Advice please: Ordered U1MB, Which Machine?

[flashjazzcat]

eBay Auction -- Item Number: 121660321429

[MrMartian]

What I used was an old CD audio cable for a PC, which I had a box full of......

[+DrVenkman]

eBay Auction -- Item Number: 121660321429

 

That would've been nice to know a few days ago, lol! I tracked down some similar auctions from the U.S. eBay site for the future.

 

What I used was an old CD audio cable for a PC, which I had a box full of......

 

If I'd had ANYTHING lying around in my bins of old electronic stuff to sacrifice in order to use the connector I would have - those four wires are all I have to to do to finish installation. As it is, I found something just like this from Amazon; I'll have it in two days, which is much faster than I can get anything from eBay, albeit at about twice the price. Still pretty darn cheap though. I'll snip the other ends, strip the wire and solder the CPU connections.

 

http://www.ebay.com/itm/40pcs-Dupont-Jumper-wire-cable-Female-to-Female-1p-1p-pin-connector-2-54mm-30cm-/291196207151?hash=item43cca5802f

[+DrVenkman]

Holy shit, you guys! It worked on the first try. Thanks for your advice and comments along the way.

 

 

 

[+DrVenkman]

Okay, slight update - the U1MB keeps losing settings and periodically resetting back to defaults. Also, occasionally the screen is garbled and filled with junk characters. Do I need to replace the battery or does it sound like I have a bad/shorted connection somewhere?

[flashjazzcat]

I'd be amazed if the battery was flat, so sounds like a bad address line somewhere. If BIOS can't read config data reliably, checksum will always be off and settings will be repeatedly reset. Probably time to get the continuity tester out, checking both ribbon cables for shorts and open circuit.

[+DrVenkman]

I'd be amazed if the battery was flat, so sounds like a bad address line somewhere. If BIOS can't read config data reliably, checksum will always be off and settings will be repeatedly reset. Probably time to get the continuity tester out, checking both ribbon cables for shorts and open circuit.

 

Will do, thanks!

[+DrVenkman]

Okay, I unbuttoned the machined and checked what I could. My HALT line soldering was a little dodgy so I re-did. I also touched up a jumper I'd installed at R63 as part of the video mod I did, and the W11 jumper for the 1200XL OS stuff. None of that stuff looked outright wrong, and I couldn't find any solder bridges or stray problems.

 

I tried to do as FJC suggested and check continuity of all the ribbon connectors but as it turns out, my test leads are all too large to make contact with the female header connectors. Any suggestions there?

 

Anyway, I put it back together and ... same thing. This time on boot it ALSO says "Ultimate Clock Not Installed" That is leading me to think that either the battery IS dead or there's something amiss with the U1MB itself. I picked up a couple spare CR2032 batteries today so I'll give one of those a try in a little while.

 

On the upside, I took my Dremel to the cartridge guide slot and case opening "throat." A few minutes with the diamond burr at 6,000+ RPM and voila. It's firm but it fits.

Edited July 29, 2015 by DrVenkman

[flashjazzcat]

Just put the probes on the solder side of the male IDC connectors with the ribbon cables attached. Since you can (hopefully) discount any shorts on the male connector, this will still reveal any issues with the ribbon cable.

 

Did you stick with the stock tin-plate sockets on the motherboard? In any case, this is still shouting out address line problems to me (especially now that the RTC has stopped working).

Edited July 29, 2015 by flashjazzcat

[AtariGeezer]

Okay, slight update - the U1MB keeps losing settings and periodically resetting back to defaults. Also, occasionally the screen is garbled and filled with junk characters. Do I need to replace the battery or does it sound like I have a bad/shorted connection somewhere?

I'd check the connectors in the battery holder to be sure there are not flattened after inserting the battery...

[+DrVenkman]

Okay guys, first off, thanks for your continued patience - I really do appreciate the help.

 

Now, here's where I stand: I checked the continuity of the ribbon cables as FJC suggested and they're all fine. I went ahead and replaced the battery just in case but there was no real change in the odd behavior.

 

After playing around with things for awhile I have noticed a few more oddities. For instance, once in awhile, upon entering the U1MB Setup screen I get garbled squares (for lack of a better terms) or garbled text on a yellow background.

 

 

 

 

 

More rarely, upon exiting the Setup screen, the computer drops into Missile Command - I tried to play it but it (oddly) doesn't respond to the LEFT on the joystick - up, down, right and fire work fine. I thought this might have been some kind of new physical damage to the controller port or board connections but they're all fine. And when I let the machine "rest" for awhile then boot up a cartridge game, the joystick port works fine again. When I disable SDX on the U1MB and use the SIDEloader on my SIDE2 cartridge, games start working fine with the joystick but then sometimes LEFT doesn't work again.

 

Perhaps related, perhaps not, I noticed today that he computer is not responding to anything on the SIO bus and cannot access real floppy drives or my SIO2USB device. During my modifications the other day, I also removed and jumpered R63 as described in the FAQ to enable +5V on the SIO line. The soldering looks good and so do the traces on the back of the board. Further, as this is a 1200XL, I removed jumpers W7, W8 and W9 and added a jumper at W11. The new jumper is clean and the solder looks good.

 

Now, on the off chance these problems are with my hardware generally and not the U1MB itself, I swapped POKEY out of another working machine. I have had other A8's with weird I/O problems and replacing the POKEY corrected them. In this case it made no difference.

 

So I'm at a bit of a loss here, folks. Any other suggestions? Could it be a bad 6520?

 

[+DrVenkman]

So I took things apart AGAIN, swapped out a 6520 from another system (not sure it made any difference) and rebooted. This time the machine could see the SIO bus properly and after several long, trouble free disk I/O operations without any problems, I took the risk of flashing FJC's latest U1MB alternate BIOS and v1.3 of the PBI ROM. I also updated SDX from from the version 4.46 that was on it to 4.47, flashed the other "BASIC" slots to the Assembler/Editor cart, the PILOT language cart and Pac-Man because, why not really?

 

After rebooting things seemed to be going fine until I had a random hang a few minutes later. Then the setup screen was all corrupted like I had seen before with the stock U1MB bios (see pics above) and I couldn't access SIO devices again. The RTC data was also lost. After unplugging the machine and leaving it off for a little while, I tried again. This time the setup screen graphics were okay but once again the RTC had been reset.

 

So here's what I'm thinking I'll do: tomorrow I'm gonna re-do my CPU jumpers from scratch. I'm gonna go over the board with a fine-toothed comb looking for ANY hint of a cold solder joint at the MMU socket, a broken trace anywhere along any of the address lines, a stray droplet of solder from from my prior mods that may have fallen from the solder pump somewhere, or any other indication of what could be amiss. If none of that bears fruit, I'm willing to consider exorcism ...

 

In the meantime, my SIDE2 cart is working GREAT in my 800XL. So at least there's that.

[flashjazzcat]

I was going to suggest looking at the wires on the 4 pin header next (those going to the CPU).

[+DrVenkman]

I was going to suggest looking at the wires on the 4 pin header next (those going to the CPU).

 

So if that fails, what do you think of my idea of the exorcism?

 

Seriously, thank you for your patience. Last night I found a thread from 2012 involving the original U1MB (not the revised 2K14 edition I have). The fellow was having remarkably similar problems and they ultimately appeared to involve the CPU connections. I'll get on that today and see if I can get a handle on things.

 

http://atariage.com/forums/topic/206560-upgrading-1200xl-with-ultimate1mb-removing-ape-warp-os-32-in-1/?hl=%2B1200xl+%2Bjumpers&do=findComment&comment=2675697

 

And that brings up an aside: in my install, I removed jumpers W7, W8 and W9, and installed one in W11. I note there's a lot of incomplete info on the 'net about what those jumpers do and which ones to install/remove. In the thread above, the fellow had installed an APE/WARP 32 OS mod that had called for removing W6 as well. I see others talking about removing W7, W8 and W9 and installing W11, W12 and W13. Can anyone provide some clarity on what all these jumpers do and why you would or would not install them, specifically in connection with this mod?

Edited July 30, 2015 by DrVenkman

[flashjazzcat]

 

And that brings up an aside: in my install, I removed jumpers W7, W8 and W9, and installed one in W11. I note there's a lot of incomplete info on the 'net about what those jumpers do and which ones to install/remove. In the thread above, the fellow had installed an APE/WARP 32 OS mod that had called for removing W6 as well. I see others talking about removing W7, W8 and W9 and installing W11, W12 and W13. Can anyone provide some clarity on what all these jumpers do and why you would or would not install them, specifically in connection with this mod?

 

I'm not sure exactly how the jumpers configure the sockets, but my 1200XL has six jumpers removed: W6, W7, W8, W9, W12 and W13. W11 has the jumper fitted. It works.

[+DrVenkman]

I was going to suggest looking at the wires on the 4 pin header next (those going to the CPU).

 

Okay I had some interesting results - perhaps someone can tell me what they mean ...

 

When I first installed the CPU jumpers, I soldered straight to the shoulders of the pin. It was easy, and I didn't have a handy reference for which pads might offer the signals I need. So today I checked continuity, as you suggested. PHI2 was fine, READ/WRITE was fine, RESET was fine and so too was HALT - from the solder point to the pin on the back of the U1MB, each one was correct.

 

So then I decided to check to see if there was something shorted on the board somewhere. PHI2 showed open when tested against all the other pins. So too did READ/WRITE and RESET. But when I checked HALT, it gave me continuity tested against the PHI2 pin. The converse was NOT true - there is no continuity from the PHI2 solder point to the HALT pin. I presume there must be a diode in the way so that current is one-way. Well, I personally have no idea if the HALT line is supposed to have continuity back to the timing signal, or if I have a short in my 1200XL board somewhere - I did remove and move around several components, added jumpers for the first three steps of the ClearPic mod, etc.

 

On the chance that there's an issue with my wiring and not my board, I removed the CPU lines from the chip and cleaned off the pins, then dug up a thread from a couple years ago (the thread I linked above where the fellow had the same problems I do). I soldered jumpers to pads this time and then measured continuity. Same exact results. Hmmm ...

 

Like I said, I don't know if there's SUPPOSED to be one-way continuity back to the PHI2 line from HALT (but not vice versa), or if that indicates a problem with my 1200XL. So I'm at a loss. If I can't get this figured out by tomorrow, I'm probably gonna pull the U1MB from the 1200XL and install it in my main 800XL instead. I'm getting pretty good at soldering those CPU lines.

 

EDITED TO ADD: I forgot to mention that after I buttoned things up tonight, I did a tiny bit of testing to see if I could trigger the issues. Essentially, first boot up is fine. With HELP held, the U1MB setup screen appears, I can set options, set the time (I'm using FJC's custom bios), turn SDX on or off, etc. Saving and rebooting gives me whatever I have set: SDX on or off, extra memory or not, etc. HOWEVER, once the 1200XL is running, if I press System Reset, all I get is a black screen. Power cycling gives me the U1MB screen again, with date/time reset and all other settings set to defaults. Clearly something is amiss with either my 1200XL or the U1MB itself.

Edited July 31, 2015 by DrVenkman