, 2x 4464 chips, add 2
more chips on the 2 places below. Or when you have an 8 chip version, solder
8x 4164 in the right row. Secondly you need to order a PAL chip at Best
Electronics about $10 you can put that into the free space on the right side
of the rams where you woulkd see 3 jumpers in a empty place where you can
put a chip. Remove these jumpers and place the PAL chip there. I just need
to look up the number for the PAL chip. Then the system will act like a
normal 130XE. But the extra chips can also be replaced by 2x 44256 or 41256
and with 1 extra chip and make the system 320Kb....
I will look for the parts later... but it's easy to upgrade...
TXG/MNX
"machf" <no_me_...@hijos.de.terra.com.pe> wrote in message
news:i22l8uoarhhin73t0...@4ax.com...
Matthias Reichl
On Sat, 09 Mar 2002 17:12:53 -0500, machf wrote:
> On Fri, 08 Mar 2002 21:15:19 GMT, usenet-...@hias.horus.com (Matthias
> Reichl) wrote:
> >
> >Does someone have a local copy of the schematics he/she could
> >upload/email/whatever?
>
You don't need to start searching, I've already got the schematics
(thanks to RoyE and Lee Barnes).
so long & thanx for the help,
Hias
machf
Reichl) wrote:
>Hi!
>On Sat, 09 Mar 2002 17:12:53 -0500, machf wrote:
>> On Fri, 08 Mar 2002 21:15:19 GMT, usenet-...@hias.horus.com (Matthias
>> Reichl) wrote:
>> >
>> >Does someone have a local copy of the schematics he/she could
>> >upload/email/whatever?
>>
>> I believe I have it, let me look for it among my Atari stuff...
>
>You don't need to start searching, I've already got the schematics
>(thanks to RoyE and Lee Barnes).
>
>so long & thanx for the help,
>
Well, I didn't find it on my HD, so I guess I already burnt it to a CD...
problem was remembering on which one! Glad to hear you already got it...
Frank Schuster
> I found schematics for the 256K upgrade done by Wes Newell at
> www.atariworld.com/newell.html. I am sure you can use it to build your
> own.
>
> Rick D.
Hi Rick, folks,
I was studying the several schematics on the weg concerning the 320K
upgrade.
The memory chips to use are 41256.
I owe an old mega st board and there are AAA2801P-10 DRAM chips on,
which
are as well 256Kx1 Bit DRAM's.
My question is, can I use these chips for the 320K mod?
Matthias Reichl
On Sun, 17 Mar 2002 22:51:04 +0000 (UTC), Frank Schuster wrote:
>
> I was studying the several schematics on the weg concerning the 320K
> upgrade.
> The memory chips to use are 41256.
>
> I owe an old mega st board and there are AAA2801P-10 DRAM chips on,
> which
> are as well 256Kx1 Bit DRAM's.
>
> My question is, can I use these chips for the 320K mod?
Maybe.
I found the AAA2801P-10 chips in the freetradezone database, but
unfortunately the datasheet is not available for free.
This means you simply have two possibilities if you want to use
your chips:
1) just build the upgrade and hope that it will work, or
2) find the datasheet and compare the pinout and specs to the 41256
Another possibility would be to look for 256k or 1M SIMMs (better
yet SIPPs, which are the same as SIMMs except they have pins
- just like normal ICs - and you don't have to search for a
SIMM socket). A good source for these modules are old 386 (and 386SX)
computers most people are just throwing away these days.
Just be sure the upgrade has built-in support for CAS-before-RAS
refresh (the Newell 1c design has it), so that RAM chips larger
than 256kBit will work.
so long,
Hias
TXG
The pal chip you need is the CO25953 put this in a 65XE or 800XE and it's
becomes a 130XE.
"TXG" <t...@the-underground.myweb.nl> wrote in message
news:3c8aa66a$0$18488$e4fe...@dreader4.news.xs4all.nl...
Rick Cortese
news:slrna9c80h.v3....@camel.hias.horus.com...
It is probably getting to be time to move to a large static RAM to do
everything. There are a few problems with things like refresh on dynamic
RAMs, even SIMMs, that make it a problematic when working with an Atari.
Since a lot of designs use multiple chips the number of connections isn't
much different between say a 256k upgrade and replacing all the RAM in a
system with a 512k static RAM. Price wise it would be about the same too I
think. Last time I priced a 512k static it was about $5-$7.
Matthias Reichl
On Mon, 18 Mar 2002 10:11:34 -0800, Rick Cortese wrote:
> "Matthias Reichl" <usenet-...@hias.horus.com> wrote in message
> news:slrna9c80h.v3....@camel.hias.horus.com...
> >
> > Just be sure the upgrade has built-in support for CAS-before-RAS
> > refresh (the Newell 1c design has it), so that RAM chips larger
> > than 256kBit will work.
>
> It is probably getting to be time to move to a large static RAM to do
> everything. There are a few problems with things like refresh on dynamic
> RAMs, even SIMMs, that make it a problematic when working with an Atari.
> Since a lot of designs use multiple chips the number of connections isn't
> much different between say a 256k upgrade and replacing all the RAM in a
> system with a 512k static RAM. Price wise it would be about the same too I
> think. Last time I priced a 512k static it was about $5-$7.
Ah well, that's a good idea. When I started building and designing
hardware extensions for my 800XL some 15 (or maybe 17?) years ago,
static RAMs (with more than 32k) were really expensive. I didn't even
dare to think about creating a 256k _static_ ramdisk :)
I just had a look into a current price-list of a german company
(www.segor.de) and it lists a 628512LP-07 for 9,80 EUR (32 pin DIP
packaging, SO32 is even 0,90 EUR cheaper). OK, now I believe it,
static RAMs are much cheaper now than back in the good old days.
Another pro for static RAMs would be that several chips (like the
41256 or the LS158 which is used for MUXing the address lines) are
quite hard to get these days.
On the contrary, a lot of people (like me, for example :) have tons
of 1MB SIMMs, taken from old PCs, resting on their shelves. So if
you can re-use one of those SIMMs, building a 1MB extension is just
a matter of a few bucks for some 74LSxx and a few hours of time.
(OK, you don't save a lot of money, so the only remaining advantage
would be the "good feeling" that you were able to re-use an otherwise
useless piece of hardware).
Coming back to the design of a static ramdisk: That could be quite
easy IMO. You could discard most of the circuits used by standard
ramdisks, only cut the "RAM-select"-trace coming from the MMU
and add a small address-decoder (using A14, A15, and PB4, maybe
also HALT and PB5 if you want separate ANTIC access) to select
between buit-in RAM and ramdisk. Connect A0-A13 from your Atari
to the static RAM, and some PB pins to A14 and higher of your
SRAM. Ah, well, and maybe a little bit of logic for Basic and
Selftest if you want 512k or 1MB and still want to control it with
PortB. So far for the rough design. Minor implementation details
(like combining the SRAM-chip-select with PHI2) are left as an
exercise to the reader :-)
BTW: any other ideas or just brainstorming about ramdisk design
would be great - I'm currently thinking about building/designing
a new ramdisk (my old AtariMagazin ramdisk has gone bad).
'til now I've analyzed the Newell 1Meg (rev 1.c) and Mathy's
design. Both designs have their pros and cons. At the moment
IMO the optimum would be a slightly modified Newell ramdisk
(adding a switch to select between a 512k 130XE compatible
ramdisk and a 1Meg ramdisk) extended with Mathy's latch for
basic and selftest (or a simple logic that switches off both
basic and selftest as soon as the ramdisk is selected - that's
mainly what my old 1Meg design was doing).
So, if you've got any ideas about which design you like best/worst,
just tell me about it. Any input would be great!
so long,
Hias
Frank Schuster
In order to get the datasheet of the AAA2801P just search for AAA2802
at freetradezone. There you get the datasheet for the AAA2800 serie.
This pdf-file contains the datasheet for AAA2800 and, what surprise,
AAA2801.
Unfortunately I am not able to tell if these chips are fully compatible
with the 41256, at least they have the same pin-out.
Regards
Frank
"Matthias Reichl" <usenet-...@hias.horus.com> wrote in message
news:slrna9c80h.v3....@camel.hias.horus.com...
> Hi!
>
----
>
> I found the AAA2801P-10 chips in the freetradezone database, but
> unfortunately the datasheet is not available for free.
>
----
Mathy van Nisselroy
> To upgrade it's easy when you have the 4 chip version, 2x 4464 chips, add
2
> more chips on the 2 places below. Or when you have an 8 chip version,
solder
> 8x 4164 in the right row. Secondly you need to order a PAL chip at Best
> Electronics about $10 you can put that into the free space on the right
side
> of the rams where you woulkd see 3 jumpers in a empty place where you can
> put a chip. Remove these jumpers and place the PAL chip there. I just need
> to look up the number for the PAL chip. Then the system will act like a
> normal 130XE. But the extra chips can also be replaced by 2x 44256 or
41256
> and with 1 extra chip and make the system 320Kb....
>
> I will look for the parts later... but it's easy to upgrade...
Use precision sockets for all chips (2 RAMs and 1 CO25953).
The xx256 RAMs are not pin compatible to the xx64 chips.
CU Mathy van Nisselroy
machf
If you're talking about the 4164s and 41256s, well, they only differ in a single
pin. Now, 4464s are another thing... (the "1" after the first "4" in 41xx or the
second "4" in 44xx stand for the way the memory is organized... 4164=64x1-bit,
41256=256x1bit, 4464=64x4bit...)
--
__________ ____---____ Marco Antonio Checa Funcke
\_________D /-/---_----' Santiago de Surco, Lima, Peru
_H__/_/ http://machf.tripod.com
'-_____|(
remove the "no_me_j." and "sons.of." parts before replying
Matthias Reichl
On Tue, 19 Mar 2002 16:18:05 +0000 (UTC), Frank Schuster wrote:
> In order to get the datasheet of the AAA2801P just search for AAA2802
> at freetradezone. There you get the datasheet for the AAA2800 serie.
> This pdf-file contains the datasheet for AAA2800 and, what surprise,
> AAA2801.
>
> Unfortunately I am not able to tell if these chips are fully compatible
> with the 41256, at least they have the same pin-out.
I just had a short look at the datasheets and as far as I can tell
the AAA2801 seems to be compatible with standard 41256 chips.
so long,
Hias
Bob Woolley
ramdisk selection. I have a 1200XL with a 512K battery-backed SRAM that
works just great. Once the files are on the "disk", you can shut off the
system for as long as you like and they'll still be there on the next
power-up. I send PB4, the ramdisk selection bit, into the MMU instead of BE
(Basic Enable) and output the ramdisk chip-select on the BASIC chip-select
line (this line has to be gated with CLK02 at the SRAM). You should be able
to wire the 512K chip into the Basic ROM site (use a little board). Works
great on my 1200XL - should be fine in the 130XE or 800XL. I can supply a
programmed PAL if you need one.
Bob
.... I'd like to know where to get 512Kx8 SRAM for $7... I pay almost $20.
"Matthias Reichl" <usenet-...@hias.horus.com> wrote in message
news:slrna9cm8g.v3....@camel.hias.horus.com...
Mathy van Nisselroy
> At the moment
> IMO the optimum would be a slightly modified Newell ramdisk
> (adding a switch to select between a 512k 130XE compatible
> ramdisk and a 1Meg ramdisk) extended with Mathy's latch for
> basic and selftest (or a simple logic that switches off both
> basic and selftest as soon as the ramdisk is selected - that's
> mainly what my old 1Meg design was doing).
With my design you do not need a switch.
CU Mathy van Nisselroy
Matthias Reichl
On Thu, 2 May 2002 13:06:24 +0200, Mathy van Nisselroy wrote:
>
> With my design you do not need a switch.
I know. But IIRC you mentioned some time ago that you had timing
problems with 800XL and 800XE machines. Another problem would
be getting the CO25953. So my first idea was to merge both designs
(Newell's and your's) since the Newell-design is known to work
with the 800XL. (I still have to find some time to do this, so
at the moment it's only just an idea).
While studying your schematics I stumbled across something strange:
If I got it right, the LS158 is clocked directly by PHI2.
I checked with my 800XL schematics and there the 158 used for
built-in RAM is clocked from a tapped delay-line (driven by PHI2):
PHI2 is sent through a schmitt-trigger (U19/LS14) and then goes
into the delay line (U29). Pin 10 of the delay line is then connected
to pin 1 of the LS158.
Maybe that could be causing the timing problems?
I've managed to loose my copy of the 130XE schematics, so I couldn't
do a double-check what Atari changed when introducing the Freddy
chip...
BTW: If someone could point me to a copy of the 130XE schematics,
I'd be very glad. IIRC I downloaded it some months ago from
somewhere but I didn't bookmark the site :-(
so long,
Hias
Matthias Reichl
On Wed, 1 May 2002 14:52:09 -0700, Bob Woolley wrote:
> You can dump the internal Basic and just re-program the PAL itself to do
> ramdisk selection.
> [...]
Wow, that's a very interesting idea!
But for now i think I'll go with a standard DRAM upgrade - I have to
get rid of all those old 1Meg SIMMs and SIPPs taking away a lot
of space on my shelves :-)
> .... I'd like to know where to get 512Kx8 SRAM for $7... I pay almost $20.
It should be possible to find a store, but it could be quite difficult.
In my case it was more or less pure coincidence when I first heard of
"Segor electronics" in Germany. Most other stores here in Austria or
Germany want to have $20 or even $30 for a single 512K SRAM...
My only idea is: keep on searching...
so long,
Hias
Mathy van Nisselroy
> I know. But IIRC you mentioned some time ago that you had timing
> problems with 800XL and 800XE machines.
The 1MB 800XE has been checked. The weak Phi2 signal is causing the
problems here. (now all I have to do is undo the damage the mail did by
shaking it to much.)
The 1MB 800XL has a different problem. It works without the BlackBox, but
with the blackbox, it stops booting somewhere between when the access LED on
the BB is ON for the second time and when it goes OFF for the second time.
Bob Puff tell me that is when some SIO stuff is handled/loaded.....
> Another problem would be getting the CO25953.
There is a file on my site called co25953-something that describes how to
replace the co25953. I haven't really looked at it yet, but it looks
promising. Although I should say that I don't know that much about GALs and
PALs etc.
And Best Electronics and B&C Computervision are still selling them for $10
and $12.50 respectively. (NOT cheap, I know)
> While studying your schematics I stumbled across something strange:
> If I got it right, the LS158 is clocked directly by PHI2.
>
> I checked with my 800XL schematics and there the 158 used for
> built-in RAM is clocked from a tapped delay-line (driven by PHI2):
> PHI2 is sent through a schmitt-trigger (U19/LS14) and then goes
> into the delay line (U29). Pin 10 of the delay line is then connected
> to pin 1 of the LS158.
>
> Maybe that could be causing the timing problems?
I'll check it out.
> BTW: If someone could point me to a copy of the 130XE schematics,
> I'd be very glad. IIRC I downloaded it some months ago from
> somewhere but I didn't bookmark the site :-(
ABBUC Bauplanservice at http://www.abbuc.de.
CU Mathy van Nisselroy
Mathy van Nisselroy
> While studying your schematics I stumbled across something strange:
> If I got it right, the LS158 is clocked directly by PHI2.
>
> I checked with my 800XL schematics and there the 158 used for
> built-in RAM is clocked from a tapped delay-line (driven by PHI2):
> PHI2 is sent through a schmitt-trigger (U19/LS14) and then goes
> into the delay line (U29). Pin 10 of the delay line is then connected
> to pin 1 of the LS158.
>
> Maybe that could be causing the timing problems?
I check. The LS14 is an inverter. It inverts the signal which is the same
as moving it have a cycle. On cycle on a 1.78 MHz machine takes about 564
nanoseconds. Half a cycle is therefore 282 nanoseconds, add to that the
approx. 10ms delay between a signal enters one of these chips and the time
it leaves these chips and you get a delay of 292 nanoseconds. It then
enters the delay chip which it leaves (according to my schematics)
260nanoseconds later. Add this up and the whole Phase2 clock is moved 552
nanoseconds back in time, of 12 nanoseconds forward. I can't believe that
12 nanoseconds are enough to cause this trouble. BTW it works in teh XEGS
too. (but the XEGS doesn't seem to have week CPU's and buggy GTIA's) If if
you think I'm wrong, please convince me.
CU Mathy van Nisselroy
Matthias Reichl
On Wed, 8 May 2002 13:36:29 +0200, Mathy van Nisselroy wrote:
> Howdy Matthias, folks
>
> > While studying your schematics I stumbled across something strange:
> > If I got it right, the LS158 is clocked directly by PHI2.
> >
> > [...]
> >
> > Maybe that could be causing the timing problems?
>
> I check. The LS14 is an inverter. It inverts the signal which is the same
> as moving it have a cycle. On cycle on a 1.78 MHz machine takes about 564
> nanoseconds. Half a cycle is therefore 282 nanoseconds, add to that the
> approx. 10ms delay between a signal enters one of these chips and the time
> it leaves these chips and you get a delay of 292 nanoseconds. It then
> enters the delay chip which it leaves (according to my schematics)
> 260nanoseconds later. Add this up and the whole Phase2 clock is moved 552
> nanoseconds back in time, of 12 nanoseconds forward. I can't believe that
> 12 nanoseconds are enough to cause this trouble.
Well, it was more or less just an idea. And it's much easier to test
it on a real Atari than it is to calculate the timing delays :-)
Anyways, my basic thought was that due to the shifted timing it
_could_ have been that the address lines are switched before the
minimum RAS/CAS hold time has elapsed. But even for quite slow DRAMs
this should not matter too much - I just checked with some current
datasheets and the hold time is approx. 20-30ns (some older RAMs
had hold times of some 80 ns which I had in mind). OK, even if the
timing would cause problems this should only affect ramdisk-accesses,
not normal use...
OTOH, I would recommend using the correct signal for LS158 switching
anyway, at least for 800XLs. It's just an additional wire from the
ramdisk to the computer and you are on the safe side considering
slight timing differences of various Ataris and ICs.
> If if you think I'm wrong, please convince me.
Well, I guess I can't :-)
IMHO it is more probable that the BlackBox itself is causing the
troubles - maybe due to an increased load at the processor signals,
maybe due to some (too) critical timings, whatever... I don't know
anything about the BlackBox design but without detailed specs and
a logic analyzer (or a scope) it will be hard to fix.
Sorry if I wasn't of too much help.
so long,
Hias