7800 BIOS bypass switch

[Flipper]

Well, I tried for this years ago before the encryption program was discovered.

 

And, not knowing much what I was doing, I rather failed.

 

However, I got my new 7800 this week, and so I got to playing with it today. Cutting pins 10 and 12 of the 74LS174 that sits two chips away from the power connector bypasses the BIOS and runs 7800 games signed or unsigned.

 

It does, however, also disable 2600 mode, as the BIOS switches into that.

 

Now, of course, it makes little sense unless you want to get around the several second wait when testing your programs. Which is nice, sometimes.

 

I'd recommend installing a DPDT switch, connection the holes for pins 10 and 12 to either +5V, or the original pins.

 

The theory is this. All the outputs of the 174 start out at 0V. Pin 10 disables the A12, A14, and A15 lines of the cartridge port and enables the BIOS ROM at the high end of memory. Pin 12 disables the HALT line to the CPU, preventing the MARIA from taking control when it needs to.

 

When the BIOS wants to read the key from the cartridge, it copies itself to RAM and fiddles with pin 10 to bring the cartridge into memory. It reads the signature and if it's valid, enables MARIA, disables TIA and locks out changes. If it is invalid, it disables MARIA, enables TIA, and locks out changes. According to the Atari docs, all cartridges must, upon startup, enable MARIA, disable TIA and lock out changes in case the BIOS ever changes. The configuration chip state is undefined at startup.

 

Cutting pin 10 causes U4 and BIOS to see a 'no connect' state, which is defined in TTL as HIGH. This disables the BIOS as CS2 goes high, and stops gating the cartridge port. The cartridge gets control at startup, sets the config chip to 7800 mode and things run perfectly, no startup delay.

 

Of course, programmers WILL break the rules. One on one basketball does NOT seem to set the config chip on startup, and this the TIA and MARIA try and run simultaneously. Galaga, Pole Position, and Centipede seem to do the right thing.

 

Pin 12 goes high after two writes to the config register, enabling HALT and thus MARIA accesses to memory. As the cartridge will only ever perform one write to the config register, pin 12 must be cut as well to allow MARIA to function normally.

 

So, it's not a perfect hack, but it would have allowed us to have been homebrewing a long time ago if someone with the schematics had ever found this. I think given the # of people that install mod chips in their NEW game consoles to play imports and that which they are not supposed to, plenty of people would have cut two pins to play homebrew games. We're all geeks anyway, those of us that would drag the old Atari out of storage.

[CPUWIZ]

Or you could just use Eckhardt's modified BIOS and not have any hassles at all.

 

http://home.arcor.de/estolberg/tools/

 

[Bruce Tomlin]

Much more useful is to do a proper EPROM mod:

 

http://xi6.com/hacks/7800eprom.html

 

Then install your favorite alternate code. Or just burn a copy of the original with address F4F5 (which is at an offset of 14F5 in an 8K EPROM) changed to zero.

[Flipper]

I think you both missed my point.

 

It took me about 5 minutes to use my method.

 

It would take me much longer to put in a new eprom.

 

And I have no eprom programmer.

 

Nor do I want to buy one.

 

Hmmm.

 

Who's method is more useful now?

[Flipper]

And my point also was about the skill level required.

 

In order to distribute homebrew code, the skill level needs to be VERY VERY low, so that the average classic gaming collector is willing to modify his console.

 

Clipping two pins and possibly installing a switch is very easy, and something that most classic gaming people can do.

 

Burning an EPROM is both expensive (if you do not already have a programmer), and above the skill level of most collectors. Building the needed inverter circuit is also above the skill level of most collectors.

 

My point was, rather than spending 5/10 years saying "No point in writing homebrew games, as no one can run them.", we should have been looking for quick and simple mods to let people run homebrew games. The EPROM circuit, while far from complicated, is more than the average collector will want to build to run a homebrew game.

[Flipper]

Yes, I am aware that we have the code signing issue solved. But really...rather than sitting on our collective asses, we should have been writing games and working out EASY console mods.

[Bruce Tomlin]

Except that the problem really wasn't the lockout, it was the lack of tools. There were no decent emulators until recently, and the Cuttle Cart 2 (for devving on real hardware) only came out a year ago. That was three years after the private key was found, so clearly the lockout wasn't the problem.

[Scooter83]

I wonder if this is a much cheaper way for me to test out prototype pcbs that are not signed. Would be worth it to get a 20 or 30.00 7800 and do this If it's gonna do the job. Opinions??