My 800's keyboard is dead

Well, as those of you have been following my hardware hacking blog know, I stopped my upgrade work on the 800 because the keyboard wasn't working and I thought I had done something to the POKEY or 4051 IC's which handle it. After extensive trouble shooting and replacement of those IC's and lots of tracing circuits with my multimeter to figure out what went wrong, I have finally discovered that NOTHING went wrong with the upgrades and working on the motherboard itself. It all tests good either in stock configuration or with the Incognito plugged in, it works. Except for the keyboard.

When I was replacing plungers on the keys to fix it, I noticed about half a dozen cold/cracked solder points on the bottom side of the keyboard, so I resoldered them to fix them. Even though the keyboard was fully working before, I suspected those keys would stop working soon. Well, some how I killed the keyboard when I did it. I don't know how, as it's a very simple circuit with only traces and jumpers on the keyboard itself, and all the IC's that make it work are on the motherboard, so I'm not sure what went wrong.

The problem is, with this Hi-Tech brand keyboard, there is no way to open it up without probably destroying it anyway. The plastic part that the plungers and keycaps are attached too doesn't seem to have anyway to come off the circuit board except by desoldering the entire circuit board and then pulling it off. Which I may have to do, as nothing on the bottom side or with the keys themselves seems to be wrong, it just won't work at all, well, except for the BREAK key. It's the only key (besides the function buttons which are on a different circuit) that works.

I checked into getting another keyboard for it, but the only place that has one for sale is B&C Computervisions (MyAtari on Ebay) and they want $100 for it! I can buy an entire new motherboard for $40! So I wish now I had broken the motherboard as it would have been cheaper to replace!

Does anyone here have any experience working with these 800 Hi-Tech keyboards? It's such a shame as the keys all work nicely now with all good plungers, no cracked corners! I suppose I should try to find an 800 Matsumi keyboard, that they started using in later years, like on the 1200XL, they are wonderful, but apparently they are ultra rare as only about 1-2% of 800's had these installed at the end of the 800's production run.

 

The Hi-Tec keyboard says just that on the bottom side of it. There are 3, IIRC, versions or brands of the Atari 800 keyboards. The earliest models, like the one you gave me, is a Hi-tec (I mistakenly added and H to the spelling before) is mechanical. which means inside the key are two brushes that come together when a key is pressed completing the circuit. They have square white plungers under the key caps which are spring loaded.

StackPole keyboards were the second version, which are also mechanical and very similar to the Hi-Tec in operation, but the plungers are yellow instead of white, and this is the worst version of 800 keyboards as they are the most prone to having the corners of the plungers crack, so the keys are loose and don't work right.

The third type, and final, but also most rare, are the Mitsumi keyboards that are like 1200XL keyboards with a mylar sheet in between the keys and the circuit board, and use rubber plungers instead of springs. These are actually the best 800 keyboards, just like 1200XL's are the best keyboards among XL/XE machines. I think they say Mitsumi on the bottom of the keyboard, but also on the inner mylar I have pictured here.

This is the top of a Hi-tec showing the white plungers. The Stackpole brand, which I could not find an image of, look identical, but have yellow plungers.


Here is the inside of a key, showing the copper metal brushes. Stackpole keyboards are similar.



This is what the bottom of a Hi-Tec keyboard looks like, though this one has been repaired with those jumper wires; the original jumper wires are inaccessible under the keyboard, so this is the way one would have to repair it (like me probably). Mine doesn't even have the screws on the bottom like this one though, but apparently this person could not open his either, so he made repairs to the bottom of the circuit board instead, like I will have too.



This is what a Stackpole brand keyboard looks like on the bottom.




This is what the INSIDE of a Mitsumi keyboard looks like, with it's mylar.


B&C Computervisions (MyAtari on Ebay) ONLY have the worst, Stackpole brand left for sale, and they want $100. Not worth it at all, especially considering they are easily broken. So I guess I'll have to attempt repairs to my Hi-Tec brand.



Now about solder points...basically the solder breaks and separates the component (in this case a keyboard key) from the motherboard so they do not make contact, or poor contact with the board they were once securely soldered too. I found a picture of such an example on-line, since I already "fixed" the ones on my keyboard. The second picture is of various good and bad solder points.

 

Luckily I found a guide on AtariAge where a member had the same issue as me, so I can use his tips and guidance in my repairs.

https://atariage.com/forums/topic/222842-my-atari-800-keyboard-repair/?tab=comments#comment-2942105

 

That soldering is a tad worrying...I didn't know Stevie Wonder did Atari repairs...

 

Stackpole is the same type as the Oric Atmos keyboard and the contacts cause connection problems later on as the little leaf springs inside stop making contact. Irrelevant to fixing an 800 keyboard I know but another reason to not buy one, unless the high tech ones have the same problem with the similarities in their design but it would be very unusual for all of them to pack in at the same time.
Theoretically it must be either the tracks,links or contacts... have you measured the impedance when you press a key?

 

Not much help but good luck with getting a good fix...Still can't get over that line of solder blobs, a true who's who of how NOT to solder..

 

My thinking all along was it isn't the keyboard, since none of the keys were working (except the BREAK key). This is the Field Service Manual's way of thinking too. All the IC's concerned have been swapped several times with new or known working chips. The computer works, with Incognito installed or reverted too stock (as now), but with the Incognito and no keyboard it goes straight to SDX, you have to have a working keyboard to get to the BIOS menu or even type CAR to run a cartridge, so I have it reverted to stock until I figure it out.

But, there was the nagging issue of "why does the BREAK key work?" The FSM says if some keys work, but not all, it's the keyboard. If all keys don't work, check the Pokey and 4050 IC's which I have. I haven't gotten around to checking continuity yet, but I find it hard to believe that an entire row of resistors or capacitors before the connector would have all gone bad at once, except one. I was about to remove the sockets on the mobo for the 4050 IC's and change them out, and make sure that no rouge solder balls or bits of wire got underneath the sockets and are causing a short.

Maybe the POKEY as well, however, in stock condition I can run cartridges and auto-load disks, and they work fine and the Pokey sounds and music play if I can start the games with either a joystick button or one of the function keys not on the same circuit as the keyboard. So I don't believe the issue is immediately at the POKEY chip. But I will do a continuity check on the keyboard the best I can without an Oscilloscope as the FSM calls for, with my multi-meter next.

 

Isn't the BREAK key hardwired differently to the other keys.....I seem to remember something about that...Will have a dig...

Could only find this

http://krap.pl/mirrorz/atari/homepa.../Atari/800XL/atari_hw/pokey.htm#keyboard scan

 

But I've also been thinking about jumping back on my 1200XL, trouble shooting it to get a break from the 800. The 1200XL is a red screen issue, and the only things I haven't checked are the OS and ram chips (brand new, recently replaced after reverting back to 64K internal and using Syscheck's 512K). So I may go ahead and reverse the 32-in-1 upgrade in my 1200XL, back to the original 1200XL OS, since I do have the Syscheck with 3-4 other OS slots available anyway. And once I get it working that way, I can always burn the 4-in-1 eproms to replace the original OS again later (and get back BASIC or other rom internally too). And I have more 64K and 256K drams on hand, NOS, so if it is a ram issue...

 

If I'm understanding the FSM schematics and flow charts correctly, which I think I am, it is part of the same circuit. It's not considered a function key like the START/SELECT/OPTION and RESET are (also HELP on XL/XE's) which are on a different circuit and controlled by the PIA, IIRC, instead of the POKEY. But I do see what you mean, at the link, it does show BREAK/SHIFT/CONTROL separated from the other keys somehow. And I can't know if SHIFT or CONTROL work without the other keys or I suppose a continuity check.

If that is the case, and looking closer into it, it does look like the BREAK/SHIFT/CONTROL are routed through one 4050 IC and the other keys through the other. Which would follow my theory that a bit of solder or wire got caught under one of the 4050's sockets and is shorting things out. Under inspection without removal, my eyesight, even aided, can't see well enough, but I *think* I saw something...

 

I feel it is the keyboard and something on the PCB from what you described. If there is a problem there it could pull the impedance down to a point that keyboard presses don’t register. If the BREAK key still works and it’s on a separate circuit that might explain the anomaly.
It will get fixed, just need to set some time aside and work through it methodically.

 

!KR2 is directly connected to the POKEY based on that diagram so there is a little difference to the rest of the keys.

 

I'm liking my foreign object under a socket on a 4050 theory more and more, since the solder/resin/brushes/wire and solder iron had all been recently involved. I

also forgot to mention that there is no issue with my keyboard's circuit board like the guy that did the AA thread I link had. I carefully went over it and verified all good solder points, no solder pads pulled or cold solder joints, traces all intact. So it really makes me want to go back to my original theory that it's not the keyboard. It's such a simple on-off circuit with no IC's on the keyboards circuit board. I know all the brushes for the keys are good, even the few I had to bend back in place during repairs. all cleaned and shiny.

 

4050's are dirt cheap and common

 

@Paul "Mclaneinc" Irvine I know, I just bought a pack of 10 of them off Amazon after already trying the 6 I had on hand from other machines.

I was just reading the circuit diagram in the PDF FSM, and I do see that the two 4050's are tied in together, and do not operate with the keyboard in parallel. So if one is out, the other will be too, no half keyboard working if it goes through either of them. And of course what @Vyper68 said, the BREAK key runs directly from the POKEY with the other two keys.

So checking continuity will be the almost next step, right after I lift at least one of the 4050 sockets and make sure there is no short.

 

Wish I could offer more help Matt, you are miles ahead of me in circuit checking, the reference to the BREAK key I made was just one of those little facts you store away and it comes in handy sometimes. As for my knowledge of the 4050's, I seem to remember them to be gate arrays but this is stuff in my head from the 80 and 90's, I'm simply remembering the picture of the catalogue page from Maplin. I used to have a photographic memory but all that went south after my mental health issues, some is still there but to be fair I just sold them or soldered them in to preexisting kits, their actual use is foreign to me

To quote a favourite old phrase I heard once, "I was born with nothing and still have plenty of it left"

Just Googled it and it's from a great source, Mr Groucho Marx himself..

http://www.mactheknife.org/Thoracic/Quotable_Quotes.html

 

I remembered, again, why I hadn't done this before; because the 1200XL OS doesn't support the PBI and then I couldn't use the Syscheck board.

 

Relax Matt, take your time, there's no race here..

Slow but sure..May all your machines shine brightly in the sun...I'm a terrible "get it fixed now" person despite having zero get up and go in me, I've had to understand that "it's done when its done" and it's a hard pill to swallow

I'm so used to doing things at 101% I forget that my body especially my head and eyes can't quite keep up..It leads to annoyance...

Your kit will be back and flying...Good luck my friend...

 

I appreciate the sentiment, but I'm hardly racing with these machines, just on the off day I get a chance to do something or test something with them. I haven't had my 1200XL running for at least a year now, and it's been a couple months for the 800. Though since it is just the keyboard with the 800, I am still able to enjoy a number of games that only need a controller or the function keys to use. So I don't have to resort to an emulator (which I really wouldn't, so for me that means no Atari).

As a programmer loves to program, I love to work on hardware, but neither likes doing debugging and being stuck on the same bug, round and round. The joy is in the creation and moving forward mostly, some in the ability to maintain for yourself, but that gets old if it's always trouble shooting and maintaining. Especially when it's a new issue that crops up when you are trying to upgrade, those are the worst of all. But even if I end up having to change out a dozen resistors and capacitors in the keyboard circuit, I'll get there eventually.

The circumstances point to a mishap in the upgrade work though, and so it's more likely a short or open circuit near one of the 4050's where it doesn't have to be across a dozen circuit traces which is highly unlikely. The more circuits that go out at once, the more likely it's at a bottleneck than across them all.

The truth is, I haven't put more than a total of an hour or two of actual hands on trouble-shooting over the weeks I've had the problem, I haven't even exhausted all the possible testing yet, like continuity across all the keys. I've started with the easiest and most obvious, which is swapping chips and visually following traces and resoldering questionable points. And searching for answers on-line.

My first issue was with the power board that I half-rebuilt, and once I at least got the Incognito SDX and/or memopad screens up, that's when I discovered the keyboard didn't work anymore (for sure-it could have been a lack of partial power from the PSU board).

I know what could happen when I start these projects, as we all do, even if you are experienced and do a mod or upgrade right, these are old machines and sometime the added stress of these projects is just to much for a component or chip, or board. My upgrades and mods over the past few years, in the end, turned out to be done right the first time, after checking over them a dozen times, and it's always been something else, not necessarily even directly related, that has cropped up troubles for me, generally something really stupid I did myself.

With the 1200XL PBI upgrade, I had one connection wrong that was something that had been revised by the project designer, but before I found that out I had swapped out several support IC's including the MMU, and it turned out that I swapped in a bad MMU, but didn't check it again until I had gone over every other possible thing! And that's one hell of a project to go over dozens and dozens of connections dozens of times!

 

Hi Matt
I've had a brief look at the diagram as well, seems "KR1" Reads all key values via the two 4051's and the control lines to each K0-K2 on one K3-K5 on the other. KR2 only reads "Break", "Ctrl" & "Shift" of which you know "Break" is working suspect the other two will be as well. (uses K5&K6 I think
"KR1" goes from pin 25 of Pokey to pin3 of Z104 and is pulled weakly high by R128 to 5V (47K) worth checking as this is the ONLY common line for the rest of the keys, should see negative pulses if working o.k that of course is not to say its not something else, hope its not POKEY.
" From Lee Avery's Altirra Hardware ref Man"
The Start, Select, Option, and Reset keys do not have key codes associated with them as they are detected differently; Shift and Control keys function as modifier keys.
This is via "KR2"
Scan algorithm
The keyboard scanning hardware consists of a 6-bit counter, a 6-bit latched compare register, and a state machine with four states. One key out of 64 total is checked per cycle at 15KHz, so a full scan takes 4ms that's the "KR1"


Hope this is of some help ....

Just checked my keyboard it's a stackpole
guessing its just waiting to break
wish you luck

 

I finally got back around to working on my 800 again yesterday.

The last time I bothered, several months ago, I had just replaced the sockets for the 4051 IC's as well as brand new 4051's, even though I already had about half a dozen on hand, just in case. After re-assembling the computer in stock configuration, I powered it up and got the "red screen of death." At the time I knew this was merely a bad connection issue with one of the 5 PC boards as it had been powering up to either the memo pad or SpartaDOS X command line depending on if I had it in stock or Incognito configurations (easily swapped out thanks to my new approach of using Dupont connectors on all wires), I just couldn't do anything else since the keyboard didn't work at all except for the break key (and function keys on a different circuit). But, I had grown tired of working on it, and set it aside, even though my trouble-shooting so far only included blindly replacing components instead of testing circuits and components properly.

So yesterday I took her apart and dis-assembled her and re-assembled her in the Incognito configuration since there's only two PC boards to worry about making proper contact. And as I suspected, it was a simple connection issue, she booted up to the SpartaDOS command line. However, the keyboard still doesn't work. But, it just so happens that I already have a back-up plan in mind, which will get me a working 800 while verifying for sure that the non-working keyboard IS a motherboard issue and not the keyboard itself.

What? You want to know my plan? OK, it's a plan that helps with more than one project as well. The other project that will benefit, which in turn gave me the excuse to take the easy way out, for now, if it is a motherboard issue for the keyboard, is my robot project. I originally planned on using a 400 for my robot's brain, but with my current situation with my 800, I decided I would give my robot an 800, but I'll still use a 400 keyboard with it. I decided to buy another 800 motherboard. I had one on my watch list for several weeks, from MyAtari, for $40 dollars. Which by the way is a good deal no matter what, these days, but it's also a way to still get extra POKEY's on the cheap, instead of buying just a POKEY chip for the going rate of about $65 or more these days (which is why I bought a Quad POKEYmax since it's less money!). And if you think buying $40 800 motherboards just for POKEY chips is a pretty good idea, you better jump on it because there's no telling when MyAtari will figure out they could sell the POKEY chips from these mobo's separately and still sell the mobo's for the same price sans POKEY! And don't think they won't start, others already have!

Which brings me to my even better deal I got on my 800 motherboard from another seller on eBay for a mere $15! But you probably just guessed that it's sans POKEY. I of course also bought a PSU board for it which was the same price. Anyway, these items just arrived in the post yesterday, which is why I started working on my 800 again last night. I already made sure that my 800's PSU board was, in fact, working properly as using the new one changed nothing with the condition of my 800. If the same happens with the new motherboard when I plug the keyboard into it, then I'll know straight away that it's the keyboard and not the motherboard. There's less than a dozen wires that I'd have to solder to the new mobo to move the Incognito over to it, so that's not a big deal once I'm 100% sure the keyboard is or isn't the problem.

After all is said and done, I'll have a working Incognito 800 soon, unless I have to wait on a keyboard, and eventually I'll properly trouble shoot the original 800 mobo in time to go into my Bender 800 robot.

 

Yay, no Red Screen .. All we need now is a working keyboard one way or another ? Are you working on this now Matt, or is this for another day ?

 

I will be working on it again this evening. First I'll start by adding the Pokey chip and a spare CPU board and OS and 16K ram board to the new motherboard, plug in the keyboard and power her up to see if the keyboard works or not. If it works, then I will install the Incognito into the new motherboard this evening. If the keyboard doesn't work, then it's safe to assume that it's the keyboard and not my original motherboard, and I won't have to do anything except start adding more upgrades and moving on with the project while I wait on getting a new keyboard. After one more attempt of finding any issue with the original keyboard. I'll have to wait until the end of the month to get one. Either way, the project can move forward after this evening.

 

I did just notice that my new motherboard has a male keyboard connector on it instead of the female connector on my original board, so I'll have to temporarily rig it up for testing, if it does work, I think it will be easier to attach a female connector onto the end of my keyboard's ribbon cable, since it's the Hi-tek keyboard with the bare wires that plug into the female connector, so instead of desoldering and soldering in a connector on the new motherboard I can solder a connector onto the keyboard's ribbon cable. I happen to have some break-away connectors on hand that I can use that should work, but if they don't match up I'll just solder Dupont connectors to the end of the ribbon cable and plug each one in individually to the motherboard's male connector. Then I'll have to order more Dupont connectors for all the other upgrades and mods...