My 610 board build.

[ronin47]

I misunderstood what you said. I thought you were saying that hitting <return> for memory size caused it to report 9471 bytes free. If it is still saying 7423, then it is not seeing the extra RAM, or it is testing bad. You could go into the Monitor and look at what it shows for location 2000. Try changing the contents, then re-entering the same address to see what it shows, e.g. .2000/AA.2000.

Yes, if you enter D with no disk drive it will hang forever waiting for a response from the drive. Entering a memory size greater than you really have just sets the top-of-memory to what you entered, with no testing. For example, if you entered 8500 but didn't have that much RAM, and you then entered an immediate-mode statement like PRINT "ABC" what would actually be displayed would be !!!, since it would show what it reads at its "top of memory" which is on page 21 hex, or decimal 33, ASCII for "!".

Based on the designations CE0 and CE1 on the schematic, I would also think that U19 and 27, and U20 and 28, would be the locations for the lowest-addressed RAM chips ($2000-$27FF). But looking just at the parts-placement guide, I would guess U51/59 and U52/60 instead. But the enable lines for those are labeled CE16 and CE17, so I guess not.

Could any of the buffer chips or the address-decode logic have been damaged when you connected the board with the bad cable? IIRC, the 8T28 (and 8T26) buffers can be easily damaged; if a good one on one board connects to a bad one on another board, the previously good one becomes bad, and both have to be replaced.

Thanks for all the above, I'm locked in a work call right now, so can't go and look, but I will later.

I assumed it was going to be the case that 7423 meant it didn't see the RAM...

I have already tried U51/U59 and U52/U60, same result. I will try and test the RAM in the main machine. Not sure how I can test the 8T28's.

And of course it might be bad RAM.

[HOWARDHB]

The 40way interconnect cable gave me a headache too, as I had to make one end "reversed" owing to the need to plug it onto the underside of the 600D, on my "reversed" pcb stack...
I initially made the "reversed" end by soldering 40 individual ribbon cable wires to a 40way socket.
I secured (for strain relief) the solder joints with hot-melt adhesive, and discovered that, combined with some remnant (hygroscopic) solder flux, it caused intermitant memory issues!
I subsequenly realised that a far easier method was to split the ribbon cable wires into pairs (only the end that needed to be reversed), and "flip" or twist each pair over, and crimp on a new 40way IDC header.

I read that either N8T26 or N8T28 can be used to buffer the data bus, but you can't have a "mix" of 28's and 26's, which implies "across the entire bus", 600D and 610 combined.
The address bus buffers on the 610 have to be N8T95, which also buffer R/W and 02 clock.

RAM chip pair U19 and U27, is the first expansion location (addressed at $CE0).
I have verified that 1k will be detected at startup with only U19 and U27 populated.

... Eagerly awaiting your first Gotek boot-up!

[ronin47]

The 40way interconnect cable gave me a headache too, as I had to make one end "reversed" owing to the need to plug it onto the underside of the 610, on my "reversed" pcb stack...
I initially made the "reversed" end by soldering 40 individual ribbon cable wires to a 40way socket.
I secured (for strain relief) the solder joints with hot-melt adhesive, and discovered that, combined with some remnant (hygroscopic) solder flux, it caused intermitant memory issues!
I subsequenly realised that a far easier method was to split the ribbon cable wires into pairs (only the end that needed to be reversed), and "flip" or twist each pair over, and crimp on a new 40way IDC header.

I read that either N8T26 or N8T28 can be used to buffer the data bus, but you can't have a "mix" of 28's and 26's, which apllies to the entire bus, 600D and 610.
The address bus buffers on the 610 have to be N8T95, which also buffer R/W and 02 clock.

RAM chip pair U19 and U27, is the first expansion location (addressed at $CE0).
I have verified that 1k will be detected at startup with only U19 and U27 populated.

... Eagerly awaiting your first Gotek boot-up!

Thanks Howard I just flipped the connectors on my cable, my stack is like yours, but I do have pin 1 on the bottom going to pin 1 on the top. Cable was a nightmare

I still need to look at the 610 setup and adjust all the pots. Pretty sure I have nothing but 28's.

Thanks for confirming the RAM setup, I do have stuff in place, but I think before I proceed further I need to test the 2K in the main board, make sure it isn't duff.

I do see activity on the RAM IC's BTW. address and data too I think. I will check.
I think I need to get the scope out before I hook any Goteks up!

[ronin47]

OK, so I managed to get a bit of diagnostics in and _WE on the RAM is under 3 volts and looks rather weird.

Then there is _CE. which looks odd too and seems to be stuck at 5v.

The manual states that RAM must be added in CE order which is U19+U27 (_CE0) and U20+U28 (_CE1) so that confirms they are in the right spot.

Address lines look OK, though some of the voltages look a bit low. I am running everything of my bench PSU (5v and just under 2.5A for both boards) but does the 610 require its own PSU? Again the manual suggests it should be powered from the 600 but wondering what other people do.

IO lines seem OK too, though ID0 looks a bit low too.

_CE0 and _CE1 seem to come from U15 (Pins 15 & 14) so I will check and see if that is OK, and U71, the 6850 (Pin 13) seems to be dealing with _WE. I'll check those out.

[HOWARDHB]

I'm running both boards from a single switch-mode PSU, with 5V, 13A and 12V, 5A outputs.
If you use a single or two separate supplies, at least connect the 600D and 610 pcb ground foils together with a suitably thick wire, as high currents through the (multiple) ground wires in the ribbon interconnect cable may cause unpredictable problems.

Most of the processors control logic signals are "biased" a little above 2.5V, and not merely "tied" high or low.
(220R to 5V and 390R to 0V)

[ronin47]

I'm running both boards from a single switch-mode PSU, with 5V, 13A and 12V, 5A outputs.
If you use a single or two separate supplies, at least connect the 600D and 610 pcb ground foils together with a suitably thick wire, as high currents through the (multiple) ground wires in the ribbon interconnect cable may cause unpredictable problems.

Most of the processors control logic signals are "biased" a little above 2.5V, and not merely "tied" high or low.
(220R to 5V and 390R to 0V)

Thanks Howard, useful info.

I did find a solder bridge earlier and got all excited, but having cleared it I still don't see the RAM

[ronin47]

Just to help out with the troubleshooting, I made a diagram of J2 on the 610.

Screenshot 2025-04-27 at 09.35.31.png (163.45 KiB) Viewed 70 times

Cross checking this with J1 on the 600, and it doesn't seem to match, am I missing something here? I thought the connector was 1 for 1 straight through.

Screenshot 2025-04-27 at 09.53.58.png (1.38 MiB) Viewed 70 times

Ot have I made a rookie error?

[ronin47]

I think I have made a rookie error, whilst J2 is correct, it isn't where the main 600 board plugs in, so I need to go back and make diags for J1 on both the 600 and 610.

I obviously got an idea in my head that J1 on the 600 plugged into J2 on the 610 and the other socket on the 610 was J3. Not sure why I thought that, but I have it correct now.

J2 should be a good point for probing signals though

[ronin47]

OK, head successfully straightened:

600 board J1:

Screenshot 2025-04-27 at 14.42.48.png (58.3 KiB) Viewed 64 times

610 board J1:

Screenshot 2025-04-27 at 14.43.31.png (58.6 KiB) Viewed 64 times

And 610 board J2:

Screenshot 2025-04-27 at 14.43.48.png (59.39 KiB) Viewed 64 times

Not sure it'll help anyone other than me, but I at least know that is is straight through pin connection from the 600 to the 610 now (I thought it was but nice to confirm) and I can start to trace the signals to see what's what.

[ronin47]

OK, so I did a test of the signals leaving the 600 at J1 and all looks fine.

Of note there is R/W activity that looks normal there and also coming into the buffers, but seems to be getting slapped down quite low but the CE0/CE1 stuff is just non-existent, especially when compared to the RAM on board the 600.

The RAM is tested and good BTW, I can't recall if I said. I have more RAM arriving today hopefully.

So yesterday I pulled all the 74 logic IC's one at a time and tested them. They are all good, but the outputs for /CE0 and /CE01 have nothing.

I need to check what inputs go together to make the /CE0 and /CE1 outputs as I am sure something is missing, and these are signals generated on the 610 itself.

Will be annoyed if my buffer ICs are bad...

But without /CE0 and /CE1 I'm going nowhere. R/W comes up to U2 just fine:


This is R/W at pin 14 of U2:

IMG20250428121118.jpg (4 MiB) Viewed 55 times

And this is what it looks like at the output as BR/W, pin 13:

IMG20250428121213.jpg (4.2 MiB) Viewed 55 times

And what it looks like at one of the RAM IC's, pin 10:

IMG20250428121304.jpg (4.2 MiB) Viewed 55 times

So it looks like the output from U2 is what the RAM ICs are seeing, but that voltage just looks way to low to work. Saying that the RAM datasheet does say those voltages are good, so high being 2v to Vcc and low being -.5v to 0.8v.

Maybe I'm worrying about BR/W too much and should focus on /CE0 & /CE1.