Major Havoc Random Resets

Yeh, scaling looks good. (y)
If you want to feel better, the vectors on your test screens look a thousand times better than mine. LOL (I need to tweak some pots).
 
I decided to try to test one theory that this might be thermal related, and produced something really cursed:


I'll cut you some slack because you said one hole was there already...

But if you want to do this again, there's already a nice big hole above the back door. You can slide a small 12V fan up into that, to help move a little more air beyond the existing convection cooling. Even a tiny fan in that hole will push all of the hot air out the top of the cab, and suck cool air in through the vent that already exists at the bottom of the cab (down where the wheels are), cooling everything inside the cab.

Drilling additional holes is not necessary, and actually just makes the airflow worse than using the existing holes, which are meant to draw air through the entire height of the cab.

And if you want to test if something is thermal, the right way to do it is to put an ice cube in a ziploc bag, and touch it to individual chips. Or buy a can of freeze spray.
 
I'll cut you some slack because you said one hole was there already...
Yeah, apparently it's part of the conversion instructions for Major Havoc. I guess the concern was that the 6100 would overheat.

1743107172516.png

I removed this fan (and its associated wiring harness) for now.
But if you want to do this again, there's already a nice big hole above the back door. You can slide a small 12V fan up into that, to help move a little more air beyond the existing convection cooling. Even a tiny fan in that hole will push all of the hot air out the top of the cab, and suck cool air in through the vent that already exists at the bottom of the cab (down where the wheels are), cooling everything inside the cab.
Noted. I'll look into finding some tiny fans next time.
 
The fans you used would probably technically have worked, as I think most of them are the same thickness. (Though those 5" ones would be pretty inefficient up in the top vent.) But they would work. You actually need very little power to turn all of the air in the cab over fast enough to keep everything cool.

I experimented with 3" ones a long time ago, up in that hole. I got ones with blue LED's, which looked kinda neat in my Space Duel. But ultimately I removed them, as extra cooling is really only needed if you're operating on location. For a few hours of home use a week, it isn't necessary if your systems are working properly.

Also, the even better way to do it is to use the vent on the top of the cab. See this thread:

 
I only have one Vector Cab with a fan - my SW upright (and it came with the fan, so I let it be).
I don't like fans. Atari didn't use them in these particular cabs and we don't run them 24/7 anyway and do use them in thermally managed environments.
More importantly, fans don't just make audible noise, they also make EM and electrical noise. Both bad for the logic board and monitor, IMHO.
Ideally any fan should also be accompanied by an appropriate line filter.

LOL - yes, fans, unshielded speakers, AC transformers, and fluorescent lights ... the ongoing noise debate.

As always -> YMMV. (laughs)
 
I'm noticing that the four minutes thing is fairly consistent.

Cold start -> four minutes -> reset/problem -> fine afterward. Weird. This time it crashed and stayed crashed until I tapped the machine and it came back up. Gotta be some kind of connection thing somewhere.
 
I doubt it's the AR. AR's either work or they don't. Never hurts to rule it out though.

I think you have another bad socket. Pay extra attention to the ones in the center of the board. Those tend to break pins under the sockets more often, because the middle of the board flexes more than the edges, when the board is handled.
 
I doubt it's the AR. AR's either work or they don't. Never hurts to rule it out though.

I think you have another bad socket. Pay extra attention to the ones in the center of the board. Those tend to break pins under the sockets more often, because the middle of the board flexes more than the edges, when the board is handled.
Yeah I suspect the same thing. I had just rebuilt and verified that AR with no change. I ordered more sockets yesterday. I was out.
 
Replace those awful sockets already. They look like crap inside where you can't see them until you pop the plastic frame off the pins. If it keeps doing it after that get some freeze spray and try to bring it back up when it's in failure by spraying and finding the suspect chip whatever that turns out to be. I doubt you will have to do any spraying though. Check your yoke connection to the deflection board for the monitor wiggle.
 
Replace those awful sockets already. They look like crap inside where you can't see them until you pop the plastic frame off the pins. If it keeps doing it after that get some freeze spray and try to bring it back up when it's in failure by spraying and finding the suspect chip whatever that turns out to be. I doubt you will have to do any spraying though. Check your yoke connection to the deflection board for the monitor wiggle.
All of the single wipe sockets have been replaced and havoc is still unhappy.

There are still some OEM sockets in place, but they're dual wipe, so I didn't think they needed replacing. Are they nearly as bad as the single wipes?
 
I have nearly had it with this machine.

The other day it was a LETA problem, and now it just won't boot the game at all.

I have no idea what the heck is wrong with it.
It boots into test mode just fine, says all tests have passed with flying colors. Vector generator tests look just swell.

But when it comes time to boot to game mode? Nah. Not about that. I'm treated to a blank screen, no deflection, and lazily flashing start lights.

Some initial troubleshooting tests:
- Swap CPUs - No Change
- Swap around Rams - No Change
- Test CPUs out of circuit - all good
- Test RAMs out of circuit - all good
- Test every single socketed chip I can out of circuit - all good.

Hooked up the cat box though, and then things are interesting. The gamma CPU section seems fine. But I can't get it to read the Alpha section. Maybe something intermittent? But I can't get it to read any of the ROMs or any of the RAMs properly. But if that were the case, how in the hell would it make it to the test mode with the actual CPU installed?
 
I have nearly had it with this machine.

The other day it was a LETA problem, and now it just won't boot the game at all.

I have no idea what the heck is wrong with it.
It boots into test mode just fine, says all tests have passed with flying colors. Vector generator tests look just swell.

But when it comes time to boot to game mode? Nah. Not about that. I'm treated to a blank screen, no deflection, and lazily flashing start lights.

Some initial troubleshooting tests:
- Swap CPUs - No Change
- Swap around Rams - No Change
- Test CPUs out of circuit - all good
- Test RAMs out of circuit - all good
- Test every single socketed chip I can out of circuit - all good.

Hooked up the cat box though, and then things are interesting. The gamma CPU section seems fine. But I can't get it to read the Alpha section. Maybe something intermittent? But I can't get it to read any of the ROMs or any of the RAMs properly. But if that were the case, how in the hell would it make it to the test mode with the actual CPU installed?

Something must be killing your bus, that would explain the LETA not working and not being able to access RAM or ROMs.
 
Very cool, if I hit the watchdog disable and go from test mode (which still works...) to game mode, it draws a box.

*Neat*

I was able to get the rams reliably testing good in circuit with the cat box. A new socket helped there. However, still can't get the rom checks to pass. Though I recall those SEEQ EPROMs have NEVER worked well with my CATBox. Only two ROMs are currently verifying in circuit- the replacements I burned when I fixed this thing the *first time*
 

Attachments

  • IMG20250426230008.jpg
    IMG20250426230008.jpg
    275.7 KB · Views: 7
Trying to think what the next thing worth trying here is. What would cause the game mode to watchdog and not boot, but wouldn't get caught by self test, which appears to fully work? My time to mess with this is about to go way down so I've got to be choosier about what I investigate.

Clearly the ROM bus and associated decoders for the alpha CPU must be working, otherwise it wouldn't be able to make it to self test.

The RAM also seems good here, no issues in self test, test in circuit (now) or out of circuit, so that bus is probably good too.

The vector generation section *seems* to be working, but I suppose there's a potential for a problem there. The scaling test works fine. I'm using a repro AVG.

I wonder if it has something to do with the gamma CPU or communication between them? Both CPUs are definitely good, and Gamma isn't reporting any errors in self test, but maybe that's a false positive?

I'm not exactly familiar with exactly how Major Havoc's self test works and what elements it precisely tests of the various components it touches.
 
Back
Top Bottom