I think I’ve killed my 800XL! Any ideas?

[bfollowell]

I've looked through the Sys-Check thread, and it looks like a device that would be nice to have, but almost impossible to get at the moment, with worldwide shipping the way it is at the moment. I am a little curious though, and can't seem to find any really good information on how it actually works. If you're using it to try and diagnose a completely dead 8-bit, or one that has no display, how do you determine what the issue is? How does the board communicate with you? I mean, the board itself has no display, and if the computer isn't able to display anything. How the heck do you get diagnostic information out of it?

 

Thanks.

 

[flashjazzcat]

1 hour ago, bfollowell said:

How the heck do you get diagnostic information out of it?

Sys-Check is able - through the innate properties of the PBI/ECI bus - to completely shut off the computer's on-board ROM and RAM. The device can then provide RAM and ROM externally, enabling the machine to boot even if the on-board ROM and/or RAM are faulty. Much of the time - unless the machine has a fault on the display circuit - the reason you get no display is because the OS crashed before initialised video output. A good thing to look for is whether the display says 'No signal' or similar after you turn the machine on. If it does, the computer is likely entirely dead (no clock running, GTIA dead, etc). But if the screen is stuck black or some other colour but you do get a video signal, it's likely that the machine crashed during the boot process, indicating a bad CPU, ROM, RAM, PIA, MMU, etc, etc.

 

Sys-Check has some other clever tricks up its sleeve (which I'm sure @tf_hh will explain if encouraged to do so), but that's my third-party precis of what's going on.

[bfollowell]

New RAM chips were fairly easy and cheap to obtain, so I figured that would be an easy possible fix to try, but no luck. My 800XL is still acting the same. I did order a Sys-Check card though. Hopefully, I'll have it in a month or so and that will help me narrow things down.

 

To remind everyone, my 800XL is only exhibiting a solid green screen upon boot. It accepts no input that I can see, there are no sounds, and nothing changes over time. A Star Raiders cartridge does nothing. I've confirmed that ANTIC, GTIA, CPU are all good. While I can't confirm that the memory chips I received are all good, changing all 8 chips has no affect on this behavior, so I'm kind of assuming they're good, and the originals probably are too, and that the issue lies elsewhere. I've checked just about all of the easy, socketed chips and they've all confirmed good. Now I'll have to move into a little harder to troubleshoot territory, unless my Sys-Check tells me something once it arrives.

 

I have a cheap rework station on the way and I'm considering ordering a starter, hobbyist scope. Hopefully, it will help me see things I can't with just a multi-meter.

[archeocomp]

It really looks like either OS or BASIC ROM are bad, with new RAM chips and (I assume) all data lines still all time low. Checked with logic probe READ, WRITE and chip select signals on the ROM and RAMS would probably not tell much, as the CPU stops very early. I would either de solder them or wait for Sys-Check.

Edited May 6, 2020 by archeocomp

[ACML]

I once had a 1200XL with a similar issue.  Turned out to be RAM, but there a twist.  If you've exhausted all other avenues right now, try this out.  Keep the same RAM chips, just arrange them differently (change position order).  Try a few combinations and see if that does anything.  Just by changing the position of one RAM chip, my 1200XL stared booting and working great again.  Have no clue why that mattered.  I thought maybe it was just pulling and re-seating chips that did it, so I put the RAM chips back in original order and it failed again.  So it did matter.  Maybe there was some critical timing issue that one RAM chip couldn't meet in that socket????

 

P.S. The position that seemed to matter was on the end (either first or last chip).

Edited May 6, 2020 by ACML

[bfollowell]

1 minute ago, ACML said:

I once had a 1200XL with a similar issue.  Turned out to be RAM, but there a twist.  If you've exhausted all other avenues right now, try this out.  Keep the same RAM chips, just arrange them differently (change position order).  Try a few combinations and see if that does anything.  Just by changing the position of one RAM chip, my 1200XL stared booting and working great again.  Have no clue why that mattered.  I thought maybe it was just pulling and re-seating chips that did it, so I put the RAM chips back in original order and it failed again.  So it did matter.  Maybe there was some critical timing issue that one RAM chip couldn't meet in that socket????

 

Since I completely replaced all 8 RAM chips, I find it doubtful that would help, but anything is possible and I have nothing to lose. I’ll play with rearranging them later and see if anything changes.

 

[+DrVenkman]

19 minutes ago, bfollowell said:

Since I completely replaced all 8 RAM chips, I find it doubtful that would help, but anything is possible and I have nothing to lose. I’ll play with rearranging them later and see if anything changes.

 

Well, there’s something you can lose, especially if you’ve already removed and reseated chips multiple times: the single-wipe sockets Atari used for most chips are not super durable. The metal contact wipes can and do lose spring-tension and they can even break. They’re really not intended for more than a few insert/remove cycles. 

 

And on that note, it’s always useful if you have a logic probe to check all the chips pins for activity and when there isn’t activity when expected, such as the refresh lines on DRAMs, the address and data pins on everything, etc., look closely to see of the socket is making good contact.

[+orpheuswaking]

21 minutes ago, bfollowell said:

Since I completely replaced all 8 RAM chips, I find it doubtful that would help, but anything is possible and I have nothing to lose. I’ll play with rearranging them later and see if anything changes.

 

I am in Indianapolis and I have a Sys-Check (older style but will work with an 800XL), I am willing to loan it to you.  

[bfollowell]

20 minutes ago, orpheuswaking said:

I am in Indianapolis and I have a Sys-Check (older style but will work with an 800XL), I am willing to loan it to you.  

Hey, that is much appreciated, but I've already ordered mine. Thanks anyway. I do appreciate the offer very much.

 

I'll keep you in mind if I need some help troubleshooting my 1088XEL build later!

[bfollowell]

I have a question about the keyboard. Can you boot an 800XL without the keyboard attached? I mean, I know you'd have no input, so you wouldn't be able to do anything, but will it boot properly?

 

[+DrVenkman]

1 minute ago, bfollowell said:

I have a question about the keyboard. Can you boot an 800XL without the keyboard attached? I mean, I know you'd have no input, so you wouldn't be able to do anything, but will it boot properly?

 

Yes, no problem. I do most of my board testing and repairs without a keyboard attached.

 

[bfollowell]

Ok, I thought so, but I’d read another post somewhere else that seemed to imply that it might not, and I just wanted to confirm. I didn’t necessarily think it was causing my issue, but I wanted to make certain I wasn’t introducing anything extra into the equation.

 

Thanks!

[bfollowell]

Well, I received my Sys-Check v2.2 BBU today and tried it out on my 800XL. No luck. I got more than I have been getting, which was nothing but a static green screen (see the first post), but the Sys-Check menu doesn't come up. I do get some garbled graphics on the screen similar, but different every time. They all seem to lean  towards the bluer end of the spectrum. I'm attaching four different boot videos so you can see what I mean.

 

 

 

 

 

 

As I said, all four similar and blueish, but not quite the same. No menu, no perceived input from pressing the keys.

 

I've verified that the dip switches on my Sys-Check are all set properly. To recap, I've already swapped ANTIC, GTIA, CPU & PIA with my 1088XEL. All of my 800XL chips work perfectly in my 1088XEL and my XEL's chips make no difference when used in the 800XL. I've swapped all of the RAM chips with new ones that I purchased. Pretty much the only socketed chip that I haven't tried swapping is the POKEY. After this I'll have to wait for my rework station and start pulling chips off the board, pick up new ones, install sockets and try the new chips until I find something that works.

 

Any ideas? Do the videos ring any bells with anyone?

 

Thanks.

 

 

Edited May 31, 2020 by bfollowell

[_The Doctor__]

not really, but I'd check video buffer/delay line... pull basic, swap OS chip... that's all I have atm...

Edited May 31, 2020 by _The Doctor__

[+tf_hh]

9 hours ago, bfollowell said:

I've verified that the dip switches on my Sys-Check are all set properly. To recap, I've already swapped ANTIC, GTIA, CPU & PIA with my 1088XEL. All of my 800XL chips work perfectly in my 1088XEL and my XEL's chips make no difference when used in the 800XL. I've swapped all of the RAM chips with new ones that I purchased. Pretty much the only socketed chip that I haven't tried swapping is the POKEY. After this I'll have to wait for my rework station and start pulling chips off the board, pick up new ones, install sockets and try the new chips until I find something that works.

 

You can remove the POKEY for a test with Sys-Check, it´s not needed. But regarding the videos you´ve sent I´m pretty sure this is a timing issue. GTIA is able to generate a stable TV picture, but ANTIC can´t show the Display List from Sys-Check. This one is located at Sys-Check´s ROM and doesn´t need any working RAM to be shown. Unfortunately without an oscilloscope to check the points of interest you can only swap chip by chip from now on.

 

[bfollowell]

I have an oscilloscope and can check those points. Which points should I check? Anything that can help narrow it down a little would be appreciated, even if I still have to wind up swapping a few chips.

[Mclaneinc]

Sorry to hear about the XL Bryon, the Syscheck is a good purchase even if it can't help here 100% and you get the extra OS's and 512mb of ram usable on its own as an upgrade. Best of luck with the fix, seems a nasty critter.

[bfollowell]

I'm not sure what was wrong about tf_hh's post, but I did go ahead and check the points he'd mentioned.

 

Here's Sync from CPU pin 7. I think it looks as expected from his original description.

 

 

These next two are Refresh from MMU pin 11 and ANTIC pin 8, respectively. I don't think these two look quite like he'd described, originally saying that they would be an almost all-time low signal. That doesn't really seem to describe what I'm seeing here.

 

 

 

The next three are PHI2 from CPU pin 39, ANTIC pin 29, and GTIA pin 30, respectively. They're not identical, but the all look fairly similar and relatively close.

 

 

 

 

These final two are PHI0 from ANTIC pin 34 and CPU pin 37. Again, not identical, but very, very close.

 

 

 

So, to me, the only ones that look suspect, at least from what tf_hh described I should be seeing is the two Refresh signals from MMU pin 11 and ANTIC pin 8. He said that if it was missing to check traces for continuity. Well, the signal isn't missing. It just isn't quite what I was expecting. Where does the Refresh signal originate and what might be causing it to look so abnormal, if in fact it does?

 

I guess it's time to read through some more of the SAM's 800XL service guide and study the schematic some more.

 

What do the rest of you think? Anything ring a bell with anyone here?

 

Thanks.

 

Edited June 2, 2020 by bfollowell
edit

[cwilbar]

Have you tried swapping the MMU ?  I did not see it listed on the list of chips you tried swapping.

 

A bad POKEY can bring the Atari to its knees....  but I'm not sure that is what is going on here... but don't count out POKEY 100%.

 

I'd say either the MMU is suspect or one of the input signals feeding it.... or something else on that line is introducing noise on that line ?

 

[+kheller2]

REF is generated by Antic.  Lift the Ref pin on Antic and test again.  Here is a whole thread on REF: 

 

 

[bfollowell]

I’m worried about “lifting” a pin. I’m worried about bending a pin and breaking it off. I’ve seen others reference bending a pin. How do you guys go about doing that then reinserting the chip into the socket without snapping it off?

 

[_The Doctor__]

very gently till it just skates by the socket, do not push all the way down if it looks like it will hit the PCB

[+tf_hh]

6 hours ago, bfollowell said:

I'm not sure what was wrong about tf_hh's post, but I did go ahead and check the points he'd mentioned.

 

Ehm... somebody has deleted BOTH posts now ? ? ?

 

I wrote one post. Then I sent by accident a second one with the same, but quoted content. Of course only the 2nd post was an error. Now both are gone...

 

 

[+tf_hh]

6 hours ago, bfollowell said:

These next two are Refresh from MMU pin 11 and ANTIC pin 8, respectively. I don't think these two look quite like he'd described, originally saying that they would be an almost all-time low signal. That doesn't really seem to describe what I'm seeing here.

 

 

This is normal when Sys-Check isn´t attached AND activated (Function Mode DIP-Switch 1 and 2 to "ON").

 

All other signals look fine. Did you check SYNC without Sys-Check? It doesn´t remain at low after a few seconds?

 

In this case, the system seems to be working internal, but video display is disturbed. Check all signals if continuity is given from ANTIC to the other points (see schematics).

 

[bfollowell]

OK, I misunderstood. All of my original waveforms were taken from my stock 800XL without my SysCheck attached.

 

With SysCheck attached, this is what I see for Refresh

 

MMU pin 11

 

ANTIC pin 8

 

 

I think these look more like what you were originally describing. All the others look more or less identical to the ones I took before, without the SysCheck attached. So, from what I can tell, nothing looks really amiss in any of the signals that I've checked so far.