Installed 74F08 in 800XL, machine no longer boots

[+x=usr(1536)]

From the 'memo to self' department:

 

Finally had some time to test U18 (74F08) for continuity.  Destination pins were determined from the schematics found in the Atari 800XL Field Service Manual.

 

Results:

 

U18 pin 4: no continuity to ANTIC (U7), pin 34

U18 pin 8: no continuity to 74LS375 (U28), pin 9

 

Pin 4 on U18 appears to be bridged to pin 5 on the same IC, so should be any easy check and fix.

 

All others were OK.  I'm suspecting lifted pads on pins 4 and 8 given that they've been reflowed a couple of times.  Will try jumpering them directly to their destinations at a later time and see if that resolves the issue; if so, will make that the permanent fix.

[+x=usr(1536)]

On 7/21/2022 at 8:48 PM, x=usr(1536) said:

U18 pin 4: no continuity to ANTIC (U7), pin 34

U18 pin 8: no continuity to 74LS375 (U28), pin 9

Fixed these, re-tested, no change.  Swapped 74F08s and 74LS08s to double-check; no dice.  Confirmed that U18 was getting power - 4.96V between pins 7 and 14, so it's being fed correctly.

  

On 7/3/2022 at 5:20 PM, x=usr(1536) said:

However, some further poking around revealed that the wire attached to (IIRC) pin 12 of the PIA going to the rear of the logic PCB used by the 256K upgrade had come adrift.

Double-checked this as well and found that I had reattached it a row lower (i.e., on a different IC) than it should have been.  Moved it to the correct position and no change.

 

At this stage I'm wondering if it might not be time to remove the 256K upgrade and try to troubleshoot from stock.  Erf.

[reifsnyderb]

19 minutes ago, x=usr(1536) said:

At this stage I'm wondering if it might not be time to remove the 256K upgrade and try to troubleshoot from stock.  Erf.

This is what I just did with a dead 800XL for which I discovered a 256k upgrade installed.  Eventually, I figured out that Sally was DOA.  

[+x=usr(1536)]

1 hour ago, reifsnyderb said:

This is what I just did with a dead 800XL for which I discovered a 256k upgrade installed.

The part of this that I'm not looking forward to is desoldering the PIA simply because it's going to mean another socket installation.  Whoever installed this upgrade originally cut pins 12 through 16 at the PCB, bent them up, and soldered directly to them in lieu of fitting a socket and just bending the legs.  Can't reuse the one that's in there as a result (well, not easily, anyway), but at least there are a couple of spare PIAs kicking around here.

1 hour ago, reifsnyderb said:

Eventually, I figured out that Sally was DOA.  

Weird - I replaced SALLY as part of this just to see what would happen.  Got rid of the half-second of red screen on powerup, at least.

 

I've also got 256K of MT RAM in the sockets, so who knows what state that's in after 35 years.

[Mclaneinc]

10 hours ago, x=usr(1536) said:

I've also got 256K of MT RAM in the sockets, so who knows what state that's in after 35 years.

Be afraid, be very afraid...

[+x=usr(1536)]

@ClausB: is there a specific set of instructions you can recommend that I look at for how to install the 5-chip 256K upgrade?  My thinking is that if I'm going to take it out, removing the logic board is the easy part - what I don't know is what needs to be reverted elsewhere on the 800XL.  Knowing how it was installed would mean that I should be able to work backwards from the installation instructions and get the machine back to (mostly) stock.

 

[1050]

http://www.atarimania.com/documents/rambo_manual.pdf

 

Is one source for the instructions. ATB in your project.

[+x=usr(1536)]

58 minutes ago, 1050 said:

http://www.atarimania.com/documents/rambo_manual.pdf

 

Is one source for the instructions. ATB in your project.

Thanks, and I do appreciate the link.  Unfortunately, how the RAMBO installs doesn't match one-for-one what I have in my machine.  There are a lot more jumpers, including ones on the underside of the RAM that just don't exist with the RAMBO.

 

I have been looking over the Byte Magazine article and I think I have an idea of what needs to be changed, but there are a couple of things I want to confirm before tearing into it.

[+x=usr(1536)]

If you happen to have a working 800XL sitting open and could check something for me, I'd appreciate it.

 

Please let me know the in-circuit value of R20 (immediately south of U18) with the power off.

 

I know what the resistance is supposed to be, but would like confirmation of the value I'm seeing.  Not wanting to poison the waters, I'll hold off on saying what the value is that I'm seeing until after someone else's result is posted - or I just say 'to hell with it' and replace it anyway

[reifsnyderb]

17 minutes ago, x=usr(1536) said:

If you happen to have a working 800XL sitting open and could check something for me, I'd appreciate it.

 

Please let me know the in-circuit value of R20 (immediately south of U18) with the power off.

 

I know what the resistance is supposed to be, but would like confirmation of the value I'm seeing.  Not wanting to poison the waters, I'll hold off on saying what the value is that I'm seeing until after someone else's result is posted - or I just say 'to hell with it' and replace it anyway

In circuit, I measured about 620 ohms.

[+x=usr(1536)]

Just now, reifsnyderb said:

In circuit, I measured about 620 ohms.

Thank you, that's much appreciated.  I was getting about 700, so it's at least ballpark-ish.  I think I have some 3KΩ resistors floating around, so may replace it just to rule it out, but have no expectations of it changing anything.

 

I really do think it's time to get that memory upgrade out of there.  Ugh.

[reifsnyderb]

4 minutes ago, x=usr(1536) said:

Thank you, that's much appreciated.  I was getting about 700, so it's at least ballpark-ish.  I think I have some 3KΩ resistors floating around, so may replace it just to rule it out, but have no expectations of it changing anything.

 

I really do think it's time to get that memory upgrade out of there.  Ugh.

There may not be anything wrong with the memory upgrade.  But until you remove it you won't have a baseline to go off of.

[+x=usr(1536)]

2 minutes ago, reifsnyderb said:

There may not be anything wrong with the memory upgrade.  But until you remove it you won't have a baseline to go off of.

Which is exactly where I'm at with it now.  I love the upgrade, but it's so delicate that every time the machine is opened up it's like tiptoeing through a minefield in order to avoid it causing problems.

[+mytek]

10 hours ago, x=usr(1536) said:

Which is exactly where I'm at with it now.  I love the upgrade, but it's so delicate that every time the machine is opened up it's like tiptoeing through a minefield in order to avoid it causing problems.

You probably mentioned it elsewhere, but what kind of memory upgrade is it?

 

[+x=usr(1536)]

10 minutes ago, mytek said:

You probably mentioned it elsewhere, but what kind of memory upgrade is it?

ClausB's quarter-meg, the 5-chip version.  He gave it a positive ID in the following thread:

 

 

[+mytek]

3 hours ago, x=usr(1536) said:

ClausB's quarter-meg, the 5-chip version.  He gave it a positive ID in the following thread:

Here's an option for you, and you can reuse the chips you already have, plus you would end up with a far less fragile situation.

 

 

You can get 3 boards shipped for $16.75 from the OSH Park link in that topic, or with about the same wait time probably around $9-10 from JLCPCB for 5 boards using their slow ship option.

 

Edited July 27, 2022 by mytek

[+kheller2]

On 7/23/2022 at 11:53 AM, x=usr(1536) said:

@ClausB: is there a specific set of instructions you can recommend that I look at for how to install the 5-chip 256K upgrade?  My thinking is that if I'm going to take it out, removing the logic board is the easy part - what I don't know is what needs to be reverted elsewhere on the 800XL.  Knowing how it was installed would mean that I should be able to work backwards from the installation instructions and get the machine back to (mostly) stock.

 

 

You remove the ribbon cable going into U27(?) and replace it with the 74LS158.

You remove all the wires from the upgrade board going to anywhere... and put the 5 pins on the PIA back into the socket.

That should revert it.

 

Of course you should mark the wires and where they go so you know where to re-attach them.  I think you mentioned you had one floating around that you weren't sure where it went, but that can be deduced from the other ones that you do trace.

 

[+x=usr(1536)]

3 hours ago, mytek said:

Here's an option for you, and you can reuse the chips you already have, plus you would end up with a far less fragile situation.

 

 

You can get 3 boards shipped for $16.75 from the OSH Park link in that topic, or with about the same wait time probably around $9-10 from JLCPCB for 5 boards using their slow ship option.

 

Yep, been considering that one as a stand-in until I can get a U1MB in there.  Deciding if I want to recycle the ICs or just keep the board intact as a souvenir of the time and buy new ones.  Not sure how much I trust what's in there now, but I'll figure that out.

58 minutes ago, kheller2 said:

You remove the ribbon cable going into U27(?) and replace it with the 74LS158.

You remove all the wires from the upgrade board going to anywhere... and put the 5 pins on the PIA back into the socket.

That should revert it.

Agreed.  However:

58 minutes ago, kheller2 said:

Of course you should mark the wires and where they go so you know where to re-attach them.  I think you mentioned you had one floating around that you weren't sure where it went, but that can be deduced from the other ones that you do trace.

The one that I wasn't sure about is on the back of the PCB, and appears to jumper pin 1 (IIRC) between every RAM IC.  I'll confirm that later when I'm at home again, but I think that was the only thing that was different.

[+kheller2]

 

That doesn't make sense that there is an additional wire on every pin 1.   Sure, that does if this was something bigger than 256.. but...

maybe there was something wrong with the board and the person fixed it that way.  The extra line should have just gone to the right pad of R32.

 

Take a pic when you get a chance...

[+x=usr(1536)]

1 hour ago, kheller2 said:

That doesn't make sense that there is an additional wire on every pin 1.   Sure, that does if this was something bigger than 256.. but...

maybe there was something wrong with the board and the person fixed it that way.  The extra line should have just gone to the right pad of R32.

R32 is completely untouched - there's nothing extra running to it on either side of the board.

 

Top view (note the 100Ω resistors between pin 2 of each of the RAM sockets and what may be a +5V via):

 

 

And the underside.  The long yellow wire is effectively tying together pin 1 on all the RAM ICs all the way down the line; it ultimately pulls back to the 33Ω resistor on the logic board.  I have no idea why the jumper between pin 5 on U10 and pin 5 on U11 is there:

 

 

Hopefully, this makes sense to someone who actually knows what it is that they're looking at   Leaving it in place for the moment.

[+kheller2]

Wow there is a lot of rework going on there. On the solder side what revision of the board does it say?  I’ll compare it to one I have.  

[TGB1718]

8 hours ago, x=usr(1536) said:

Top view (note the 100Ω resistors between pin 2 of each of the RAM sockets and what may be a +5V via):

Those are meant to be there, they are in series with the Data bus lines pins 2 and 14

[+x=usr(1536)]

11 hours ago, kheller2 said:

Wow there is a lot of rework going on there. On the solder side what revision of the board does it say?  I’ll compare it to one I have.  

I'll let you know when I'm back at the house this evening.  I want to say it's Rev. D, but won't swear to that 100%.

5 hours ago, TGB1718 said:

Those are meant to be there, they are in series with the Data bus lines pins 2 and 14

Interesting.  So were those installed as part of the 256K upgrade?  I'm not finding mention of them in the article I have, which is Claus' original one from Byte, and they don't look like they were put there by the factory.

21 hours ago, mytek said:

You can get 3 boards shipped for $16.75 from the OSH Park link in that topic, or with about the same wait time probably around $9-10 from JLCPCB for 5 boards using their slow ship option.

Went with the OSH Park option.  Decided that for as deep as I'm probably going to have to get into this, I may as well just kill two birds with one stone.

[TGB1718]

33 minutes ago, x=usr(1536) said:

Interesting.  So were those installed as part of the 256K upgrade?

No Factory fitted, Sams Computerfacts shows them in a photo of the motherboard and the circuit diagram

 

[+x=usr(1536)]

7 minutes ago, TGB1718 said:

No Factory fitted, Sams Computerfacts shows them in a photo of the motherboard and the circuit diagram

 

Much appreciated.  At work right now so don't have access to all the usual documentation.