Dead New-in-Box 600XL?

[flashjazzcat]

This is a puzzler. I received a virtually NIB 600XL a month or so back as payment for some bespoke modding. What a beauty. Box in good shape, 600XL looking absolutely factory fresh and still in its polythene wrapper (albeit minus the protective film on the console keys)... a real stunner. Swooning too much to think of testing it at the time, I ordered some RAM chips for a 64KB upgrade (which arrived last Friday), and this evening settled down to perform the simple mod. It's then I discovered the 600XL boots to a black screen.

 

I have gone ahead and performed the 64KB upgrade, as much to rule out bad RAM as anything else - to no avail. Chroma and Luma are properly hooked up to the monitor jack pins on the underside of the board. RAM upgrade wiring triple-checked.

 

So far, I have:

 

• Pulled and reseated every IC on the board
• Swapped - one by one - every major IC on the board with equivalents from a known working 600XL

I certainly don't have time to do much more on this over the next couple of days, but it's puzzling that this happens to a machine which has ostensibly been sitting in its polystyrene cocoon for twenty-five years. I have no gripe with the client who sent me the machine, but I've sent him an email to ask him if he tested the 600XL at all. It may be that he just moved it on untested, having picked it up on eBay. However, if he did test it, something odd's happened during transit.

 

Anything to watch out for with machines which have received little use and sat untouched for many years? I know I'm probably grasping at straws here...

 

 

 

[Rybags]

The one I got last year... dead 6502, so easy fix.

 

Have you tried stuff like Delay Line, MMU, LS* ICs involved in RAM selection?

 

Another idea - if you've got an AtariMax flashcart, make an image with a diagnostic mode cart that just flashes some colours up without actually needing any RAM.

Although I'm not sure if the flash software allows you to provide the default bank's code, might have to hack the thing into a generated ATR file.

 

 

ed - also, with the 64K upgrade, you can plug the 64K chips in without doing the remainder, and it'll function just fine as a 16K machine.

Edited March 2, 2011 by Rybags

[flashjazzcat]

Well, I'm trying to fit this in between other soldering jobs, but I swapped the RAM upgrades in the two 600XLs, lock, stock and barrel, and the NIB machine still isn't working. Apparently it was functional before I took delivery, so it's either been fried in transit or it didn't take kindly to being dismantled. Wish I'd tested it before I took the motherboard out.

 

So, all the areas I've worked on have been isolated and eliminated. The machine actually isn't booting at all (judging by the audio), although it puts out a (black) video signal, so all I've managed to establish is that the system's getting power.

 

 

 

[sloopy]

hook up a logic analyzer to it and see what it has to say...

 

doesnt everyone have one? ;')

 

serial tho, pull the OS out, and see if you get something different, if the video doesnt sync, or you get the PMG on screen (vertical bar(s) ) then the machine is atleast starting to run the OS, and clearing out mem etc...

 

sloopy.

[flashjazzcat]

I tried whacking an IntSDX in the OS ROM slot (without making the hookups), and the computer stuck on a pale orange screen. So I can change the behaviour of the machine slightly by changing the OS ROM. A known good XL OS ROM chip boots to black, however. I do get the same momentary red as the working 600XL displays, however, the instant the power is turned on. I just reflowed the solder on the OS ROM socket but no improvement.

[Almost Rice]

are u using a known good PS or the one from the casket?

[flashjazzcat]

are u using a known good PS or the one from the casket?

Known good, which I'm also testing the working 600XL with when swapping parts.

[Rybags]

Probably the RAM or select logic then.

 

Have you tried that diag flashcart idea yet?

[flashjazzcat]

Not yet: I spent the evening working on Phil's 800XL, and I'm a bit pumped with that s-video board.

 

I doubt it's the RAM, since I switched the chips between this machine and the working 600XL, which works with either set of chips. Do you think the RAM sockets might have a dry joint (possibly opened up when I prized the chips out)? I already switched the LS158s - same story - and reflowed their sockets.

Edited March 2, 2011 by flashjazzcat

[flashjazzcat]

Been doing some soldering tonight, so I decided to have another look at the 600XL while I was about it. Remembering that RAM 320XL completely supercedes the on-board RAM, I tried it in the 600XL. The results were interesting: half the time, the machine powers up with a brief glimpse of dark-red screen (normal for an XL in my experience), but then the screen goes black again, generating an audible "click" from the CRT. The RAM 320XL's light stays lit. The other half of the time, the machine boots straight to the self-test, but with a corrupt display and frozen, thus:

 

 

Sometimes the self-test goes straight into the memory test, and depicts - amidst a corrupted display - a failed ROM test. Swapping out the OS ROM and BASIC doesn't help. I've reflowed all the relevant sockets, swapped out the MMU, RAM and LS158s (again), but I can't get any further with it. Without the RAM 320XL installed, the machine won't boot at all.

 

Interesting that it goes straight to the self-test every time it does boot. This is a real puzzler: it's a nice board, and I know that it was working before I received it.

 

F**K... looks like the 1084S has just died (sounds like flyback transformer, judging by high-pitched whine). Not my day...

Edited March 18, 2011 by flashjazzcat

[spookt]

That pretty much matches the symptoms I got on my good 600XL with RAM320 before I did the timing fix.

[flashjazzcat]

That pretty much matches the symptoms I got on my good 600XL with RAM320 before I did the timing fix.

Ah - thanks Paul! I'd never considered the timing issue. If that fix works, I guess there's a RAM problem to trace. I'll test tomorrow.

Edited March 18, 2011 by flashjazzcat

[mimo]

I have an LOPT out of a 1084S, I thought that it was the problem as the picture had collapsed so I fitted a new one, did not fix the problem. On further inspection it turned out to be a bloody big crack in the PCB (since repaired)

So, I have a spare LOPT of uncertain functionality that you can have for cost of shipping

[Marius]

I would say go swap ic's one at a time and start with CPU!

[flashjazzcat]

I have an LOPT out of a 1084S, I thought that it was the problem as the picture had collapsed so I fitted a new one, did not fix the problem. On further inspection it turned out to be a bloody big crack in the PCB (since repaired)

So, I have a spare LOPT of uncertain functionality that you can have for cost of shipping

Sounds great - thank you. I'll try and get the case open next week and see if there's anything obvious. I may well take you up on the offer.

 

I would say go swap ic's one at a time and start with CPU!

Been there, done that. Following Spookt's suggestion, I think a Phi2 fix will see the machine work with RAM 320. That points to the memory being duff, which is odd, since the MMU and RAM work in my other 600XL. So - problem with the board?

[Rybags]

Did you try other less substantial stuff such as Delay Line and some of the generic ICs used in the select logic?

[flashjazzcat]

I'll tell you what I swapped out/resoldered so far:

 

RAM

74LS158s

MMU

74LS51

74LS375

BASIC

OS

CPU

 

Can you suggest others worth checking?

[Rybags]

75LS138 - individual selects for I/O pages.

 

74LS74 - clock divider.

 

Maybe a continuity check between connections on various paths of the Address and Data buses would reap more. It'd only take one intermittent line to foul the whole system up.

[Defender II]

I'd do a good visual inspection checking for a solder bridge. Sometimes all it takes is a small drop of solder in the right/wrong place.

[mimo]

Replace the RAM sockets

[Rybags]

With the RAM 320XL you should be able to run with the RAM totally removed - probably worth a try too.

[flashjazzcat]

Hmmm... decided to fit my other VBXE to the other 600XL; after fitting the precision socket for ANTIC and reinserting chip, the display is screwed up. Continuity tested the socket... all appears well. I've got two 600XLs and neither of them work now... don't say it's going to be one of those weekends.

[sloopy]

Hmmm... decided to fit my other VBXE to the other 600XL; after fitting the precision socket for ANTIC and reinserting chip, the display is screwed up. Continuity tested the socket... all appears well. I've got two 600XLs and neither of them work now... don't say it's going to be one of those weekends.

 

ok... i will say 'its gonna be one of these weekends...'

 

:')

 

sloopy.

[Mclaneinc]

Message to self, never lend Jon any hardware, he's a Terminator!!

 

Seriously, sorry to hear Jon, hope the luck improves and they all fix, I think we all have had times like this and its a real ball buster.

[flashjazzcat]

Thanks folks... I had lifted a couple of traces (despite my best efforts) when removing the original Antic socket, but I'd checked it over before I fitted the new one so I was forewarned and jumpered them elegantly: four short tracks under the top right corner of the chip, easily fixed. Fortunately, if I put my index finger on the top right corner of the CPU, the display rights itself, so I'm assuming something isn't earthed properly. Should be fairly easy to track down.

 

Naturally I'm a bit reluctant to do any more cutting work on the latest XL Style SIO2SD conversion today... mangled, scarred plastic beckons.