Is my 600XL doomed?

[Grumps]

7 minutes ago, candle said:

one more thing

if your machine boots up, try twisting motherboard

if there will be any sort of reaction - random hangs for example - try to localise this with more subtle methods and then replace faulty socket

single wiper sockets used back then by Atari don't get high scores in my book

 

Does removing and re-inserting the chips (all of them) clean any corrosion from the pins?

[candle]

you can burn gal16v8 for MMU replacement

jed file should be somewhere (pigwa perhaps?)

 

but what i don't like about this is how many chips gone bad in this unit - this looks like voltage regulator failure - you do use new power supply now, don't you?

 

about corrosion -it will clean some - it all depends on the damage done - if pins have gone green - there is no hope for this to happend

there are also cleaners for such cases - disolving sulfides and oxides that forms on these pins

look for "kontakt chemie" product line - you should find some, but i would rather replace them

 

Edited February 11, 2020 by candle

[Grumps]

My PROM programmer doesn't support GALs.

 

I don't know how many chips are bad. All I can tell is that the OSROM reads OK in the programmer.

And there is the correct clock from the crystal to GTIA, then from GTIA to ANTIC, and then from ANTIC to 6502.

 

The original old power supply is giving 5.1V and appears smooth using the scope.

 

[Rybags]

If you've got a programmer then couldn't you make up a custom OS?

 

It'd not need a big programming effort, just a bit of code to do colour bars or something.

But my suspicion is that you have a dud CPU or MMU which is preventing the OS from starting up.

[candle]

ok, please try with that rom based tester i've posted link to - it should give some valuable information on how to proceed

 

[Grumps]

12 minutes ago, candle said:

ok, please try with that rom based tester i've posted link to - it should give some valuable information on how to proceed

 

Will do. I've downloaded the latest ROM file. Should be able to try later today.

Ta.

[Grumps]

So I've tried both a new OSROM and that ROM based tester. Still only have a red screen.

 

I have checked continuity from the IC pins (directly on the devices) to the solder side of the board, and ALL connections are good. There is obvious corrosion/discoloration of the pins though, so maybe the next step is to replace ALL of the sockets.

 

[shoestring]

Have you tried another MMU ?

[Grumps]

I'd like to. I'd also like to try another CPU. But am having difficulty finding spares in UK/EU.

 

I'd also like to try and burn the MMU (PAL/GAL) myself, but my programmer needs an upgrade to do that.

 

[candle]

you could test mmu for basic operation - this is only combinational logic so all you need is some kind of test vector (jumper wires) and breadboard

i know you already replaced RAM, but if you do have a scope, please check RAS to CAS to PHI2 timings and WE line on ram chips

try also to run it without BASIC chip installed - it might be dead aswell

perhaps some bits are stuck either low or high 

 

[mimo]

1 hour ago, Grumps said:

I'd like to. I'd also like to try another CPU. But am having difficulty finding spares in UK/EU.

 

I'd also like to try and burn the MMU (PAL/GAL) myself, but my programmer needs an upgrade to do that.

 

https://www.ebay.co.uk/itm/NEW-Atari-800-XL-XE-computer-console-MPU-CPU-IC-chip-C014806-6502C-Sally/162081111036?epid=28007777933&hash=item25bcc97bfc:g:9esAAOSwoydWooXL

[shoestring]

I have a few spare GALs on hand and can burn one for you.

 

Inbox me if you don't have any luck. 

[Grumps]

Thanks all for you kind offers of help.

 

Update:

I re-soldered ALL IC pins and then gave it a good clean using a foaming cleaner designed to remove flux/rosin residue, then a rinse and dry.

Screen is still red - as expected really.

 

Then I removed all chips and used Kontakt Chemie 60 on the sockets and IC pins. Waited. And re-tested. Screen still red.

 

So, then I dug out my old (unused for a very long time) HP logic analyzer. Of course I must've put it away in a non-working state as it needed a new fuse.

 

So, I have the analyser looking at the address and data busses, and all of the PAL outputs, and IRQ, and RST.

I can see that immediately after a RST (the reset pin going high) that Sally starts to push 3 bytes onto the stack. The stack address is most often 0x14C, but sometimes 0x1FF. It looks like she's not sure what she's doing!

Then she reads the RST vector from 0xFFFC/0xFFFD. I can see the correct vector coming from the ROM (I've got a rev02) as address 0xC2AA. Sometimes Sally will start to execute from this address, and sometimes she will execute from address 0xC0A8. This looks like she didn't see data bit 1 correctly from the vector location.

She also does 3 more pushes to the stack, then reads the IRQ vector from 0xFFFE/0xFFFF. The IRQ pin is definitely not active. And then sometimes just stops executing, or just goes off into oblivion.

 

It's been quite fun, but also tricky to see exactly what's going on as ANTIC is on the bus regularly reading display stuff I assume.

 

So, if the panel agrees, I'm going to try a new Sally.

[SS]

I am still leaning toward the idea that an ANTONIA board would be the right answer for this 600XL.  The ANTONIA would completely replace the SALLY, the MMU, the RAM, the DELAY, the OS, and the BASIC chips.  

[shoestring]

It most likely would. But I understand the compulsion behind wanting to know what is broken on the board, half the fun stems from troubleshooting and what you learn from it.

[Rybags]

The spurious stack push at the start is probably valid - the OSes do a LDX #$FF / TXS as part of the reset routine so it's likely the 6502s all have an indeterminite stack value at that time.

 

The IRQ thing though is still very strange.  The OS clears the hardware registers before IRQs are enabled, so you should have a black background before any IRQs can occur.

[Grumps]

On 2/15/2020 at 12:34 AM, SS said:

I am still leaning toward the idea that an ANTONIA board would be the right answer for this 600XL.  The ANTONIA would completely replace the SALLY, the MMU, the RAM, the DELAY, the OS, and the BASIC chips.  

I'm not even sure an Antonia board would solve the numerous problems I'm seeing.

 

The more I investigate (which I am enjoying), the more questions/problems I have.

 

E.g. When the system is powered on, immediately IRQ becomes active. No delay.

If I remove both the PIA and Pokey, then IRQ is pulled high as it should be.

Sally still doesn't read the RST vector from the OSROM though (it gets data bit D1 wrong), but I can see using the scope that D1 is good and proper.

According to the PIA datasheet, all registers are cleared when RST - this would mean it should not assert IRQ, but it does. So this means a new PIA (in addition to Sally which is on order).

 

Pokey is not blameless either. Without the PIA present, it looks as if Pokey is doing something bad with IRQ. Not low, nor high, but wobbling about mid-state!

And, the whole data bus is continuously pulled up about 0.8V when Pokey is present.

 

How far through boot does it go without a PIA and Pokey present?

 

[+DrVenkman]

Check your sockets. 

[Grumps]

I have previously tested continuity between IC pins and solder side of the board. This took quite some time. All was OK, so I assume the sockets are too.

But in addition I also removed all chips and use Kontakt Chemie 60 to remove any corrosion.

 

If, for example, the PIA has power and ground (which it has), and has a valid RST signal (which it also has), and the other control signals are correct (which they are), then it should not assert IRQ (which it sadly does).

All signals have been scoped directly on the PIA's pins.

 

Does anyone know of a ready supply of PIA (or compatible) parts?

 

[candle]

there should be pullup resistor on IRQ line, so what you see could not happend

if you remove PIA, configuration for PORTB is done through resistors on PORTB data lines, thus it should boot without this chip

 

as for pia supply, try mc6821

https://www.aliexpress.com/wholesale?catId=0&initiative_id=SB_20200219021944&SearchText=6821

 

[Grumps]

Yes, there is a pull up. If Pokey and PIA are not inserted, then IRQ is pulled high.

 

Also, I put my new Sally in and there is some progress.

First time power on and I got the standard blue screen with a white block cursor. No READY though.

Then I tried that diagnostic ROM (link posted earlier). The attached shows the results.

 

But, after a few more power cycles I cannot get the board to boot even a little. Just a flickering screen, but no discernible image.

[candle]

you won't get to ready without pokey chip in place

 

[Grumps]

OK. If Pokey is in place then the machine doesn't get to the blue screen. Looks like it's time for a new Pokey.

And I'm pretty sure PIA is dead(ish) too. Constantly asserts IRQ.

 

Why are there so many dead chips on this board?! Power supply is OK, so I don't know why so much has died.

 

[Grumps]

1 hour ago, candle said:

there should be pullup resistor on IRQ line, so what you see could not happend

if you remove PIA, configuration for PORTB is done through resistors on PORTB data lines, thus it should boot without this chip

 

as for pia supply, try mc6821

https://www.aliexpress.com/wholesale?catId=0&initiative_id=SB_20200219021944&SearchText=6821

 

Ah, the aliexpress gamble. Have you had good experience with chips from them? Or anybody else?

[+DrVenkman]

2 hours ago, Grumps said:

Yes, there is a pull up. If Pokey and PIA are not inserted, then IRQ is pulled high.

 

Also, I put my new Sally in and there is some progress.

First time power on and I got the standard blue screen with a white block cursor. No READY though.

Then I tried that diagnostic ROM (link posted earlier). The attached shows the results.

 

But, after a few more power cycles I cannot get the board to boot even a little. Just a flickering screen, but no discernible image.

I’ve never used Shoestring’s test program, but does it usually cause those vertical PM/G bars? If not, you’ve got GTIA problems too, looks like.