ANTIC P9 / GTIA P20, Troubleshooting.

[kazriko]

I recently got my old first computer out of storage. It's a 130xe and it's been dead for 30 years, and I wanted to repair it. It shows a solid black screen when running.

 

I ran a logic analyzer across a variety of pins on the system to see what seemed to be working. The address and data lines across all ram chips, the sally, antic, and gtia all seem to be good. The BASIC chip also has solid connectivity to these lines.

 

The OS Rom chip was socketed by the former owner of the machine (I bought it in 1986 used,) and most of the pins on this socket aren't making contact to the address and data lines. This is likely the cause of the fault in the system. The socket doesn't seem to be gripping the pins very well and the chip comes out a bit too easily.

 

But the other thing I found is that P5 on ANTIC, which goes to P20 on GTIA, and P23 on GTIA are all solid logic low, when the Sam's guide says they should be going back and forth between low and high. Most of the other pin the wave form is provided for seem to have good signals on the analyzer. (I don't have a scope to verify the waveforms though.)

 

Would the OS chip being disconnected cause it to be a solid black screen all on its own, or could there be some other problems with the GTIA and ANTIC?

 

I'm going to be ordering a new socket for that (and maybe a new ROM, just to be safe,) and wanted to order any other chips that could be bad at the same time so I don't need to wait around twice if the first thing doesn't fix it.

 

(I already have sockets for the RAM chips, and new ram chips on the way, since I figured this is a common failure mode with a MT ram based 130xe's.)

Edited November 22, 2018 by kazriko

[Rybags]

Antic Pin 9 is /HALT and actually goes to pin 26 on GTIA. GTIA only uses it as a reference to know when to pull PMG data off the data bus after a HSYNC.

Regardless of what the CPU or anything else is doing, /HALT should always be active around 9% of the time due to refresh cycles, with an active display typically closer to 1/3rd active as it also gets pulled low for video memory accesses. But a black screen would suggest no screen is active, and if the OS Rom isn't connected then there'll be no screen set up anyway.

 

The 40 pin DIP sockets are still a common part and likely buyable for a couple of bucks at electronics stores. Some people insist on more expensive precision machined sockets but for something you'll likely plug in once and just leave the cheap ones are fine. Upgrades like 32in1 OS also work OK in them.

 

Without an OS the Atari is an oversized paperweight and not exactly easy to do diagnostics on. The alternative method is if you have a known good machine, swap each chip into it to test functionality. Just note there's 2 Antic versions, with 130XE using the later one as it caters for refreshing an extra set of rows.

[kazriko]

Antic Pin 9 is /HALT and actually goes to pin 26 on GTIA. GTIA only uses it as a reference to know when to pull PMG data off the data bus after a HSYNC.

Regardless of what the CPU or anything else is doing, /HALT should always be active around 9% of the time due to refresh cycles, with an active display typically closer to 1/3rd active as it also gets pulled low for video memory accesses. But a black screen would suggest no screen is active, and if the OS Rom isn't connected then there'll be no screen set up anyway.

 

The 40 pin DIP sockets are still a common part and likely buyable for a couple of bucks at electronics stores. Some people insist on more expensive precision machined sockets but for something you'll likely plug in once and just leave the cheap ones are fine. Upgrades like 32in1 OS also work OK in them.

 

Without an OS the Atari is an oversized paperweight and not exactly easy to do diagnostics on. The alternative method is if you have a known good machine, swap each chip into it to test functionality. Just note there's 2 Antic versions, with 130XE using the later one as it caters for refreshing an extra set of rows.

 

Oops, I meant P5 on the Antic not P9. Reading the schematic zoomed out when I sat down at my computer and got those numbers mixed up.

 

There's basically no electronics shops within 250 miles of me, so everything has to be ordered in. Of course, the cheap socket that was in my system was mostly just plugged in and not touched, but ended up failing anyway, so I'd probably want to go with something at least above the bottom of the line. (It did have a 3-chip ROM switcher board sitting on it, and that was pushed down by the RF shield and keyboard, so the stress of that might have worn it out as we were typing on it years ago.) I imagine the $2 gold plated ones would likely be fine. I also have an OmniviewXE OS chip that the former owner had put in, but I don't know if either chip is still any good. When I was a kid, I had no idea what that switch on the side of the computer did...

 

Will order in a new 28 pin socket soon though for that OS rom.

 

I have a working XEGS as well, but I would hate to take that one apart if I don't have to. My son uses that to play Frogger and Ms. Pacman.

[Rybags]

Pin 5 is AN2 which is part of the command bus that Antic uses to tell GTIA how to draw the playfield graphics.

 

IMO best sign of life test on Antic is probably what you originally posted - there should always be constant activity on /Halt and also on /Refresh pin 8.

 

Re XEGS, or just about any XE for that matter - they're generally not socketed at all and when there is a socket or two present, the theory is that it was a failed component that was pulled from the factory line and socketed before replacing the faulty chip.

[kazriko]

Pin 5 is AN2 which is part of the command bus that Antic uses to tell GTIA how to draw the playfield graphics.

 

IMO best sign of life test on Antic is probably what you originally posted - there should always be constant activity on /Halt and also on /Refresh pin 8.

 

Re XEGS, or just about any XE for that matter - they're generally not socketed at all and when there is a socket or two present, the theory is that it was a failed component that was pulled from the factory line and socketed before replacing the faulty chip.

 

I checked Pin 9, and it's getting high pulses every second or so, as is pin 8.

 

Yeah, the only socketed thing on either system was the OS rom, and that's just because the former owner modded it to switch OS images.

 

Ugh. I pulled the Omniview rom out and put it in after spraying some contact cleaner on both sides, more pins were making contact, but some of the pins that aren't making contact on the chip, also aren't making contact where the socket is soldered into the board. Looks like the stress on that socket lifted some of the traces. I'm probably going to need to install some bodge wires to get this thing going again.

[kazriko]

Ugh. I pulled the Omniview rom out and put it in after spraying some contact cleaner on both sides, more pins were making contact, but some of the pins that aren't making contact on the chip, also aren't making contact where the socket is soldered into the board. Looks like the stress on that socket lifted some of the traces. I'm probably going to need to install some bodge wires to get this thing going again.

 

And I pull the board out of the bottom case and find 8 bodge wires already installed... the former owner did a bad job installing this socket.

[kazriko]

Thanks for the help! Cleaning it up, using contact cleaner, reseating everything, then reinstalling the triple OS module managed to get the thing up and talking again. It makes a lot of noise on startup, but it functions after that. I need to try the SIO port though, and the other expansion ports just to make sure that all works, but I accomplished the goal of getting this thing running again.

[BillC]

 

Oops, I meant P5 on the Antic not P9. Reading the schematic zoomed out when I sat down at my computer and got those numbers mixed up.

 

There's basically no electronics shops within 250 miles of me, so everything has to be ordered in. Of course, the cheap socket that was in my system was mostly just plugged in and not touched, but ended up failing anyway, so I'd probably want to go with something at least above the bottom of the line. (It did have a 3-chip ROM switcher board sitting on it, and that was pushed down by the RF shield and keyboard, so the stress of that might have worn it out as we were typing on it years ago.) I imagine the $2 gold plated ones would likely be fine. I also have an OmniviewXE OS chip that the former owner had put in, but I don't know if either chip is still any good. When I was a kid, I had no idea what that switch on the side of the computer did...

 

Will order in a new 28 pin socket soon though for that OS rom.

 

I have a working XEGS as well, but I would hate to take that one apart if I don't have to. My son uses that to play Frogger and Ms. Pacman.

The 3-chip ROM switcher board is most likely the Newell Ramrod XL, I don't think they had a different version for the 130XE. Newell Industries is the same company that made the Omnimon and Omniview OS chips, and also RAM upgrades up to 4MB. I believe when Newell Industries shut down ICD may have acquired their product line.

Edited November 22, 2018 by BillC

[kazriko]

The 3-chip ROM switcher board is most likely the Newell Ramrod XL, I don't think they had a different version for the 130XE. Newell Industries is the same company that made the Omnimon and Omniview OS chips, and also RAM upgrades up to 4MB. I believe when Newell Industries shut down ICD may have acquired their product line.

 

An image search online shows that to be an accurate assessment. What I have matches the Ramrod XL exactly.

 

https://nextcloud.arkaic.com/nextcloud/index.php/apps/gallery/s/NMor4dw57Zk2BFE The unit in my system, you can see where the label was pushing against the RF shield and rusted it a bit.

Edited November 22, 2018 by kazriko