atari 7800 compatibility fix
John Soper
in revisiting a fix I first made a couple years ago.
Thought I'd write it up as a new post so everyone would
see it.
An expansion port 7800 has the following circuit:
_____
11 | | 10 10 |}
D2-----|D Q|-----------| }
(6502) | | | } 8
| | 9 | }------A12 (cart)
|> | A12----| }
|_____| (6502) | }
|}
74174 7408
Flip Flop AND
A later model 7800 has an extra flip-flopish circuit
between the two pin 10's above:
,---------------<-------------,
| |
| _____ |
| | | |
P10 ----+-| | |
74174 | | |---, |
|_| | | |
| | | ______ |
|_____| | | | |
| | | |
`-->--| | 4 |
_____ | |-->-+--->---P10
| | | | 7408
Phi2 -----| | ,-->--| |
(out) | | | | |
| |---' |______|
D1 via ---| |
flop | |
|_____|
7402 NOR gate (ALL THREE)
The fix I did was as follows:
1. Cut trace coming from 7408 P10 with an exacto-knife
2. Verify cut trace with ohmmeter
3. Wire up a SPDT switch so that 7408 P10 can be connected
to 74174 P10 (expansion circuit) or 7402 P4 (later
circuit)
After I did this my space shuttle and robot tank carts
started working, however decathlon still wouldn't. I then
added a 1uf capacitor to ground onto the line and decathlon
came up fine. In my final assembly, I used a second switch
to connect the cap but decathlon worked even without the cap.
Seems like the loading of the switch itself was enough.
If you do this mod, you do so at your own risk, yadda yadda
yadda. If anyone wants to put this into a FAQ, go ahead.
I've seen posts that dark chambers sometimes doesn't work on
the expansion 7800's (which prompted the circuit change) and
that player paddles 3 and 4 are not connected in some units.
Have not seen any evidence of this in the 4 7800's I own.
Had a supercharger go bad in my expansion model 7800, and
wrecked two others modifying them for 8k games (even had one
working for a night). Got so sick of buying new ones that I
don't experiment with them anymore. You're on your own there.
That's about all I know. Wanted to fix the color problem in
my new portable 2600 but this ate up the whole evening, grumble
grumble :)
Sure hope these ascii drawings don't get garbled,
John Soper
Bob Colbert
<snip 7800 fix>
Thanks for the info, I have a couple of 7800's that I might try this on!
: Had a supercharger go bad in my expansion model 7800, and
: wrecked two others modifying them for 8k games (even had one
: working for a night). Got so sick of buying new ones that I
: don't experiment with them anymore. You're on your own there.
Get a Cuttle Cart! They are awesome, and they will load almost any game.
: That's about all I know. Wanted to fix the color problem in
: my new portable 2600 but this ate up the whole evening, grumble
: grumble :)
Grumble, grumble, I wish I had a portable 2600.
: Sure hope these ascii drawings don't get garbled,
They didn't!
Bob
--
Collector of Atari 2600 games - Want to write a game for the 2600?
Check out http://members.home.com/rcolbert1
John Soper
> John Soper <jso...@tripath.com> wrote:
> <snip 7800 fix>
> Thanks for the info, I have a couple of 7800's that I might try this on!
You're welcome Bob. I'm a big fan of your work too. John
C Schell
I understand exactly what this is doing, and why it will kill FE and
Supercharger games. I would say do NOT do the modification that John
has suggested, as there are much simpler alternatives. But let me look
into a little more and post again.
Great work John! Thanks very much for doing this and posting this
information. (I never got around to tracing that extra chip's
function.)
Chad
John Soper wrote:
--
========================================================
Chad Schell
Want to write your own Atari 2600 game? Or play
rare or homebrew games on an Actual 2600? Check out the
Atari 2600 Cuttle Cart - http://www.schells.com/
C Schell
Why Atari installed this circuit, I have no idea. Apparently they felt
that there would be bus contention problems between the 7800's internal ROM
and 2600 cartridges during the time when the address bus is unstable. Or
maybe they had planned some future activity that would make it matter. But
it doesn't exist at all on the original models, and on later models, even
though the pads for the 74LS02 are on the board, they didn't install it and
instead just shorted Pad 2 to Pad 4, thus bypassing the circuit and making
it act like a 1984 7800.
What this circuit does is drive the A12, A13, and A15 lines of the
cartridge port to zero during the time the address bus is changing. The
board is actually laid out so that a high pass filter will connect the NOR
gate in the lower left corner of John's diagram to the NOR gate on the
right. This effectively puts A12 to zero during the start of every machine
cycle, for about 150-200 ns on my 7800. If it were shorted as in John's
diagram, it would drive the A12 line low for 1/2 of the machine cycle.
This in itself could cause problems with older 2600 games, as the ROMS in
the cart are 450 ns ROMs, and they would not have 450 ns after they were
selected to the time the data was read. So I suspect that not many 7800s
have a direct short in place of the high pass filter.
So why does this break FE (Decathlon / Robot Tank) games, the Supercharger,
and the Cuttle Cart? Because these are the carts whose bankswitching
circuits use address bus driven state machines. I know for certain the
Cuttle Cart implements these state machines with a latch structure, and I
assume the others do as well. All three carts are sampling the bus before
or at about the same time the additional circuit in the 7800 is cutting
out. This means that they are not reliably reading the A12 line of the
bus, so their bankswitching schemes are not working correctly. For
example, when the supercharger should see 1FF8 to latch a control byte, it
is instead seeing 0FF8 and the latch does not occur.
I don't believe this circuit is necessary in any way. I removed it from my
7800 and the 7800 games I tried all worked fine. The facts that the 1984
models don't have this circuit, and Atari removed it on other 7800s
strengthen my belief. Thus I don't think there is any need to install a
switch, as the circuit does not need to be restored.
If the high pass filter is in place, this circuit can be removed by cutting
a lead on a single capacitor. If it is not, then the short in place of the
capacitor would have be to cut, and another short installed. I will take a
picture of my 7800 and post in on my website showing what I'm talking
about. I'll post again when the diagrams are ready. Of course attempting
this modification will be at your own risk. I'm working with a sample set
of only one incompatible 7800.
Chad
C Schell
I've put up a quick webpage with some photos and my suggested modification,
which involves the removal of a single capacitor. I've only tested it on one
7800 though, so if other people give it try, let me know how it goes.
It's definitely got to be considered experimental, as there are SO MANY 7800
variations out there. At least 4 different mother boards, most of which have
some amount of hand rework on them.
The site is at
http://www.schells.com/7800mod.shtml
Assuming your 7800 has the extra circuit formed by the 74LS02, and it contains
the highpass filter created by R66 and C64, removing C64 will turn the circuit
into nothing more than a delay line.
Chad
Genki
the keyboard with a hammer and typed:
>If the high pass filter is in place, this circuit can be removed by cutting
>a lead on a single capacitor. If it is not, then the short in place of the
>capacitor would have be to cut, and another short installed. I will take a
>picture of my 7800 and post in on my website showing what I'm talking
>about. I'll post again when the diagrams are ready. Of course attempting
>this modification will be at your own risk. I'm working with a sample set
>of only one incompatible 7800.
I'll be interested in that. I have 2 7800's that doesn't work right
with some 2600 games and I have a Supercharger to blow up if
necessary.
Just in case, anyone know how many *working* and *original*
Suercharger are left in the world? When i said original, I mean not
the mod to expand the RAM.
John Soper
> Alright,
>
> Why Atari installed this circuit, I have no idea. Apparently they felt
Thanks for grabbing the ball and running with it Chad.
I actually started looking into this after seeing a 1997 post by Jay Tilton.
You can all access this by searching for "tilton 7800 compatible" on deja.com
but these are the critical lines:
>There are some conspicuous differences between the early,
>fully-compatible 7800's and the later dodgy ones. Most noticeable is
>the presence on later ones of a 7402 (right side of the board, between
>a 7417 and a 7432)..
>I just realized that a recently acquired late 7800 (made in 1988) *is*
>fully compatible with all the finicky carts, and wouldn't you know it,
>the socket for the 7402 is empty except for a shorting resistor
>between pins 2 and 4. I'll see if there are anything else was changed
>so that any solder-jockey can doctor up his own late-model 7800 for
>full compatibility.
Love that phrase, "solder-jockey". Only took people 4 years to follow up
on this, kinda reminds me of things at my company.
Does anyone have an original 7800 that won't run some games?
http://home1.gte.net/smedley/atari.htm#7800
mentions problems with newer games like Dark Chambers but it runs fine
on my expansion port 7800
John Soper
Msorman
>http://home1.gte.net/smedley/atari.htm#7800
>mentions problems with newer games like Dark Chambers but it runs fine
>on my expansion port 7800
Hmm, I just tried 2600 Dark Chambers and it doesn't work on my two '84 models
but it works fine on my three later 7800. I can however play 2600 Dark Chambers
on the '84 models if I use the cuttle cart. I'll have to try modding one of my
later 7800's and see if it will still work with 2600 Dark Chambers.
John Soper
Man, this is like layers of an onion!
John
> Hmm, I just tried 2600 Dark Chambers and it doesn't work on my two '84 models
> Mitch
> http://atari7800.atari.org
VهدrXpêrt
>
>Just in case, anyone know how many *working* and *original*
>Suercharger are left in the world? When i said original, I mean not
>the mod to expand the RAM.
I got one non modified
One modified
one still sealed in the box
and the original I bought back in the early 80's which got fried in
the late 80's when I plugged it into my new 7800.
Msorman
>Man, this is like layers of an onion!
>John
>
OK, I did some more testing. I tested Model 2 Rev's A, B, and C. I also tested
a Model 1 for comparison. My test carts were the Cuttle Cart, 2600 Dark
Chambers and picture label versions of Space Shuttle, Robot Tank, and
Decathlon.
Model 1
C worked
DC didn't work
SS worked
RT worked
D worked
Model 2 Rev A
Before Mod
C worked
DC worked
SS screen rolls
RT screen rolls
D didn't work
After Mod
C didn't test
DC didn't work
SS worked
RT worked
D didn't work
Model 2 Rev B
Before Mod
C didn't work
DC worked
SS worked
RT didn't work
D didn't work
After Mod
C worked
DC worked
SS worked
RT didn't work
D worked
Model 2 Rev C
Before Mod
C didn't work
DC worked
SS worked
RT worked but very touchy
D didn't work
After Mod
C worked
DC worked
SS worked
RT worked but very touchy
D didn't work
Is this crazy or what?
A few notes. The Model 2 Rev C board was slightly different than the Rev A and
Rev B boards. The Model 1 board didn't having any "missing" components, all of
the Model two boards had some factory modifications done to them.
From what I could find the factory mods were as follows:
Rev A was missing R25, C10, R7, and Q4. Q4 had a jumper replacing it. The
second memory chip also had the trace going from pin 21 cut and had a 1k ohm
resistor between pin 21 and the other side of the cut trace.
Rev B was missing R25, C10, R7, CR1, R57, and Q4. Q4 and R57 had a jumper
replacing them.
Rev C was missing CR14, CR15, CR16, CR17 and had open jumpers at W4 and W5.
All of the boards had jumper W1 open (used with W2 to determine the OS ROM
size).
Robot Tank seemed to have a slightly larger case and didn't fit into the Rev C
slot too well, so it only worked sometimes.
Whew, I hope I didn't make any typos. If you have any questions let me know and
I'll try to answer them as best as possible.
John Soper
<snippety-snip>
Wow, good job on the data-taking Mitch!
It looks like the web-page at:
http://home1.gte.net/smedley/atari.htm#7800
was basically accurate. There are a few (how few?) 84/expansion_port
models that will not work with 2600 Dark Chambers. Atari added a
circuit which fixed this but broke Space Shuttle,Decathlon,Robot
Tank,Supercharger (do you get the feeling they only tested their own
products?). Then they took it back out.
By the way, if anyone goes to that webpage, click on "infamous
controllers" in the 5200 section. It will take you to the "Hall of
Shame" which is a pretty funny read.
From the Model 2 Rev A data, it seems like Chad's fix can break Dark
Chambers on SOME units. My fix (post #1) wouldn't have this issue
because it can switch between the different circuits. Think it will
fix Decathlon too, but not sure.
Of course Chad's mod is a lot quicker than mine so it's a definite
judgement call on time spent versus usefullness. Dark Chambers may
work fine after doing Chad's fix and even if it doesn't, you can
probably play it on the Cuttle Cart (or just play the 7800 version).
By the way my 7800 has six switches mounted on it:
1) Select 1984 or later compatibility circuit
2) Add 100uf cap onto same line (for decathlon)
3) Short stereo outputs together for mono
4) Short Chroma and Luma together for composite video (don't like the
looks of this, edges are kind of smeared. The transistors may be
pulling on each other)
5) 2600 color/black_and_white select
6) Select 2600 chroma or 7800 chroma (so the 2600 games don't have
dingy colors)
You might not want your 7800 looking that wacked out.
Someone else posted awhile ago that a good fix for 7800's with tight
cartridge slots is to get one of those Rom-switchers that lets you
plug in six carts and select which one with a button. Seems these are
kinda rare.
Regards, John
C Schell
> Chambers on SOME units. My fix (post #1) wouldn't have this issue
> because it can switch between the different circuits. Think it will
> fix Decathlon too, but not sure.
Yep, I wonder what silly thing they did in Dark Chambers to break it.
It's not like it uses a unique bankswitching format. From looking at
Kevin Horton's sizes.txt document I'd guess it has something to do with
the extra RAM.
If you want to keep compatibility I would cut one leg of the cap and
install a switch between the now loose leg and the pad that it used to be
soldered to. That way you could still restore it to its premodified state
with the flick of a switch, yet it's still a pretty simple modification.
But I'm pretty sure you could play Dark Chambers using the Cuttle Cart
after performing this modification. I believe Mitch mentioned that to me.
Chad
Msorman
>cartridge slots is to get one of those Rom-switchers that lets you
>plug in six carts and select which one with a button. Seems these are
>kinda rare.
To get around this problem, I use my 7800 high score cart. It works great.
Msorman
>after performing this modification. I believe Mitch mentioned that to me.
Yes, it works fine using the Cuttle Cart but not with the actual cart. Weird.