exxos's Atari, Amiga & retro forum

As you've deduced, the black screen with a white border essentially means the machine sees no RAM.

In my experience when you see multiple failures of different bits with the diag cartridge, it more likely means a problem with data and/or address lines.

I'd suggest that you start off by doing a continuity test of them to make sure that they have a connection at all of the sockets.

Check the RAS/CAS lines while you're at it.

I've experienced a specific problem while upgrading one of these to 1MB by adding in the extra chips, and I found A3 and A7 didn't connect all the way through the bank.

I've had similar issues on a Mega ST where D2 and D4 were broken somewhere between the CPU and MMU.

As well as checking the address & data lines, I'd also check all the RAM has power & GND OK, and also make sure there are no shorts between any of the pins.

tzok wrote: ↑Wed Jan 13, 2021 11:19 pm

wietze wrote: ↑Wed Jan 13, 2021 9:21 pm I am using RAM simms that I have succesfully tested using Yaart on another machine.

I see no SIMMs there (SIMMs are used in the STe, here you have just chips in DIP packages), but what I see is the AT-Speed hardware emulator. Have you tried removing it?

You are right; apologies; I mean DIP chips.

Re: AT-Speed hardware emulator; no I have not tried removing it yet.

rubber_jonnie wrote: ↑Wed Jan 13, 2021 11:23 pm As you've deduced, the black screen with a white border essentially means the machine sees no RAM.

In my experience when you see multiple failures of different bits with the diag cartridge, it more likely means a problem with data and/or address lines.

I'd suggest that you start off by doing a continuity test of them to make sure that they have a connection at all of the sockets.

Check the RAS/CAS lines while you're at it.

I've experienced a specific problem while upgrading one of these to 1MB by adding in the extra chips, and I found A3 and A7 didn't connect all the way through the bank.

I've had similar issues on a Mega ST where D2 and D4 were broken somewhere between the CPU and MMU.

As well as checking the address & data lines, I'd also check all the RAM has power & GND OK, and also make sure there are no shorts between any of the pins.

Awesome, thank you for these pointers. I will get to this tomorrow after work!

Thanks for the response so far guys! You have been really helpful so far!

@wietze No problem.

You may also want to look here: https://www.exxosforum.co.uk/forum/viewt ... 220#p36270

It'll drop you in the middle of some RAM troubleshooting for one of my Megas, it may help.

Ok, Ive started tracing continuity, and I found a first problem.

I traced:
R59 to pin 4 to the first 2 banks of ram (ok)
R60 to pin 15 to the first bank of ram (ok)
R61 to pin 15 to the second bank of ram (ok)
R62 to pin 1 of all (ok)
R63 to pin 9 of all (ok)
R64 to pin 13 of all (ok)
R65 to pin 5 of all (ok)
R67 to pin 6 of all (ok)
R68 to pin 12 of all (ok)
R69 to pin 11 of all (ok)

R70 to pin 7 and 10 of all (nok!)
R66 to pin 7 and 10 of all (nok!)

So there is an unwanted connection somewhere.

Pin 7 and pin 10 are exactly the 2nd pin of the left from both the top and bottom of the soldering points, so I suspect there is a connection as a result of my soldering the sockets, somewhere between the holes.
I have subsequently visually checked my soldering, but I realized I made a mess out of the board that it makes visual inspection quite difficult. I tried to detect the connection by touch, by scanning the area between soldering holes with my multimeter, but so far, no detection yet.

I will first remove all chips from sockets and inspect some more; but I suppose without having a clue as to how to detect the short; I should probably remove the sockets so I have a clean slate again.

Ok, admittingly, after a lot of cleaning, the board looked less damaged than I initially thought. It was not so much that the board was corroded or something, it was just a lot of flux that I didnt clean and left a kind of dusty mess. After some cleaning I had this:
From visual inspection this the back side of the board shows no shorts between said pins. Im not sure how likely it is that the solder would get through the hole and short on the other side of the board, given the large amount of flux I used when soldering it...

Alternatively, the short is made elsewhere. I will go have a look.

I can see a few damaged vias.. as to if they are connected to anything or not ...

exxos wrote: ↑Thu Jan 14, 2021 9:39 pm I can see a few damaged vias.. as to if they are connected to anything or not ...

Would you please mark so I can learn what to look at / how to identify?

The pads are damaged or missing totally... I just list 3, there are probably more.

Thank you for pointing that out to me. I checked those pads, but they dont seem to be causing the short. I tried to check the routes from the resistors into the memory area, but I can not find any kind of shorts there.

If there are no other suggestions, I will start desoldering the sockets tomorrow and see if the short still remains.