Problems with Attack of the PETSCII Robots

[prorok]

Well, I've run the next rounds of experiments and not much closer to getting PETSCII running yet.

 

EXPERIMENT 1: I plugged in the over-spec'd 2500 mA power supply and PETSCII Robots fails in the same consistent way after the first screen.

 

EXPERIMENT 2: I completed a significant "Rite of Passage" and successfully placed and received my first-ever order from Bradley at Best Electronics. Included in that order was my lifetime maximum of ONE POKEY chip: a sealed box copy of Ballblazer. Given all my Atari stuff as a kid came second, third, or fourth hand from garage sales, this was the first time in my life I had a brand new Atari game box I could tear the shrink-wrap off of. Plugged in Ballblazer, played a few rounds, and it appears to work just fine, complete with the groovy music.

 

I had been operating from a theory that maybe the 7800 wasn't getting the power it needed to run the extra HW on the PETSCII carts. That's why I went with a beefier power supply and another title that has extra HW and extra current draw. If I want to keep beating down the power supply theory, I could take a stern look at the internal voltage regulator, but I'm starting to doubt this is the problem since I'd think the problem would be more intermittent, vary with temperature or playing time or something...and why would it have anything to do with the corruption of the Atari Age logo on the initial screen?

 

Nevertheless, I'm getting ready to open up my 7800 and start probing the regulator. In the meantime, is there anything else I should take a look at while I'm inside?

 

Dan

[+-^CrossBow^-]

10 minutes ago, prorok said:

Well, I've run the next rounds of experiments and not much closer to getting PETSCII running yet.

 

EXPERIMENT 1: I plugged in the over-spec'd 2500 mA power supply and PETSCII Robots fails in the same consistent way after the first screen.

 

EXPERIMENT 2: I completed a significant "Rite of Passage" and successfully placed and received my first-ever order from Bradley at Best Electronics. Included in that order was my lifetime maximum of ONE POKEY chip: a sealed box copy of Ballblazer. Given all my Atari stuff as a kid came second, third, or fourth hand from garage sales, this was the first time in my life I had a brand new Atari game box I could tear the shrink-wrap off of. Plugged in Ballblazer, played a few rounds, and it appears to work just fine, complete with the groovy music.

 

I had been operating from a theory that maybe the 7800 wasn't getting the power it needed to run the extra HW on the PETSCII carts. That's why I went with a beefier power supply and another title that has extra HW and extra current draw. If I want to keep beating down the power supply theory, I could take a stern look at the internal voltage regulator, but I'm starting to doubt this is the problem since I'd think the problem would be more intermittent, vary with temperature or playing time or something...and why would it have anything to do with the corruption of the Atari Age logo on the initial screen?

 

Nevertheless, I'm getting ready to open up my 7800 and start probing the regulator. In the meantime, is there anything else I should take a look at while I'm inside?

 

Dan

Yes and I mentioned it already I thought? Look to see if you have the extra timing circuit in place. If you do, there is an easy way to disable that to see if it helps.

 

 

Next is to look for the presence of a capacitor (will look like a resistor but usually green or bluish in color) soldered on the bottom side of the PCB usually off pin 25 of the 6502 CPU. If so, you could try to disable that deglitch cap by removing one of the leads from it and see if anything changes.

 

 

I had a 7800 here that wouldn't play the new popeye release and I found that factory added deglich cap was at fault. Once I removed it, Popeye was working just fine.Prior this it was doing something similar when starting it up as you have described for Petscii. The only risk I know of for removing that degltich capacitor is that it could cause Ballblazer (possibly Pole Position too) to not play stable anymore from an actual cartridge. But it seems to effect one release version of Ballblazer more than the other. And if it did cause issues with those games, both games end up working fine through flash carts I've found.

 

 

Edited September 29, 2023 by -^CrossBow^-
Added pics for clarity

[desiv]

59 minutes ago, -^CrossBow^- said:

Look to see if you have the extra timing circuit in place.

"The more you know!!!"
I hadn't heard about this. 

And as a result I just spent the last half hour researching this and testing some 2600 games on my 7800 to see how my 7800 handled them. (They worked fine on mine)  
It will be interesting to see if this is related to the OPs problem...

[+-^CrossBow^-]

28 minutes ago, desiv said:

"The more you know!!!"
I hadn't heard about this. 

And as a result I just spent the last half hour researching this and testing some 2600 games on my 7800 to see how my 7800 handled them. (They worked fine on mine)  
It will be interesting to see if this is related to the OPs problem...

Well, you won't really know if you have that in place or not unless you open up the console and look. The whole point of it was to improve compatibility with later made 2600 titles. But it did so at the cost of removing compatibility with a few Activision games as a result as well as incompatibility with Supercharger games. Like frogger not showing the sprites on the game play screen for instance.

 

Earlier made 7800s in 84 - 86 timeframe don't usually have the extra timing circuit. But starting in late 86 and through most of 87 that part of the board has the extra chips and that capacitor marked as C64. You only need to clip or de-solder one leg of it to disable the timing circuit. I usually lift the right side leg on the consoles I service. Newer consoles after this will still have the markings on the board for the extra ICs and capacitor, but they won't be populated and instead have a vertical standing 0Ω resistor that is usually tan in color with the single black band across the middle that is soldered into play to bypass all of that. 

 

I've only confirmed that disabling that circuit effects 2600 Dark Chambers (But not always...) and 2600 Kung Fu Master with it disabled. I've been told that Stargate and 2600 Double Dragon could also be effected, but the little testing I've done with those games that hasn't been the case for me. But 2600 Kung Fu Master never works on the 7800 without that circuit on consoles that the rest of it in place. Just gives a black screen as I recall.

 

The deglitch cap I talked about is what caused issues on one specific 7800 sent to me for the audio issues on their previous AV mod they had in place. I was using my actual 7800 Popeye cart to test it as that was the main game the owner was complaining about and found it was a glitchy mess that would lock up shortly after starting a game. So I don't know if there have been any other reports of that or not, but I did encounter it on one 7800 where it did NOT like that factory bodge in place so I removed it. Interesting, my daily driver 7800 has that same factory bodge on it as that is the pic from the bottom of my normal 7800 and it doesn't have an issue with that same Popeye cartridge. So it very much varies from console to console.

 

[prorok]

2 hours ago, -^CrossBow^- said:

Yes and I mentioned it already I thought? Look to see if you have the extra timing circuit in place. If you do, there is an easy way to disable that to see if it helps.

 

 

Next is to look for the presence of a capacitor (will look like a resistor but usually green or bluish in color) soldered on the bottom side of the PCB usually off pin 25 of the 6502 CPU. If so, you could try to disable that deglitch cap by removing one of the leads from it and see if anything changes.

 

 

I had a 7800 here that wouldn't play the new popeye release and I found that factory added deglich cap was at fault. Once I removed it, Popeye was working just fine.Prior this it was doing something similar when starting it up as you have described for Petscii. The only risk I know of for removing that degltich capacitor is that it could cause Ballblazer (possibly Pole Position too) to not play stable anymore from an actual cartridge. But it seems to effect one release version of Ballblazer more than the other. And if it did cause issues with those games, both games end up working fine through flash carts I've found.

 

 

Thanks for reiterating the suggestion to go exploring for the timing circuit. It appears mine was intact: C64 was in place. I went ahead and desoldered one leg and lifted up, but, alas, PETSCII is no different...same AA logo corruption on first screen and crash thereafter.

Meanwhile, there's nothing on the bottom of the board in the vicinity of the 6502. The only thing on the top of the board in the vicinity of the 6502 is a yellowy-orange thing jumpered across two pins of the MARIA chip.

 

 

Any thoughts on next steps? Thanks for your help so far!

[+-^CrossBow^-]

1 minute ago, prorok said:

Thanks for reiterating the suggestion to go exploring for the timing circuit. It appears mine was intact: C64 was in place. I went ahead and desoldered one leg and lifted up, but, alas, PETSCII is no different...same AA logo corruption on first screen and crash thereafter.

Meanwhile, there's nothing on the bottom of the board in the vicinity of the 6502. The only thing on the top of the board in the vicinity of the 6502 is a yellowy-orange thing jumpered across two pins of the MARIA chip.

 

 

Any thoughts on next steps? Thanks for your help so far!

That is just a capacitor and is supposed to be there. Have you reached out to Al to get a replacement yet? At this point I don't really have any other ideas. And....yes...

 

You can leave that capacitor disconnected on the extra timing circuit for now if you want and it should be fine. Again, the only games that does effect for sure is 2600 dark chambers and 2600 kung fu master that I can confirm first hand.

 

Wait a sec... both your MARIA and CPU ICs are in sockets? Are the other two large ICs opposite them also in sockets? It doesn't look like it from your picture of the timing circuit? That would indicate those chips were replaced out at some point...

 

[marauder666]

CPU definately swapped, its 1984.

[+-^CrossBow^-]

26 minutes ago, marauder666 said:

CPU definately swapped, its 1984.

It is odd how much older than the other ICs it is. But that 6502 could have been the first thing that was grabbed from the bin to install into it. But the fact that both MARIA and the 6502 are in sockets indicates that system might have failed initial QC and had to be reworked possibly before it was sold out as new.

 

[prorok]

1 hour ago, -^CrossBow^- said:

That is just a capacitor and is supposed to be there. Have you reached out to Al to get a replacement yet? At this point I don't really have any other ideas. And....yes...

 

You can leave that capacitor disconnected on the extra timing circuit for now if you want and it should be fine. Again, the only games that does effect for sure is 2600 dark chambers and 2600 kung fu master that I can confirm first hand.

 

Wait a sec... both your MARIA and CPU ICs are in sockets? Are the other two large ICs opposite them also in sockets? It doesn't look like it from your picture of the timing circuit? That would indicate those chips were replaced out at some point...

 

 

Yes, Al has sent me a replacement and it performs consistently with the original cart. I just tested the replacement cart with the timing circuit disabled and the bad behavior is still the same.

 

Yes, only the MARIA and the 6502 are socketed.

 

Dan

 

[prorok]

4 minutes ago, -^CrossBow^- said:

It is odd how much older than the other ICs it is. But that 6502 could have been the first thing that was grabbed from the bin to install into it. But the fact that both MARIA and the 6502 are in sockets indicates that system might have failed initial QC and had to be reworked possibly before it was sold out as new.

 

Maybe, or it could have been done later. I purchased the system with several mods to have all the bells and whistles: UAV S-video out, blue power LED, recapped with new power regulator, and a normal barrel connector for the power supply. It's possible that as part of the Cadillac refurbishment treatment, it got sockets added for the MARIA and the 6502.

 

[+-^CrossBow^-]

2 minutes ago, prorok said:

Maybe, or it could have been done later. I purchased the system with several mods to have all the bells and whistles: UAV S-video out, blue power LED, recapped with new power regulator, and a normal barrel connector for the power supply. It's possible that as part of the Cadillac refurbishment treatment, it got sockets added for the MARIA and the 6502.

 

Its possible but you wouldn't normally do that unless there was an issue possibly with those chips. I've seen a few factory made 7800s with the Maria in a socket, or the CPU in a socket. But not usually mixed like that one and especially with the CPU dating from '84. It might be worth a shot to try a different 6502 in that console as I just don't see anything else that might cause this at this point.

 

It is worth a check to make sure all of the cartridge port pins look to have good solder joints on the board. 

[prorok]

3 hours ago, -^CrossBow^- said:

Its possible but you wouldn't normally do that unless there was an issue possibly with those chips. I've seen a few factory made 7800s with the Maria in a socket, or the CPU in a socket. But not usually mixed like that one and especially with the CPU dating from '84. It might be worth a shot to try a different 6502 in that console as I just don't see anything else that might cause this at this point.

 

It is worth a check to make sure all of the cartridge port pins look to have good solder joints on the board. 

Tried another experiment since everything on my 5200 is socketed in. I swapped the CPU from the 5200 into the 7800, but unfortunately, no change to the 7800 behavior. At least I know that older 6502 is fine.

 

Probably should have tried this long ago, but I have a fresh can of Deoxit to try on the cart port and I will take a careful look at the cart slot solder joints.

[prorok]

Sprayed some Deoxit in the cart slot and took a look at the cart slot solder joints...all looks okay to my eye. Alas, still not working.

 

[firebottle]

Just wondering... does the 7800 use a stock 6502 or the 6502c "Sally" chip?

[prorok]

41 minutes ago, firebottle said:

Just wondering... does the 7800 use a stock 6502 or the 6502c "Sally" chip?

Even though people call the Sally chip a 6502C, my understanding is that a real 6502C is something different, a 6502 that runs at 4 MHz.

 

The Atari 7800 uses the same custom Atari C014806 "Sally" 6502 variant used in the 5200 and the 8-bit Atari computers. You can see the C014806 markings in the picture I posted earlier in the thread.

Edited September 30, 2023 by prorok

[+-^CrossBow^-]

Well, that pretty much just leaves possibly the RIOT but still this is the ONLY game to give you any issues correct?

 

[prorok]

11 hours ago, -^CrossBow^- said:

Well, that pretty much just leaves possibly the RIOT but still this is the ONLY game to give you any issues correct?

 

Yes, this is the only game I've noticed a problem on. However, I generally play my 2600 games on the 2600 and only own about two dozen pretty common 7800 games, of which I play a minority. PETSCII is my first Homebrew and, in an effort to see if a "challenging" cart might be the problem, just procure Ballblazer this week as a test.

 

Any suggestions for specific titles to try to see if I can repro the problem on anything else?

 

In the meantime, I guess it's time to source a new RIOT and a socket for it. Any suggestions on good sources for that?

Dan

[prorok]

Did a little more testing today to see if I have problems when any other games and just started walking through my collection in alphabetical order. Everything works in a way that doesn't affect gameplay, but I was being hyper sensitive to glitches that normally I'd just shrug off as artifacts of playing 8-bit hardware on a CRT.

 

Asteroids: seems to work fine without any glitches

Ballblazer: there's a brief flash of static between the Lucasfilm title screen and the Ballblazer menu select. The music at some points seems to stutter when it loops

Choplifter: a few pixels here or there are glitchy

Dig Dug: similarly, a few glitchy / glittery pixels here and there on the screen

 

Again, I guess I'd have categorized all of this as normal, but maybe not. The image is all composite video out with a UAV mod running to a Sony Trinitron TV that could use a tune up.

 

All of this seem normal or does it clue into where to look for the "hard fails" of PETSCII Robots?

[RevEng]

The glitchy/glittery pixels and glitches are a sign that something is out of spec, either in terms of bus timing, or something interfering with the bus. Maria is sometimes getting bad data when it's fetching the graphics data.

[prorok]

Okay, more "newbie" questions from a guy who has been playing on a 7800 since the 90's...

 

Is there an establish approach to 7800 debugging? I'm approaching the point where I'll start collecting all the IC's, socket EVERY DIP package on the board and just start swapping parts until I find some love. It would be cool if there were some diagnostics I could run that would exercise the CPU, MARIA, TIA, RAM, etc, so I wasn't just randomly swapping parts with the precision of a shotgun blast. I seem to recall there's some Uber-rare official 7800 diagnostics cart that is essentially unobtanium. Has someone developed a Homebrew test suite either on its own cart or that can be loaded on a Game Drive?

 

Thanks for the guidance.

 

And if you're still reading the thread, Al, let me know when time's up and you want me to send one of the two PETSCII carts back to you. I'm pretty sure it's not the cart's fault, but I'm happy to hold on to them until this problem is solved if you're cool with that.

[+-^CrossBow^-]

3 hours ago, prorok said:

Okay, more "newbie" questions from a guy who has been playing on a 7800 since the 90's...

 

Is there an establish approach to 7800 debugging? I'm approaching the point where I'll start collecting all the IC's, socket EVERY DIP package on the board and just start swapping parts until I find some love. It would be cool if there were some diagnostics I could run that would exercise the CPU, MARIA, TIA, RAM, etc, so I wasn't just randomly swapping parts with the precision of a shotgun blast. I seem to recall there's some Uber-rare official 7800 diagnostics cart that is essentially unobtanium. Has someone developed a Homebrew test suite either on its own cart or that can be loaded on a Game Drive?

 

Thanks for the guidance.

 

And if you're still reading the thread, Al, let me know when time's up and you want me to send one of the two PETSCII carts back to you. I'm pretty sure it's not the cart's fault, but I'm happy to hold on to them until this problem is solved if you're cool with that.

The diagnostic cartridge ROM is out there and can be put onto a custom cartridge. The main issue with that diagnostic is that I believed it also required some loop back adapters that plugged into the controller ports similar to the Kludge plugs for the 2600 diagnostics.

 

I believe you stated that you don't have the extra bodge capacitor on the bottom of the board off pin25 of the 6502 CPU correct? While it shouldn't be needed, and I did state that removing it fixed 7800 Popeye for another AA member, you might try adding one in to see if it helps? I believe the original cap was only like 420pf. I don't keep anything that exact on hand, but have had luck using 1nf mlcc caps in that place on a few consoles. Worst that will happen if you do install it is that the console won't boot up any games at that point but it won't do any permanent damage. 

 

 

[GranitePenguin]

2 hours ago, -^CrossBow^- said:

The diagnostic cartridge ROM is out there and can be put onto a custom cartridge. The main issue with that diagnostic is that I believed it also required some loop back adapters that plugged into the controller ports similar to the Kludge plugs for the 2600 diagnostics.

 

This page appears to have info on what those loopback adapters need to look like:  https://pages.cs.wisc.edu/~harvey/7800/diag.html

[+-^CrossBow^-]

3 hours ago, GranitePenguin said:

This page appears to have info on what those loopback adapters need to look like:  https://pages.cs.wisc.edu/~harvey/7800/diag.html

Yes I've seen that before but just never bothered to make up a set. Honestly I'm able to pretty much troubleshoot most issues on the 7800 between the 7800 Utility cartridge, Paul Slocum's Test Cart, and the 2600 diagnostic cart plus the troubleshooting pages for the 7800. The good thing about the 7800 having a 2600 inside it, is that most of the stuff and circuit for the 2600 consoles is mimicked on the 7800. An example is the paddle lines. There is no difference in the circuit schematics for those so the same tests for the 2600 paddle line issues can be used on the 7800.

 

[prorok]

The debug continues...

 

While cycling through my Atari games today, I observed odd behavior with BOTH copies of Tower Toppler I have. I can put in the game and I can select 1 player or 2 player and sound on/off on the main menu. If I choose a 1 player or 2 player game with sound on, both cartridges have the same off behavior that they load the level and after the frog is ready to move on the platform, the level ends and progresses to the next level. This proceeds all the way through all the levels so that it cycles back to the first level. Interesting stuff.

 

I have also taken the big step to figure out this mystery by procuring a second (purportedly working and stock) Atari 7800. It has arrived, but I haven't plugged it in to start more testing of the now two bizarre games: PETSCII and Tower Toppler.

 

I foresee lots of desoldering, soldering and chip swapping in my future while I hunt for the problem.

[+Mitch]

That is normal for Tower Toppler. Check your difficulty switches.

 

Mitch