exxos's Atari, Amiga & retro forum

@exxos

I had 2 takes on this kind of report, but the 000008 I believe is the start address for the RAM tests. Based on this I think you're looking at bank 0.

Interpretation 1 was that the first block is the test value, and the second block is what's returned.

Interpretation 2 was that the first block is bank 0 and the second block is bank 1.

On thing I am confident of is that both blocks represent bits 0-15, with bit 0 at the RHS and bit 15 at the LHS of each block.

Problem is my service manual is missing the section on interpreting!

The RAM/Data line error definitely means a data line problem, in my Mega it was D2 & D4, there wasn't a problem with the RAM.

In the end I used a logic probe to test against what I'd expect from a working machine then buzzed the lines out when I found the bad data lines. Looks more like an address line problem though.

rubber_jonnie wrote: 27 Apr 2020 23:52 Problem is my service manual is missing the section on interpreting!

Yeah know what you mean... biggg post to follow any second...

So A0 tested short to GND. The waveform on my scope was basically just noise which is why it caught my attention. On first inspection I couldn't see anything wrong... So back to the low resistance meter...

Testing GND to A0...


TEST POINT 1 - 0.441R


TEST POINT 2 - 0.396R - Getting closer...


TEST POINT 3 - 0.357R - Getting closer...


TEST POINT 4 - 0.316R - Getting closer...


TEST POINT 5 - 0.275R - Getting closer...


TEST POINT 6 - 0.229R - Getting closer...


TEST POINT 7 - 0.185R - Getting closer...


TEST POINT 8 - 0.141 - Getting closer...


TEST POINT 9 - 0.182 - GONE UP!

So ah-ha! TEST POINT 8 is where the short is.. On the second to last socket ont he bottom right of the board... sooo... Magnifier time...


Marked A0 with marker pen to see where it was going.. Something looks suspect...


Zooming in.. so tiny....


Re-solder....


and...........




:cheer: :girldance: :cheer: :girldance: :cheer: :girldance: :cheer:


So what have I learned in all this ? I'm glad I don't do this for a living anymore :lol:

LOL, well done and a very informative way to test for shorts.

Re ram test, I always presume that the bit indicated with a 1 has a problem and work from there as the manual is less than hopeful with its descriptions. Presume that left 16 bits is 1 bank and right is another. This is how I found the short on my mega the other day.

PhilC wrote: 28 Apr 2020 06:46 Re ram test, I always presume that the bit indicated with a 1 has a problem and work from there as the manual is less than hopeful with its descriptions. Presume that left 16 bits is 1 bank and right is another. This is how I found the short on my mega the other day.

Guys, there's no need for guesswork. It's as simple as this with the initial diagnostic cartridge RAM test:
Expected value (bits D15-D0), value actually read (D15-D0), address (in hex).
The affected bank you have to decode yourselves from the faulty address.

@czietz thanks.

@czietz Cool, so the first interpretation I had was the correct one, that's good to know. The service manual(s) I have don't seem to describe this at all well, other than working out which bank it is.

@exxos Good job, and proves that a multimeter is sometimes all you need :D

:girldance: :girldance: :girldance:

rubber_jonnie wrote: 28 Apr 2020 08:46 @exxos Good job, and proves that a multimeter is sometimes all you need :D

Yeah, I think the diagnostic tests "lead me up the garden path" into thinking there was a faulty data bit, when actually there wasn't, it was a address fault.

I3 RAM address check 0000000000001000 0000010000001010 000008

8 = 1000 so that must be the address. it gets back 1010 = 10. Even so it doesn't make sense. As if A0 is shorted to GND, then "1000" is still A0=0.. So it shouldn't have mattered anyway at that point.

I have been talking to @Icky off and on about doing a better low level diagnostic board for the past year. Mostly intended to debug a dead H4 but could be used on ST's as well. As when the machine is "dead" nothing runs from ROM anyway.. It would be like a "CPU EMULATOR" which would have a bank of LED bits to relay data on the bus properly.

exxos wrote: 28 Apr 2020 10:27
I have been talking to @Icky off and on about doing a better low level diagnostic board for the past year. Mostly intended to debug a dead H4 but could be used on ST's as well. As when the machine is "dead" nothing runs from ROM anyway.. It would be like a "CPU EMULATOR" which would have a bank of LED bits to relay data on the bus properly.

@exxos A proper 'Dead Test' cartridge would be a great idea.

I know when my Mega 4 was first on the workbench with the dry joint on the ROM socket GND, even with the diag cart, I got nothing out of the serial port, presumably because the diag cart couldn't be mapped into the ROM space.

A cart that will work and report back, even on a dead board would be a great tool to add to the collection.