Barn find 800 has only one working key: Break

[Britishcar]

I've acquired what I would almost literally call a "barn find" 800. The poor thing was the filthiest computer I've ever encountered and I've seen quite a few. The inside was so crusty with crap that the power switch had something partially blocking it from switching on. Once I rattled that out of the way, the machine powered up perfectly to an incredibly clean and stable Memo Pad. The outside was no better with filth ground into it. The only working key was the BREAK Key which was 100% functional. The entire rest of the keyboard was 100% non-functional. Not even an intermittent click. RESET, OPTION and SELECT work too but alas START seems to be sleeping (tested using a Pac Man cart).

 

This 800 turned out to be a sort of "diamond in the rough" since after a FULL disassembly, a toothbrush, hot soapy water, a miracle sponge, etc. it looks almost new from '82! No cracks, scrapes, gouges not even discoloration! I think it was kept in a very, very dirty but very dark place.

 

Here's the problem: I gave the keyboard (a Hi-Tek) a full keys-off wash and clean along with everything else. I fully expected it to come to life but no luck. The exact behavior as before. I checked all of the traces for continuity (everything seems connected) but it refuses to do anything.

 

I know about the hitting keys over and over to get them to reconnect but I'm not sure that is the problem here since I cleaned just about everything including spraying contact cleaner on each of the little gold claws beneath the keys. BREAK, RESET, OPTION and SELECT continue to work and everything else is still dead.

 

I don't think it's POKEY as the computer plays games great, the sound it great, etc.

 

I know that the console keys are on a separate circuit to I will have to deal with the START key separately when or if I get the main keys working.

 

The only keyboard matrix layout I can find applies to the 800XL, not the 800. And I cannot find anything that discusses the keyboard pin connector layout (the rainbow one in this case) that connects right under the 800's keyboard to the motherboard.

 

Am I missing something here? Does anyone have any advise on what this might be? I feel as if I'm missing something very simple here.

 

Thank you!

[+kheller2]

This might help, the hardware manual section 3: http://atariage.com/forums/topic/127483-atari-os-and-hardware-manuals-get-them-here/

Also perhaps a better image of the 800 keyboard connector: http://atariage.com/forums/topic/177885-atari-800-ram-selection/?p=2243860

[Britishcar]

Ok, I've got a bit deeper and here's what I've learned (I've sprayed the hell out of the switches with contact cleaner and they look great...but...?).

 

Popping off the keycaps, you can access the little gold "claws" that close when the key/switch is pressed.

 

Using a multimeter and leaving the claws separated (key not pressed), you can touch the positive side of a multimeter to the right claw and the negative side to the left claw.

 

Doing this key by key gives a voltage reading on the multimeter of between 4.2 and 4.4 DC Volts. I don't know if this is normal but it's fairly consistent.

 

ALL of the keys register this voltage EXCEPT 6, 7, Y, U and N. These five keys are "dead" -- no voltage touching the "claws."

 

Despite ALMOST all of the other keys showing 4.2 - 4.4 DCV, only BREAK continues to work perfectly. Everything else is dead.

 

Well, at least this is more information. I'm not sure what to do with it.

 

Help!?

 

Thank you!

[Kyle22]

Have you tried gently removing and re-seating the POKEY and the 2 4051 keyboard multiplexer chips?

[Britishcar]

Kyle22, I'll try that. I know where POKEY is, but I don't know where the 4051's are offhand. I'll search online. If you know a schematic link, please let me know.

 

Thank you!

[+David_P]

Kyle22, I'll try that. I know where POKEY is, but I don't know where the 4051's are offhand. I'll search online. If you know a schematic link, please let me know.

 

Thank you!

 

More info on this, including the location of the 4051s, in this thread: http://atariage.com/forums/topic/152687-800xl-dead-keys/

[Britishcar]

Thanks, Kyle22 and David_P! I located the pair of 4051's and swapped them. I also pulled POKEY up and sat it back down. All pins were shiny clean, sockets clean, etc. All three chips went down nice and crunchy and were a tight fit.

 

And...nothing. No change. So perhaps I'll order a pair of 4051's and see if they are both blown for some reason...?

 

Are all 4051's created equal or does the 48K 800 use a particular type? Several have different 2 letter prefixes...?

[Kyle22]

CD4051 is a common one.

[Britishcar]

Thanks Kyle22. I have two on order. It's amazing how this 800 cleaned up. It's actually one of the nicest machines I've come across whereas before it looked like something rats had been living in for decades. What a change. I'd really like to bring it fully back to life.

Edited March 31, 2015 by Britishcar

[Kyle22]

When you get your 800 fixed, check out the Incognito board. There have been 3 production runs, probably won't be a 4th, unless lots of people beg for it, anyway, here is info on it: http://atariage.com/forums/topic/199490-incognito-board-preoder/?hl=incognito

 

I have one, as do many others here, and it is truly awesome

Edited March 31, 2015 by Kyle22

[Britishcar]

Arg! I swapped in two 4051's from a working donor A400 to no avail!

 

So, I've swapped a known working keyboard and known working 4051's with the same results (the 400 was fine): every key dead except BREAK. Others may be working such as SHIFT and CTRL but I can't tell without another key working to confirm.

 

POKEY seems great (plays music, sounds fine, etc.). Is it possible that part of POKEY could fail or should I be looking for something else on the motherboard? A blown resistor or capacitor?

 

I found this user with the EXACT same problem but no final solution is ever discussed on the thread:

http://forum.digitpress.com/forum/showthread.php?143398-Atari-800-Keyboard-issue

 

Since this thread was from 5 years ago and he last posted 3 years ago, I'm not sure I can contact him any longer for an answer.

 

Thanks to anyone in advance for any help you can provide!

[Kyle22]

It could be only the keyboard part of POKEY that's bad. Resistors and non-electrolytic caps don't blow very often. I had a 1200XL that SIO suddenly quit on. Keyboard and sound were fine. I tried the PIA first, but it was POKEY that was bad.

 

Try the POKEY from the 400 and see if that makes a difference.

[Britishcar]

Kyle22, you've been very helpful. I'll try the 400's POKEY and see what happens.

[Britishcar]

No joy. The POKEY transplant from the 400 to the 800 yielded nothing. Same dead keyboard with only BREAK working. So the issue is somewhere not related to 1) the keyboard, 2) the 4051's nor 3) POKEY.

 

I'm lost.

[+DrVenkman]

It might be something as simple as a bad keyboard connector on the 800 chassis. Maybe one of those pins is damaged or has a broken solder trace.

[Britishcar]

That's a great suggestion. I'll try to see if I can trace pin by pin from the connector back through the motherboard to see if I can find something wrong.

[ClausB]

Did you wash the circuit board itself? If so, did you rinse and dry it thoroughly before use? If not, could some dirt be trapped under a socket?

 

Since the break, shift, and control keys are multiplexed into the KR2 signal, and all other keys are multiplexed into the KR1 signal, I would suspect a problem with KR1 or with the Z104 chip.

 

Remove POKEY and the 4051s from their sockets and use an ohmmeter to check continuity between all the ground pins at those sockets. Then do the same for all the 5V Vcc pins. Then check KR1 continuity between POKEY pin 25 and Z104 pin 3 and pullup resistor R128. Also check R128's resistance (47K) and that its other end connects to Vcc.

[Britishcar]

Thank you ClausB. The machine fired up untouched and still dirty with the keyboard issue the first time I opened it out of the box. I did wash the motherboard but with no change in functionality. However, some bit of hidden dirt could have been there the entire time. The motherboard looks very clean indeed but of course that means nothing.

 

I believe I can follow your advise above. I won't be at the machine again until Monday.

 

The only thing I'm not clear on is the "Vcc" nomenclature. What motherboard structure are you referring to? Another chip or a set of lines or the connector itself?

 

Thanks to everyone giving this great advise! I will update you all when I know more. I feel like this is becoming an entire community helping to revive this old girl!

[ClausB]

Vcc is the techy name for the +5V supply. It's on pin 16 of the 4051s and on pin 17 of POKEY. Good luck.

[Britishcar]

I MAY have found something but need advice. If you can follow the thread above and read the progress below and know something, please let me know!

 

1) POKEY pin 17 is +5V

2) Both 4051's show +5V on pin 16

3) R128 has a perfect 47K Ohm across itself

4) Grounds and +5's between POKEY and both 4051's all have continuity with 0 Ohms.

 

Here is where I found a discrepancy. R128 runs north/south with the joystick ports pointed at you, just to the right of POKEY.

 

5) POKEY pin 25 (KR1) shows a connection to R128 this way: north side of resistor 0 Ohms, sound side (closer to joysticks) 46.9K Ohms

6) POKEY pin 16 (KR2) shows a connection to R128 this way: north side of resistor 94.7K Ohms, sound side (closer to joysticks) 47.6K Ohms

 

I don't know if this is normal?

 

A further discrepancy:

 

7) Z104 (the right-most 4015) pin 3 shows 0 Ohms on the north side of R128 and 46.9K Ohms on the south side

8 ) Z103 (the left-most 4015) pin 3 shows 61.6K Ohms on the north side of R128 and 12.84K Ohms on the south side

 

Further info. All long motherboard keyboard pins tested left to right (1 to 18?) while powered and keyboard disconnected. I don't know what these values should be but I think it is normal for 18 to have 0v. No?

 

Pin 1 = 12.1v Pin 2 = 13v Pin 3 = 12.9v Pin 4 = 13v Pin 5 = 12.5v Pin 6 = 12.5v Pin 7 = 11.5v Pin 8 = 11.2v Pin 9 = 4.97v Pin 10 = 4.75v Pin 11 = 4.77v Pin 12 = 4.77v Pin 13 = 4.79v Pin 14 = 4.8v Pin 15 = 4.79v Pin 16 = 4.75v Pin 17 = 4.79v Pin 18 = 0v

Edited April 7, 2015 by Britishcar

[ACML]

Britishcar,

 

You have quite the adventure on your hands. It can be great fun to fix and restore these wonderful machines. You seem committed to finding the culprit, which is admirable and can prove to be a challenge. There are schematics out there that isolate the keyboard circuit. See the top center J106. If you give it the old college try and decide to replace, I'm sure you are aware a new tested populated 800 motherboard can be had for $20. I like the challenge, and its half the fun, getting a broken machine working again that is. Sometimes even after exhaustive effort, victory still eludes us. At least it broke hard and smoking gun is easier to fix than an intermittent issue. So at least you have that going for you. I'll be watching the thread for progress and look forward to your triumph.

[Keatah]

All the crusted crud and dirt.. it acted as a protectant against the elements!

[ClausB]

1) thru 7) all normal. might be normal but what counts is whether Z103 pin 3 is connected to ground. It should have 0 ohms to ground. Your voltages on keyboard pins 9-17 look good but 1-8 do not. They should be near 0V. Are you sure your meter was not reading mV?

 

With chips in and power on, touch a wire to any keyboard pin 1-8 and touch the other end of the wire to any keyboard pin 10-17, and see if a character appears on the screen and you hear the speaker click. If not, try another pair of pins (not all 64 combinations are valid keys). If you get characters then the circuit is working and you should check key continuities on the keyboard itself.

[Britishcar]

Thank you to everyone. You are all helping me to move along on this challenge.

 

ClausB, particular thanks to you for guiding me down new paths to test. And thanks to ACML for the schematic. Thanks to Keatah for the laugh.

 

I have a couple of definite updates:

 

To address ClausB's ideas:

 

1) Z103 pin is well connected to ground, i.e. it has 0 Ohms.

2) I was WRONG on keyboard pins 1-8 as ClausB predicted. My meter was reading in mV and I didn't notice the auto-switching from the V readings on the other pins. So 1-8 are near 0 and the others are very near 4.8V so all good to go.

3) ClausB's brilliant idea to touch motherboard keyboard pins together with a simple wire yielded great results: lots of healthy 800 speaker clicks with lots of random keystrokes appearing for the first time (for me anyway) in Memo Pad. Upper, lower, inverse, space, click click click.

 

However, a known working keyboard yields nothing but BREAK as usual. How is this even possible? Isn't the keyboard just creating shorts as well?Arg!

 

INTERESTING FIND: I found a large discrepancy between resistance output from Z103 ground and R128 on the troubled motherboard and a known working motherboard.

 

Here are the details:

 

Known working motherboard:

 

Z103 PIN 3 TO R128 northside lead (closer to cart slots) 93.7 K-OHMS

Z103 PIN 3 TO R128 southside lead (closer to joysticks) 48.3 K-OHMS

 

Troubled motherboard:

 

Z103 PIN 3 TO R128 northside lead (closer to cart slots) .89 M-OHMS

Z103 PIN 3 TO R128 southside lead (closer to joysticks) .82 M-OHMS

 

Note the swap to mega Ohms from Kilo Ohms. I assume this may the the core of the problem somehow? If so, why would shorting keyboard pins together still produce characters?

 

HAL 9000: My mind is going. I can feel it.

 

Thanks to anyone who can help.

[Kyle22]

It looks to me (from the fuzzy schematics) like you should measure pin 3 of Z103 to ground, should be 0 ohms. The 800XL schematic (which is clear) shows this pin going through 100 ohm resistor to ground.

 

If you have any more than 100 ohms (with chip removed from socket), then you may have a bad solder connection, or a broken trace.

 

Pin 3 of Z104 should have 0 ohms resistance to PoKey socket pin 25 (with PoKey removed), and they both should read 47K ohms to +5.

 

Good luck.