1200XL + U1MB: New Stability Woes

[+DrVenkman]

After two-plus years of rock-solid stability, my restored Ugly Duckling 1200XL has suddenly started glitching on me and I haven’t yet been able to track it down. Things started going pear-shaped after I updated the machine to the latest version of Jon’s BIOS for the U1MB. I’m not saying Jon’s BIOS is to blame, nor am I insinuating anything of the sort! I think pulling the machine off the shelf, connecting it up and putting into use after a couple months sitting on the shelf may have uncovered a latent problem.  Alternately, it may just be pure coincidence. 

At any rate, when I pulled the machine apart, the jumper wire I had installed under the board between pin 23 of SALLY and the now-vacant W6 jumper location seemed to be fragile, for lack of a better word. At first it looked intact but when I touched it, it came off the CPU pin almost instantly. So thinking I had discovered the issue, I resoldered the jumper wire, powered the machine up and it ran without issue for a couple hours. Problem solved, I thought. 

 

Last night I powered it up again and the problems recurred - the machine crashed after about ten minutes. So I disassembled it again and inspected everything internally pretty closely and all seems well including the newly soldered CPU jumper. I wonder if I have an IC going bad somewhere, or perhaps one of the U1MB jumper wires might be to blame, though all four of them look good, solder joints intact, no stray flakes of solder anywhere I can see on the board, etc. My U1MB cables between the U1MB and the OS and MMU sockets look good and again, the whole machine has been perfect until recently for over two years.

 

Anyone have any ideas? 

 

[Roydea6]

temperamental and pissed at being put on the shelf for months. 

[+DrVenkman]

This is weird. It’s been up and running trouble free for almost an hour since my last check, including running some graphically intense or busy programs (the old Atari CES demo first for 10 minutes and now the demo screens and music for The Goonies). And I’ve got it out on the workbench in my un-air conditioned garage too, so if it was going to overheat it I would think it should have. 
 

So I’m confused here unfortunately. 

[flashjazzcat]

I'd try swapping out the ribbon cables, at least if you can do so with 5-10 mins work. The ribbon cables are the weakest link in the chain and when they go bad, odd things can happen. Not saying they're guaranteed to be the problem, but they're always the first thing I try to rule out on an U1MB machine. If the ribbon cables and the four jumpers are good, you can probably start to look for culprits other than the U1MB itself (although my second troubleshooting step is... you guessed it: a different U1MB).

 

Edited September 15, 2019 by flashjazzcat

[+bob1200xl]

How does it normally fail? Do you normally see wild characters in your screen data? That should not crash the system, but it is an indication that memory access is flakey.

 

Bob

[+DrVenkman]

6 minutes ago, bob1200xl said:

How does it normally fail? Do you normally see wild characters in your screen data? That should not crash the system, but it is an indication that memory access is flakey.

 

Bob

Yeah, screen garbage. I can usually recover with a system reset and get back to the U1MB BIOS with RESET + HELP as usual. 

 

And as a required nod to the old saw about the car making a funny noise EXCEPT when the mechanic is driving it, now that I posted this topic, the machine has been running for close to two hours now out on my workbench in 84 degrees F summer heat with absolutely no issue, idling at the Action! screen editor, with Action! loaded from the SIDE2 cart via Jon's new SIDE2 firmware, with a simple Hello World program loaded from an ATR file mounted via the SIDE loader in the U1MB.

 

So ... who knows? Maybe I've scared the gremlins out of the machine? (But somehow I think not).

 

Anyway, I'll report more if something else interesting appears in terms of symptoms. If things go flaky again I do have a set of new and never-used U1MB cables around here somewhere in a box from Lotharek (that U1MB went into my XLD prototype build so the cables were surplus). I do not have a spare U1MB however, so hopefully it's something to do with this 1200XL, the U1MB ribbon cables or something along those lines. 

[flashjazzcat]

2 minutes ago, DrVenkman said:

I do not have a spare U1MB however, so hopefully it's something to do with this 1200XL, the U1MB ribbon cables or something along those lines

You could also disconnect the U1MB entirely (put an XL OS ROM and MMU back in) and see if the issue still crops up (assuming issues return at all).

 

[+DrVenkman]

3 minutes ago, flashjazzcat said:

You could also disconnect the U1MB entirely (put an XL OS ROM and MMU back in) and see if the issue still crops up (assuming issues return at all).

 

Good suggestion! I thought about that this morning as I was working and then in typical middle-aged fashion, the idea drifted off to parts unknown until your suggestion.  

[+bob1200xl]

You could try freezer spray - sparingly and localized to just a few suspects.

 

Do you see screen corruption before the crash, or is it simultaneous?

 

Bob

 

[+DrVenkman]

1 minute ago, bob1200xl said:

You could try freezer spray - sparingly and localized to just a few suspects.

 

Do you see screen corruption before the crash, or is it simultaneous?

 

Bob

 

Usually there would be some weirdness before - such as the photos above in my initial post. Sometimes it was just crashing with little or no warning at all. 

 

And of course now, it's still up and running - about 45 minutes ago I unplugged it at the bench in the garage and brought it back inside to its usual home on a computer desk in the AC and powered it back up. It's been working non-stop just fine. These gremlins have decided to hide on me now that I've called them out.

[flashjazzcat]

It won't fail again until half-way through a BIOS update.

 

[+DrVenkman]

13 minutes ago, flashjazzcat said:

It won't fail again until half-way through a BIOS update.

 

It’s sitting there plotting and biding its time ... 

 

 

 

(EDITED TO ADD: Fun fact - this desk was built by my step-dad circa 1982 to hold first my and my step-brother's Atari 400 and then our 800 when we upgraded from the 400. It's got a circular cutout in the right-side riser sized for an SIO plug and room enough at right for a disk drive - we had a 410 there at first - plus a printer, which we added a couple years later. That power switch is wired to 4 x dual-plug grounded outlets properly mounted and wired inside metal boxes and a single pigtail to the wall - it was an 80's version of a power strip in the pre-surge protector era). 

[_The Doctor__]

I'd still bet on the ribbon cables, as the cables settled back into place they probably regained full contact... not unlike those flaky sockets some have come to find from time to time...

Edited September 15, 2019 by _The Doctor__

[+DrVenkman]

3 hours ago, _The Doctor__ said:

I'd still bet on the ribbon cables, as the cables settled back into place they probably regained full contact... not unlike those flaky sockets some have come to find from time to time...

This machine is the same one that required no less than 3 replaced sockets due to either outright corrosion (one socket wipe was broken completely off) or loss of spring tension in a number of individual wipes. 

 

At any rate, the computer has now been up all afternoon and into the evening running various things without a single hint of a problem, through several reboots and a few power cycles. I've copied files to and from SIDE-mounted ATR files, files loaded into an SDrive-MAX, and now I've spent a couple hours running a few different ATX-format game disks through their self-demo mode. 

So I have no idea what was causing my earlier issues nor why they have disappeared for the nonce. Hopefully they will not reappear during Jon's suggested most-likely time. 

[+Nezgar]

If the machine still has the original capacitors, maybe this is a rare case of old capacitors needing to be 'reformed' after long idle periods by using them again. After a while of poweron time they may be "good" again.

 

I know the general consensus is Atari 8-bits don't normally need recapping, but maybe this is the beginnings of symptoms that might warrant it...

[Gunstar]

I've had stability problems with my 1200XL on 3 occasions. The first time was a bad ribbon cable I had going to the 32-in-1 OS back when I was attempting to use a POKEY stereo board instead of my final solution of the original DIY piggy-back. The second time was just after doing my PBI mod, and it had me stumped and going over a perfectly good mod a dozen times before re-checking the MMU which was flakey and replacing it solved the issue. The irony of that issue was that I introduced the problem by swapping out a good MMU for a bad one when I was having stability issues, initially, from having my PHI2 signal coming from the wrong place which @flashjazzcat pointed out to me. The last time actually wasn't the 1200XL, so sorry there, since it was a bad flash chip for SDX, which I have on cartridge with MyIDE II piggy-backed. 

 

I've had other issues over the years, of course, but those were all complete non-working issues, not stability issues. Except with bad dram in extended memory, which only had issues when using extended ram. These days I've reverted my 1200XL to 64K dram and have external extended ram via the PBI. Less ram issues that way, and also a nod to the 1400XL that had only 64K internally and a PBI, for my "Poor Man's 1400XL."

Edited September 16, 2019 by Gunstar

[+DrVenkman]

So tonight, after working all day yesterday and into yesterday evening flawlesssly, I came home tonight and powered it on to a black screen, lol. Gremlins are back, it seems. Next weekend I'll take it apart again and start scoping things to figure out what's what. 

 

Aren't old computers fun?  

 

EDITED TO ADD: And the culprit just accidentally revealed itself to me: the power jack!!! Arghhh! Sometimes it's the simplest things ...

 

I was sitting at my desk about to disconnect the computer and set it aside for a few days when I decided to cycle the power again on a lark. Nada. Then my phone rang and I got busy for a minute. I came back to the machine, pulled it forward and tilted it back to clear the shelf on my desk and the machine suddenly flickered to life! I had forgot to flip the power switch to off and when I tilted the computer, it put pressure one the power jack. So I took the machine out to my bench and looked closely: there was surface corrosion all around the outer tab of the power jack. For those who read along with my long repair thread a couple years ago, this poor machine had been very badly abused in a prior life; this corrosion at the power jack should not be a surprise. So now eventually I will remove and replace the jack, but for now, I sprayed Deoxit into it and then scrubbed it well with a nylon toothbrush. Upon letting it dry, plugging it back in and powering up, the computer came to life instantly. 

 

Have I mentioned that old computers are fun?

[+Stephen]

1 hour ago, DrVenkman said:

Have I mentioned that old computers are fun?

That they are - same as old cars, old houses, etc.  Something satisfying about keeping them going though.

[tjlazer]

On 9/15/2019 at 8:51 PM, Gunstar said:

I've had stability problems with my 1200XL on 3 occasions. The first time was a bad ribbon cable I had going to the 32-in-1 OS back when I was attempting to use a POKEY stereo board instead of my final solution of the original DIY piggy-back. The second time was just after doing my PBI mod, and it had me stumped and going over a perfectly good mod a dozen times before re-checking the MMU which was flakey and replacing it solved the issue. The irony of that issue was that I introduced the problem by swapping out a good MMU for a bad one when I was having stability issues, initially, from having my PHI2 signal coming from the wrong place which @flashjazzcat pointed out to me. The last time actually wasn't the 1200XL, so sorry there, since it was a bad flash chip for SDX, which I have on cartridge with MyIDE II piggy-backed.

 

I've had other issues over the years, of course, but those were all complete non-working issues, not stability issues. Except with bad dram in extended memory, which only had issues when using extended ram. These days I've reverted my 1200XL to 64K dram and have external extended ram via the PBI. Less ram issues that way, and also a nod to the 1400XL that had only 64K internally and a PBI, for my "Poor Man's 1400XL."

What problems were you having with the bad ribbon cable?  I am having some issues with my 1200XL and 32-in-1 OS.   Can't determine if it's a bad flash chip, bad rom image, or bad connection.  I get some corrupt graphics on some games on a cold boot.  After a few resets, it goes away.

[+DrVenkman]

22 hours ago, tjlazer said:

What problems were you having with the bad ribbon cable?  I am having some issues with my 1200XL and 32-in-1 OS.   Can't determine if it's a bad flash chip, bad rom image, or bad connection.  I get some corrupt graphics on some games on a cold boot.  After a few resets, it goes away.

You should start a dedicated thread, preferably with good photos of your installation. I'm sure folks would be happy to offer suggestions. Unfortunately, it's easy for posts to get buried if they're not part of a dedicate thread. 

 

*****

As for my own 1200XL, it's been perfectly stable again since I figured out the issue with the power jack.

[tjlazer]

You should start a dedicated thread, preferably with good photos of your installation. I'm sure folks would be happy to offer suggestions. Unfortunately, it's easy for posts to get buried if they're not part of a dedicate thread. 
 
*****
As for my own 1200XL, it's been perfectly stable again since I figured out the issue with the power jack. [emoji4]

Thanks for your reply, I already started one. I think it may be the flash chips I’m using. 29F040B. Ordered some non B’s to test!

[Gunstar]

I'm sorry I some how missed your inquiry @tjlazer . Simply put, bad connections and/or broken wire in the ribbon cable caused system crashes and corrupt graphics if the computer was even slightly jarred, or just out of the blue after a while. I decided not attempt a second time and went another route, ensuring the 32-in-1 was installed properly, directly in the socket.  It all worked perfectly fine at first, but months down the road it started exhibiting this behavior. I guessed it was a bad chip on the 32-in-1 at first, after narrowing it down to a OS rom issue, had it replaced for free from Atarimax, with the same issues anyway. That's when I decided to get rid of the POKEY stereo board so I could install the 32-in-1 the right way.

Edited September 21, 2019 by Gunstar