Atari 800 keyboard stopped working

[PoDuck]

I managed to get the atari 800 I had when I was a kid from my mother's house, after she passed away.  It had a bad space bar from way back, and my little brother mangled the little metal fingers under the key trying to fix it.  Long story short, I replaced the keyboard with a new one.  It worked great for about 10 minutes, then suddenly no keys are working.  Putting the old keyboard back in didn't clear up the problem, so I have to assume it's something on the motherboard.

 

After doing a bit of research, it seems like the Antic chip might be the issue, as absolutely none of the keys work, but I don't know how to diagnose it for sure.  Further, even if I decided to buy one of those fast disappearing chips to replace it with, what caused it to go out in the first place?  It seems rather coincidental that it goes out 10 minutes after replacing the keyboard, after surviving for 30 years with a broken space bar, and a year of use since I got it back, using it with the broken space bar.  I'd like some ideas on what to test to maybe figure out why it would have gone out.

I appreciate any help someone can give me.

[griff3125]

That's very odd, it's a long shot but remove all but one of the memory boards, and swap it with another to be sure.

Also try cleaning the ROM board contacts.

That machine's a tank, I'd be very surprised if you blew a chip just swapping the keyboard.

 

[Panther]

CD4051 chips?

 

[PoDuck]

10 minutes ago, Panther said:

CD4051 chips?

 

Those are just jelly bean multiplexers.  I had a few of them and just swapped them both out.  I also bread boarded them, and they tested out fine.  It was a long shot anyway, since all the keys stopped working, which would mean both died at the same time.

[PoDuck]

59 minutes ago, griff3125 said:

That's very odd, it's a long shot but remove all but one of the memory boards, and swap it with another to be sure.

Also try cleaning the ROM board contacts.

That machine's a tank, I'd be very surprised if you blew a chip just swapping the keyboard.

 

I'll try that tomorrow.

[Panther]

26 minutes ago, PoDuck said:

It was a long shot anyway, since all the keys stopped working, which would mean both died at the same time.

I learned not to discount long shots, but sorry it wasn't something so simple.

 

[DavidMil]

6 hours ago, PoDuck said:

After doing a bit of research, it seems like the Antic chip might be the issue, as absolutely none of the keys work, but I don't know how to diagnose it for sure.  Further, even if I decided to buy one of those fast disappearing chips to replace it with, what caused it to go out in the first place?  It seems rather coincidental that it goes out 10 minutes after replacing the keyboard, after surviving for 30 years with a broken space bar, and a year of use since I got it back, using it with the broken space bar.  I'd like some ideas on what to test to maybe figure out why it would have gone out.

I appreciate any help someone can give me.

Have you checked the Pokey Chip?  That is where the outputs of the two 4051's go.

 

DavidMil

[PoDuck]

9 hours ago, DavidMil said:

Have you checked the Pokey Chip?  That is where the outputs of the two 4051's go.

 

DavidMil

That's right.  I meant pokey.  Sorry for the confusion.

As for checking it, how do I test it?

[PoDuck]

Okay, I contacted a guy here that deals in old arcade machines.  He also repairs them.  He gave me a pokey chip that he had gotten by mistake long ago that should be the correct one for the home computers.  I replaced it, and it seems to be working mostly correctly, other than some odd behavior.  I have an sio2sd, where you should be able to long press x to get it to boot the image, but instead of booting, I just hear it beeping on repeat, as though I am pressing the key multiple times.  It doesn't boot the image, and the only way I am able to get it to boot is to press the fire button on the joystick.  Also, if I type too fast, and press a second key before releasing the first, the second key won't register.  I don't remember this being the case previously.  Is this a sign that it isn't the right chip?  I know they made multiple pokey chips, and that they were different for the arcade machines than they were for the home computers.

[DavidMil]

C012294 is the Atari part number for this chip.  If it has anything like a dash number after this number it may be one of the multicore versions

that was made for arcade games.  I understand that they went all the way up to a quad core version (CO12294B-4).  I have never seen

anything higher than a dual core version of the POKEY chip.  I have no idea of how these chips would react in an 8bit computer.  Sorry.

 

DavidMil

[+slx]

Don‘t throw away the old POKEY. It might still work as a second POKEY for stereo or in a 7800.

[Sugarland]

PoDuck, single core POKEY's are the same between arcade machines and the home computers.  It could be a failing socket or cracked solder joint or corrosion on the replacement POKEY's legs.

[PoDuck]

On 6/6/2021 at 1:31 AM, DavidMil said:

C012294 is the Atari part number for this chip.  If it has anything like a dash number after this number it may be one of the multicore versions

that was made for arcade games.  I understand that they went all the way up to a quad core version (CO12294B-4).  I have never seen

anything higher than a dual core version of the POKEY chip.  I have no idea of how these chips would react in an 8bit computer.  Sorry.

 

DavidMil

They were both the -3 versions.

[PoDuck]

On 6/6/2021 at 1:39 AM, slx said:

Don‘t throw away the old POKEY. It might still work as a second POKEY for stereo or in a 7800.

I didn't think about that.

[PoDuck]

On 6/6/2021 at 1:39 AM, Sugarland said:

PoDuck, single core POKEY's are the same between arcade machines and the home computers.  It could be a failing socket or cracked solder joint or corrosion on the replacement POKEY's legs.

I'm considering replacing the sockets on this thing with machined sockets.  The legs on the chip don't look too bad.  They are all nicely tinned, with no signs of oxidation.  This computer lived in the desert all its life, so that helps too.

[PoDuck]

Another question for everyone.  Does anyone know if the Pokey chip determines a long press from key repeat?  It seems like that could be the same thing to be determined in software instead.

[PoDuck]

I found this datasheet, which makes me wonder if it may be the debounce circuit that is having trouble.  It seems that would be an internal issue to the chip, and not an external issue.  It appears that the debounce circuit has a comparator that detects if there is a new key being pressed.  Given the behavior I am seeing, I have to wonder if that comparator is not storing the value of key being pressed, it would always think it has a new key pressed at every HSYNC, and after 4 HSYNCs, it passes off the pressed key as a new keypress.  It also ignores it if two keys are pressed simultaneously, which would explain why it doesn't regester a new key, and halts any keyboard recognition when I press a new key while holding down the old key.

Does this sound about right?

[Faicuai]

5 hours ago, PoDuck said:

Does this sound about right?

Read the thread, here... 

 

...In practical terms, it will take quite longer to troubleshoot what seems a strange problem, without additional / stand-by parts.

 

I am not sure if you have additional parts, but I do have an extra 800 mother-board you can have for this purpose, no worries. Just PM for details, and hopefully you may be able to get to the bottom of this rabbit-hole much faster.

 

Good luck!

[PoDuck]

11 minutes ago, Faicuai said:

Read the thread, here... 

 

...In practical terms, it will take quite longer to troubleshoot what seems a strange problem, without additional / stand-by parts.

 

I am not sure if you have additional parts, but I do have an extra 800 mother-board you can have for this purpose, no worries. Just PM for details, and hopefully you may be able to get to the bottom of this rabbit-hole much faster.

 

Good luck!

I appreciate that, and I will keep it in mind.  I have a working 800xl with a good pokey chip in it, I'm just leery of frying it, considering I don't know what killed the original in the 800.  Also, I'm considering putting the original Motorola 4051 chips back in it, since they didn't fix the problem to begin with, and one of the issues I'm having could conceivably be caused by a fault in the multiplexers, namely the situation where it doesn't seem to like multiple keys pressed at the same time.  I don't see any differences that would matter between Motorola MC14051Bs and Texas Instruments CD4051Bs on their datasheets, but there could be, and it couldn't hurt.

The major problem is, it's getting a bit hot, or at least humid, and the air conditioner is out in my lab, so I only work in there in the mornings.  I won't know if that fixed anything until then.  I'm also going to try and do some probing with my scope and see what I find.  I now know what pins I should expect to see any signals on.  I just wish I had an MDO.  Sadly, I only have a lowly oscilloscope and a logic probe.

[Sugarland]

Try the 800's pokey in the 800XL.

[PoDuck]

2 hours ago, Sugarland said:

Try the 800's pokey in the 800XL.

Not a bad idea.

[+DrVenkman]

8 minutes ago, PoDuck said:

Not a bad idea.

That's actual standard troubleshooting practice - try the parts FROM the suspect machine IN the good one; that way, if something external to that part damaged the part to begin with (like a short somewhere), it won't damage known-good parts that you install in the bad machine. If the POKEY from you 800 works in the 800XL (all CO012294 are completely interchangeable, whether in an arcade machine, a 5200, a 7800 BALLBLAZER or COMMANDO cart, or any A8 computer), then you will have the unenviable task of tracing the relevant signal paths from the the keyboard lines on the POKEY chip from the pins on the chip all the way to the keyboard to verify they're good. More realistically, you can check for activity with a meter, logic probe or scope to find any bad signals or simply "brute-force it" by replacing the socket. Socket problems are more common than people realize, btw. 

[DavidMil]

23 hours ago, DrVenkman said:

If the POKEY from you 800 works in the 800XL (all CO012294 are completely interchangeable, whether in an arcade machine, a 5200, a 7800 BALLBLAZER or COMMANDO cart, or any A8 computer)

So a dual or quad core POKEY can be substituted in place of a standard single core POKEY?  I wondered about that...

 

DavidMil

[+DrVenkman]

1 minute ago, DavidMil said:

So a dual or quad core POKEY can be substituted in place of a standard single core POKEY?  I wondered about that...

 

DavidMil

No, that's not what I said. CO12294 chips are all standard, interchangeable POKEY chips. The arcade division dual- and quad-core POKEYs have different part numbers. But many millions (tens of millions?) of ordinary single-chip POKEYs were used in arcade machines, which is where the impetus has come in recent years for some of the FPGA replacement projects. Folks like you and me can go to Brad at Best, hat in hand and wallet extended, for that one chip we need to fix a treasured machine. Guys who own an arcade and need 10 of them to maintain their cabinets have to go elsewhere.