Atari 130XE boots to Self Test, does not test

[Wumbologist]

Thought I'd ask here before attempting to desolder a >30 year old board.

 

I've got an Atari 130XE that shows a blank blue screen on boot for a few seconds before going straight to the self test menu but doesn't start any test. All buttons on this menu are unresponsive. (video)

 

The only cart I have is an UnoCart, which the 130XE will boot to instead of the self test and everything seems to be working fine, keyboard included.

 

I uploaded Basic onto the UnoCart and am able to launch and use Basic in this way, so I guess that's a workaround for now if I really want to use Basic. From here, I can type in the "BYE" command to launch the self test manually, but unfortunately in this application all keys are unresponsive still.

 

One possible culprit is the infamous "Ingot" which is the power brick that came with it. The first few times I powered the 130XE on, it booted to Basic as normal until one time it started booting to self test instead and has ever since. I've since replaced my Ingot with a newly-built power supply, so my Ingot shouldn't be damaging any more Ataris in the future.

 

Another possible culprit is my careless use of the video out. I bought a cable that splits the output to composite, luma, chroma, and audio. The connectors seem like they're colored all wrong, so I was pretty confused when plugging it into my TV and got the wires crossed a few times. I didn't think that this could hurt my 130XE at all, but I might've been wrong.

 

I took the thing apart to clean and inspect the insides. The most notable thing I found is that there's these two loose wires going across the board. Looks quite unprofessional, why not just put this lead in the PCB? Moreover, the pins they're connected to are soldered together on the underside.

 

Any idea what's going on?

[Roydea6]

Nothing is wrong plug it for about 4hrs to normalize all the electronics. the two wires are factory installed. the Video plug may have the wrong id on the cable if so just take off and use the try each RCA in one TV RCA plug and isolate , Sound, Chroma, Luma, Composite and re label the wires..

[Sugarland]

Also check the voltages on the PSU. You can use heat shrink tubing to keep the probes from touching other pins.

[1050]

Ingot is not the problem, those don't boot at all.

Didn't cause the problem either since there likely

isn't anything wrong with ingot except for the potential

issues.

 

I'm suspecting a bad OS ROM chip because of the

flaky way it does work. Not a normal situation ever

is what you are presented with yet most of it works

well enough to be recognized. MMU might be a fault

also but my money is on yet another OS ROM chip has

been found to be slightly insane.

 

Very common amongst XEGS these days and good bit more

rare for 130XE, it can still happen from time to time.

[xrbrevin]

I think the ingot got a bad rep in the US but the European ones dont seem to suffer the same issue but saying that, modern equivalent PSUs are cheap and reliable

If you have access to an EPROM burner, the 27C128 IC is pin-compatible with the OS ROM. Search the forums here or the following page for OS ROM image files and take your pick http://ftp.pigwa.net/stuff/collections/atari_forever/ROM/Rom%20-%20OS/

[+Nezgar]

As xrbrevin said, a 27C128 EPROM is a drop in replacement. If you're interested and don't have a programmer, send me a PM as I can mail you a programmed chip for my cost. (S/b less than $5 mailed) And since it's a 130XE you'll probably need a socket for it too which I can include.

[Wumbologist]

Nothing is wrong plug it for about 4hrs to normalize all the electronics. the two wires are factory installed. the Video plug may have the wrong id on the cable if so just take off and use the try each RCA in one TV RCA plug and isolate , Sound, Chroma, Luma, Composite and re label the wires..

Left it running for about 4 hours like you said but no dice. It was interesting to see my Basic prompt cycle through a bunch of colors, didn't know old 8-bit machines like this had screensavers. Those labels I put on the cable are mine and are correct, it was the colors of the connectors that confused me. Like composite being red and audio being yellow.

 

Also check the voltages on the PSU. You can use heat shrink tubing to keep the probes from touching other pins.

Heat shrink! That's so obvious, that would've made it way easier to test. From what I can tell, my Ingot is fine. The voltages are okay and it powers the 130XE on just like my new power supply. Though, I've also been told the Ingot isn't very trustworthy and can start and stop working on a whim.

 

Ingot is not the problem, those don't boot at all.

Didn't cause the problem either since there likely

isn't anything wrong with ingot except for the potential

issues.

 

I'm suspecting a bad OS ROM chip because of the

flaky way it does work. Not a normal situation ever

is what you are presented with yet most of it works

well enough to be recognized. MMU might be a fault

also but my money is on yet another OS ROM chip has

been found to be slightly insane.

 

Very common amongst XEGS these days and good bit more

rare for 130XE, it can still happen from time to time.

Really? Ingots just sometimes don't deliver power? I was under the impression they could kill Ataris. The OS ROM going wrong makes sense.

 

I think the ingot got a bad rep in the US but the European ones dont seem to suffer the same issue but saying that, modern equivalent PSUs are cheap and reliable

If you have access to an EPROM burner, the 27C128 IC is pin-compatible with the OS ROM. Search the forums here or the following page for OS ROM image files and take your pick http://ftp.pigwa.net/stuff/collections/atari_forever/ROM/Rom%20-%20OS/

I don't have an EPROM burner, but I've got about six other non-working vintage machines sitting in my closet that I haven't gotten to yet, so that might be a good investment.

 

As xrbrevin said, a 27C128 EPROM is a drop in replacement. If you're interested and don't have a programmer, send me a PM as I can mail you a programmed chip for my cost. (S/b less than $5 mailed) And since it's a 130XE you'll probably need a socket for it too which I can include.

PM'ed

[1050]

Really? Ingots just sometimes don't deliver power? I was under the impression they could kill Ataris. The OS ROM going wrong makes sense.

2nd try

Ingot is not the problem, ingots with problems don't

boot at all. They just kill Atari without you having

a clue that a murder just took place. No screen, nothing.

Ingot will never recover on it's own and neither will the

Atari.

 

You have a good screen, so ram is still good and not

shorted. Ingot is fine still, OS kinda works but it

kinda doesn't at the same time. Usually a more basic

fault will rip the screen up or give red, green screen,

etc. Just a feeling I have that makes me point at the

OS but we all play this game. Only the very clever

snag the proper clue and call it right, I see flashjazzcat

is reading this thread - he's one of those clever ones.

 

OS seems to be the agreed highest likely fault so far

and maybe too obvious to start feathering my cap for

the victory lap. We are for sure waiting for more clues

after the OS chip change because too often it's not the

obvious. And better clues help immensely.

[_The Doctor__]

Actually, ingots can produce humm bars and distortions on your CRT letting you know it's all going wrong and damage is occurring. Sometimes the gremlins like to torture their prey!

[Wumbologist]

Alright so I gut a 27C128 OS replacement chip from Nezgar, was finally able to socket it in, and there's absolutely no change. So I guess it's safe to say that the OS is fine.

[+Nezgar]

So looking back at your original post and video, you show that the system can boot to self test, and can also boot to BASIC from your UNO cart, and type BYE<Return> also takes you to self-test, but no function keys respond. Once in self test either you stated that function keys do not work.

 

Normal keyboard typing is handled by the POKEY chip via 2 4051's, so that part would seem OK, but the function keys are handled by the GTIA. Booting to Self-Test is also the behaviour if the system is powered on with OPTION key pressed, with no disk drive turned on.

 

If the system thinks the OPTION key is stuck, can you try turning on the computer with the keyboard ribbon cable completely disconnected, to see if you get to the BASIC READY prompt, and not the Self-Test?

 

To confirm the (mis)function of the function keys from BASIC, run the following program:

 

10 ? PEEK(53279):GOTO 10

 

You should see changes in the number presented for all various combinations of Start/Select/Option/Nothing pressed. (0-7) Check if you see no change on any of the keys. I think this value should also not change just by disconnecting/reconnecting the keyboard connector while it's running.

[Wumbologist]

With the keyboard unplugged it boots to self test just as before.

 

I ran your program and it just prints 1, no matter what combination of function keys I press. Even if I unplug the keyboard it still just prints 1.

[+Nezgar]

OK, so I just tested this on my own machine, and 1 means the computer thinks you have both SELECT & OPTION pressed. No keys depressed = 7

 

So given it is the same even with the keyboard physically disconnected, that indicates something on the motherboard, or your GTIA chip.

 

So I think first order of business is getting your GTIA socketed, and test the IC in another machine. Maybe just checking continuity / reflowing the solder on the existing pins on the solder pads will help, or as a positive result of socketing it, but I'm not sure if a 'not connected' pin results in the chip thinking the key is depressed? (Maybe someone can chime in about that)

[ChildOfCv]

The GTIA technical documents say that the chip connects each S port to a pull-up resistor for input. So it should read high when nothing is connected.

 

Just for fun, try adding the following line to the test program above:

 

5 POKE 53279,0

 

Just to make sure it's not that something mis-programmed the register.

 

Also try

5 POKE 53279,15

 

to see if it makes a difference.

 

 

Just a note, but if anybody has the original GTIA internal schematics on paper and would be willing to subject them to another scan that has better resolution and doesn't look like a fax of a fax, that would be quite beneficial.

[+Stephen]

The GTIA technical documents say that the chip connects each S port to a pull-up resistor for input. So it should read high when nothing is connected.

 

Just for fun, try adding the following line to the test program above:

 

5 POKE 53279,0

 

Just to make sure it's not that something mis-programmed the register.

 

Also try

5 POKE 53279,15

 

to see if it makes a difference.

 

 

Just a note, but if anybody has the original GTIA internal schematics on paper and would be willing to subject them to another scan that has better resolution and doesn't look like a fax of a fax, that would be quite beneficial.

POKing to that register will not change its value upon PEEKing it. It will make the GTIA produce a click. Bit bang that register fast enough, and you get sample playback.

[Wumbologist]

Socketed the GTIA. No changes. I also re-socketed the small chip to the left of it which came to me already socketed. I doubt this is from the factory though because there were scorches all over the pins on the underside. It's cleaned up now and looking much better, but still doesn't work.

 

Also tried adding ChildOfCv's suggested lines to the program. No change.

 

Completely unrelated to the topic, but does anyone know where I can find screws like the ones found in these old Ataris? I've got a few with missing screws. None of my local hardware stores have anything like them. Machine screws are small enough but the threading is too narrow. Wood screws come closer to having the right threading but they're always too long.

[+Stephen]

Socketed the GTIA. No changes. I also re-socketed the small chip to the left of it which came to me already socketed. I doubt this is from the factory though because there were scorches all over the pins on the underside. It's cleaned up now and looking much better, but still doesn't work.

 

Also tried adding ChildOfCv's suggested lines to the program. No change.

 

Completely unrelated to the topic, but does anyone know where I can find screws like the ones found in these old Ataris? I've got a few with missing screws. None of my local hardware stores have anything like them. Machine screws are small enough but the threading is too narrow. Wood screws come closer to having the right threading but they're always too long.

See the message above. You cannot write to that register and effect a change. Many of the registers in the custom chips have 2 different behaviours, depending on whether or not you read from them, or write to them.

[Wumbologist]

Okay so I have another 130XE that will power on (red light on keyboard) but won't display anything. I tried swapping the GTIA chip from it into this one we've been troubleshooting and now it works perfectly. Boots right to BASIC. I can run the self test and it comes back without any issues.

 

There is one thing I noticed though and that's that the output picture is a bit more green than it was with the original GTIA. For example, the BASIC terminal looks teal instead of blue.

[+kheller2]

Okay so I have another 130XE that will power on (red light on keyboard) but won't display anything. I tried swapping the GTIA chip from it into this one we've been troubleshooting and now it works perfectly. Boots right to BASIC. I can run the self test and it comes back without any issues.

 

There is one thing I noticed though and that's that the output picture is a bit more green than it was with the original GTIA. For example, the BASIC terminal looks teal instead of blue.

 

Great to hear! Seems that the GTIA was bad. As for the color, adjust the color POT.

[ChildOfCv]

POKing to that register will not change its value upon PEEKing it. It will make the GTIA produce a click. Bit bang that register fast enough, and you get sample playback.

 

The GTIA manual says differently: Poking 1's to it turns it into an open-drain that pulls the pin low. Of course then it will definitely read low. Poking 0's to it pulls up the resistors, so then reading from it will read a high. It's a slick way to make the ports configurable simply by programming the right value and adding the correct external logic to support the intended direction.

 

So the experiment was to see if someone misconfigured the port. But apparently not.

[Wumbologist]

Ah I didn't even know that existed. That's cool.

 

So that's one 130XE down. Someday I'll get to repairing my other. Anyone know what chips are compatible replacements for the GTIA that I can flash with my TL866?

[Rybags]

GTIA isn't a Rom, it's a custom processor chip. There's been modern day implementations done but it's just easier to get a legacy one from a supplier.

[+Nezgar]

The GTIA manual says differently: Poking 1's to it turns it into an open-drain that pulls the pin low. Of course then it will definitely read low. Poking 0's to it pulls up the resistors, so then reading from it will read a high. It's a slick way to make the ports configurable simply by programming the right value and adding the correct external logic to support the intended direction.

 

So the experiment was to see if someone misconfigured the port. But apparently not.

I just tested this on a real machine.

 

5 POKE 53279,0

10 ? PEEK(53279):GOTO 10

 

indeed makes a click, but the next PEEK is 15, then the next all consecutive peeks then resume reading the function keys as normal.

 

5 POKE 53279,15

10 ? PEEK(53279):GOTO 10

 

With this, first PEEK returns 0, then resumes reading function keys as normal.

 

Does this mean that a program bit-banging the speaker might have issues trying to simultaneously trying to read the function keys? It would have to wait a small period of time before being able to properly read them again..

[+DrVenkman]

I just tested this on a real machine.

 

5 POKE 53279,0

10 ? PEEK(53279):GOTO 10

 

indeed makes a click, but the next PEEK is 15, then the next all consecutive peeks then resume reading the function keys as normal.

 

5 POKE 53279,15

10 ? PEEK(53279):GOTO 10

 

With this, first PEEK returns 0, then resumes reading function keys as normal.

 

Does this mean that a program bit-banging the speaker might have issues trying to simultaneously trying to read the function keys? It would have to wait a small Herod of time before being able to properly read them again..

There’s a thread here in the forum from a couple years ago with a program demonstrating how to play music from bit-banging the GTIA. Pretty sure the machine isn’t capable of much else while doing so. I think the player read the regular keyboard buttons or something to pause or stop playback. Can’t remember for sure. Fun trick but probably not terribly useful.

[ChildOfCv]

Yeah and the other thing is, if you only want to bit-bang the speaker, use values of 0 and 8 instead. This keeps the other 3 pins as readable inputs.