exxos's Atari, Amiga & retro forum

One may be working... though the original may have taken out some of the SRAM chips.

Gah! Where did I put those SRAM chips I ordered from Mouser?!!!!

Found the chips, hidden under something in a place I'd usually not put chips. (It was near where my work bag sits.)

Replaced 4 of the SRAMs, no difference to cache error. Replaced the other set of four original chips with the ones I'd taken out and... The cache is now workin! Cachechk.exe shows no issues. :-)

So, now I need to diagnose the start-up issue. The first thing I'll check is removing the CPU and see if the RESET is working correctly. I can also see what the values of the jelly bean tants are and order a whole lot probably.

Let's see what Vogons users say. I've posted on there. Hopefully there's a guru or two who know the PC architecture backwards can help with the reset issue.

Well, with some hints from the Vogons people pointing in the right area of the board I think that I've found the problem...

If I press down on the 82C806 chip then the system works. Looks like a dry joint on one of the SMD pins, so I'm going to have to do a reflow of the solder.

stephen_usher wrote: 29 Jan 2022 12:43 Well, with some hints from the Vogons people pointing in the right area of the board I think that I've found the problem...

If I press down on the 82C806 chip then the system works. Looks like a dry joint on one of the SMD pins, so I'm going to have to do a reflow of the solder.

Nice find, hopefully an easy fix too :)

Except it wasn't, it only seemed that way...

Reflowed the solder on the 82C206, along with some pull-ups and capacitors which seemed make a difference if i touched them, but no change. Then I noticed that if I held the ROM in a specific way it would start, until it didn't, so I changed the ROM socket with a new 3M one... no change.

I think what was happening was that my fingers were changing the capacitance in the area, influencing how the system behaved.

So, the next thing I'll do is order the replacement 10uF 16V tantalum jelly bean caps and replace those.

stephen_usher wrote: 29 Jan 2022 16:17 Except it wasn't, it only seemed that way...

Reflowed the solder on the 82C206, along with some pull-ups and capacitors which seemed make a difference if i touched them, but no change. Then I noticed that if I held the ROM in a specific way it would start, until it didn't, so I changed the ROM socket with a new 3M one... no change.

I think what was happening was that my fingers were changing the capacitance in the area, influencing how the system behaved.

So, the next thing I'll do is order the replacement 10uF 16V tantalum jelly bean caps and replace those.

Don't you just hate it when that happens :(

At least the caps are cheap, though I suppose you could just stick your hand inside the PC whenever you want to boot it :)

If only sticking my hand in worked all the time. For a few resets it'll work in one position, but never again... Then try another and another and hope. :-)

I'm coming more and more to the conclusion that this is an issue with one of the three custom chips.

When I press the reset button the amplitude of the processor clock decreases markedly and the frequency doubles. In fact it looks a lot like the raw output of the crystal. The 82C482 chip deals with the system clocks, reset and some of the memory management.

If the CPU isn't being given a clock signal which is within specification during the time the reset signal is high then it's likely that it's not resetting properly as it needs 15 clock cycles to do so, which is probably why it's not properly trying to read the ROM.

The trouble is that without either a working reference system or a schematic it's a case of shooting in the dark.

I may pack the machine away and get back to the Apricot stuff.