Reconstructing an auctioned Lisa

[berskyboy]

Yours is part number 341-0348, which is for Macintosh XL computers that boot only Mac OS. If you want to boot LOS, you'll need to remove that PROM and replace it with 371-0072, and disconnect the transformer between the Lisa's Video Board and the CRT - thereby enabling Lisa Office System's native resolution.

This is what shows up in Mac XL with the black borders on each side, and when I put the CRT P1 and P2 it went directly into the Lisa Video Card, so maybe it's still stock CRT?

https://www.dropbox.com/scl/fi/bour2rbj1jygmhpvmg7fd/IMG_4174.jpeg?rlkey=ecybjape2jj2gg65qdu3mzf0l&st=2fpr9klf&dl=0

[berskyboy]

There is an adapter design to use an EPROM in lieu of the one-time-programmable 20 pin bipolar PROM, but I've not tried it.

Note that the Video PROM contains the Lisa's Serial Number, so there are numerous variants of the programmed contents of this part (containing different serial numbers).

I believe you are in Canada with me, so I may be able to provide these; contact me privately to discuss.

sure thing, I did drop you an Message.

[TorZidan]

I tried the Piggyback method still getting
00000186 0000 8000 0000 0000 0000 0000 0000 0000

00000186  0000 8000 0000 0000 0000 0000 0000 0000
https://www.dropbox.com/scl/fi/kvfsgapov3rf3fizzh8ik/IMG_4151.jpeg?rlkey=ke4waclbeexuouu5rm4jqhzee&st=e0t9slf7&dl=0

every time I ask AI to tell which chip it says different chips.  Is there a lookup document? 
I tried this website but it's down

https://lisafaq.sunder.net/single.html#lisafaq-hw-mem_error

Here is a copy/paste from my records:

Each memory board has 4 banks (of 16 chips each), named "A,B,C,D".
Each "word" (4-digits in your post) corresponds to a memory board + "bank".
Look above for banks that show non-0000: they have faulty ram chip(s).
If there is one memory board in slot MEM 1, then the 8 words are for the following banks:
  B C D E  x x x x          (x means "unused")
If there is one memory board in slot MEM 2, then the 8 words are for the following banks:
  E D C B  x x x x
If there are two memory boards installed, then the 8 words are for the following banks:
  MEM 2    MEM 1
  E D C B  B C D E

Now that we determined the board and bank(s) that have bad ram chipss, we need to determine the ram chip column number in each bank:
Word  Bit#= Column Number
  0001  0  =  22
  0002  1  =  21
  0004  2  =  20
  0008  3  =  19
  0010  4  =  18
  0020  5  =  17
  0040  6  =  16
  0080  7  =  15
  0100  8  =  1
  0200  9  =  2
  0400  10  =  3
  0800  11  =  4
  1000  12  =  5
  2000  13  =  6
  4000  14  =  7
  8000  15  =  8
If there is more than one bad chip in a given bank, there will be more than one bit set and you'd get a word different that the ones above.
In this case, proceed to identifying the bad chip similarly to how it's done  on a Macintosh 128k.


So let's assume you have one RAM board in slot MEM1 (verify this!!!). The error is 8000 in Bank C on the board, which is the ram chip in column 8. So pull tooth C8.

[berskyboy]

So let's assume you have one RAM board in slot MEM1 (verify this!!!). The error is 8000 in Bank C on the board, which is the ram chip in column 8. So pull tooth C8.

I put in piggy back chips in C8 and D8 (both one at a time and together)

got
00000186 0000 8000 0000 0000 0000 0000 0000 0000

and
00000186 0000 0000 8000 0000 0000 0000 0000 0000

and even
00000186 0000 8000 8000 0000 0000 0000 0000 0000


AI said this

The Real Problem: Shared Logic Components

When multiple entirely separate blocks across two different boards fail on the exact same bit, it points directly to the logic gates handling that bit array line. You need to pull out the boards and inspect the Logic Multiplexers / Transceivers that drive Column 8.

Look at the logic chips at the bottom or sides of the memory board PCBs (typically 74-series logic). Specifically, look for:

1. 74LS244 or 74LS245 Octal Buffers/Transceivers: These handle the data bus lines coming off the RAM columns to send them back to the Lisa motherboard. A single failed gate inside a 74LS244 handling Data Line 15 will cause every single row block in Column 8 to fail simultaneously on that board.

2. 74LS153 or 74LS257 Multiplexers: These route the address strobe lines (RAS/CAS) across the row selections.

Not sure where 74LS244 is but I did examine and touch each chip RAM card still looks OK, and when I boot into Mac XL is shows 908K with a known good 512RAM card and my failing 512RAM.  But it doesn't load with just the 512RAM card (but it was in slot 2, with slot 1 empty).

[berskyboy]

So let's assume you have one RAM board in slot MEM1 (verify this!!!). The error is 8000 in Bank C on the board, which is the ram chip in column 8. So pull tooth C8.

so I did the following

RAM      Slot        Error      Mac XL                  Service Code
B, A      1, 2         70          Works show 908k  00000186 0000 8000 0000 0000 0000 0000 0000 0000 (D8)
B          1             passes   Works show 304k  00000186 0000 0000 0000 0000 0000 0000 0000 0000
A          1             70          two beeps              00000186 0000 0000 0000 0000 0000 0000 8000 0000 (D8)

so it seem D8 is the culprit, I did piggy back it, here's the photos

Also I got an error 75 once and sound cut out twice also.

https://www.dropbox.com/scl/fi/4v4sdpmig7z0m5hokv48e/IMG_4180.HEIC?rlkey=jhvf8iqqthiq81sveanoociq5&st=6woi5hrl&dl=0
https://www.dropbox.com/scl/fi/5ie0w2jxpttjy83tf8uy8/IMG_4181.HEIC?rlkey=4q2ju9g0drmdiwmagc0grqvlc&st=748nxi4s&dl=0

 

[berskyboy]

Is there a longer test to run on the RAM other than Apple - S, option 1 and 186 20?  I tried the piggyback method and it didn't fix my bad RAM card.

[sigma7]

I tried the piggyback method and it didn't fix my bad RAM card.

Please provide a link to the method you tried... is it Lisa specific?

Regardless, it doesn't always work as its success depends on the failure mode of the chip underneath (assuming the chip underneath is indeed the problem).

The lisafaq description of how to interpret the self-test error breadcrumbs in memory is, AFAIK, reliable. So unless those breadcrumbs are not giving consistent results, I'd suggest proceeding to remove the faulty chip, install a socket, and try a new chip. It is very difficult to desolder the power and ground pins, you may want to cut them and desolder them after desoldering the other pins and removing the chip. You may need two irons, one on each side of the board; that requires 4 hands of course.

[berskyboy]

This was the website that I used to try and piggy pack and now desolder and put in a new chip.

https://www.classic-computers.org.nz/blog/2011-04-02-fixing-lisa-memory-cards.htm

I bought this on Amazon,

https://www.amazon.ca/dp/B082FMHB8G?ref=ppx_yo2ov_dt_b_fed_asin_title

and it beeps when I put it in and gives the checkered screen

https://www.dropbox.com/scl/fi/7qvb7sf109y2ogthg9bdb/IMG_4208.jpeg?rlkey=pp80rcsjxdffii1w4cesl7ipd&st=eexwt6xp&dl=0

black arrows point to no conductivity (but seems to match other pins on other chips)
https://www.dropbox.com/scl/fi/v0gmisbvq8g7n1hb48j38/IMG_4207.jpeg?rlkey=johw3ip0ua33umk0qi3zps9yl&st=flv474rs&dl=0

new chip (I put in before I read your post, but I can take it out and put in IC socket)
https://www.dropbox.com/scl/fi/udasvw1ldexl6zkv071fr/IMG_4198.jpeg?rlkey=s4e504ha17lwagzohia7axocq&st=w8wbx23r&dl=0
https://www.dropbox.com/scl/fi/2q4ow98nm5mhc28zuonud/IMG_4197.jpeg?rlkey=3ovmpupfwgr5rofmgeg3rqtj4&st=kb05mjpj&dl=0

[berskyboy]

Here's an YouTube Video

I did a continuity test, which matches a known good chips

https://www.youtube.com/shorts/YHyCjWwFuY4

[berskyboy]

I just took out the chip and will put a IC Socket but I wanted to see what I need to fix for my traces, and if I need some jumper wires.

https://www.dropbox.com/scl/fi/r4yvw5suqukk743096xae/IMG_4213.jpeg?rlkey=sllkjv6k48uzj65j0jneflmaw&st=ku1zfgbh&dl=0

https://www.dropbox.com/scl/fi/5ypvuvtqe06ut4fhipsju/IMG_4214.jpeg?rlkey=0yxlk8nba1d9s0uhkubg60r6g&st=vgdvok9t&dl=0

[berskyboy]

UPDATE:  I did put in a IC Socket and checked the continuity and they reach all the way down to the bottom of the board.

Pins 1,2, 4,5,6,7,8,9,10,11,12,13,14 all beep to the corresponding chip.  Leaving 3, 15,16 that don't but I think that's by design. 

I put the problem RAM board in and it gave the following error (1 512K RAM in slot 1)

00000186  0000 0000 FF00 FF00 0000 0000 0000 0000

then I tried with two RAM modules and it gave a checked screen.  Took out the known good RAM leaving the problem RAM still checked screen.  With the know good RAM in slot 1, it boots fine.  So still and issue with the problem RAM.  I wish there was a better way to test the troubled RAM board.  (or maybe I need a replacement?)

[sigma7]

From the notch/dot at one end of the package, viewing from the top, the pins count counter-clockwise (reverse of what you show in your video). On the bottom side, they count clockwise.

On some boards (including this one), pin 1 of each chip has a square pad and the rest have round or oval pads.

On your pic of the bottom, it looks like there is a broken trace beside pin 11, and it looks like the wide trace connected to pin 16 may not reach the pad any more.

On your pic of the top, the trace that runs through the gap between pins 7 and 8 looks suspect, there is a potential solder bridge from pin 10, and it looks like the wide trace connected to pin 16 may not reach the pad any more (might just be lighting).

On an earlier pic of the newly installed chip, the solder on the bottom of the board looked ok, but from the top of the board, the holes did not contain solder, indicating there was not enough heat for the solder to wick through the board. Some of the solder joints in the PSU repair pics also look like there might not be enough heat applied for the solder to flow properly; reviewing a DIP soldering tutorial might be helpful (or I might be misinterpreting the pictures... it is hard for me to discern what shape a shiny 3d surface is in a 2d picture).

Due to the difficulty of desoldering on these heavy copper 4-layer boards, you want to minimize the amount of re-work, as the damage to traces and pads and delamination will accumulate. This is one of the reasons for using a socket (when there is room to do so). In this case there is also the concern that maybe the problem is not isolated to that one part, maybe the replacements are bad or mis-marked, etc. so being able to change the part easily is valuable. Sockets can cause intermittent problems, so you want to use a high quality one.

In addition to checking for continuity from a pin of the new ram chip to the same pin on another chip, it can also be helpful to check that adjacent pins at the replaced chip are not shorted together due to an unseen solder bridge.

Using the Lisa Memory Board Schematic, you can determine where all of the pins of the replacement chip are supposed to be connected (not just to the adjacent memory chip, but elsewhere on the board) and so you can verify them all.

[berskyboy]

On your pic of the bottom, it looks like there is a broken trace beside pin 11, and it looks like the wide trace connected to pin 16 may not reach the pad any more.

On your pic of the top, the trace that runs through the gap between pins 7 and 8 looks suspect, there is a potential solder bridge from pin 10, and it looks like the wide trace connected to pin 16 may not reach the pad any more (might just be lighting).

hi Sigma7,

I decided to try again with another IC socket, but before I do that, I think I have issue with my lifted broken traces.  I see from the photo problem areas.  (Pins 9, 10,11)

https://www.dropbox.com/scl/fi/z660yxfq5twpkzsxud2t8/IMG_4233.jpeg?rlkey=ldghw6pws4zuj3zo0uaj7oz56&st=ag7pc1xx&dl=0

https://www.dropbox.com/scl/fi/5uyvbthtz8ecpktuo49kd/IMG_4234.jpeg?rlkey=ahddlbbqhci6yea3zrsj1loc7&st=eu8vc57z&dl=0

and that might be the issue previous to my. first attempt. 

Question is it salvageable?  I think I would need to put jumper wires 9, 10 and 11. ( they seem to be just connected to the adjacent chip) so D7 in this case so I could run those wires under the board.  Is there any other problem areas?

YouTube Video (Top of the board)
https://www.youtube.com/shorts/FUjApAEzW-w

YouTube Video (Bottom of the board)
https://www.youtube.com/shorts/YZtorRMpPoE

I'm looking for the  Lisa Memory Board Schematic,
are they these ones?

The Vintage Compu85 Mirror: You can access the direct circuit GIF images via the [Compu85 Lisa Schematics Index](https://compu85.homeip.net/stuff/Apple/lisa/Schematics/ "Apple Lisa "Memory Board Schematic" pdf"). On this server, look for the files specifically named ram512-1.gif, ram512-2.gif, and ram512-3.gif. These three files contain the complete multi-page logic layout for the 512K memory expansion board.

[sigma7]

Question is it salvageable?

Yes.

I think I would need to put jumper wires

Yes.

Note that some traces are daisy-chained across the row of chips, so you need to make sure that those traces connect to the chips on both sides. eg. if the trace is broken, re-connecting to the chip on one side with a jumper may still leave the chip on the other side disconnected.

Since/if you have another working memory board, comparing continuity tests with it may be helpful.

I'm looking for the Lisa Memory Board Schematic

You can use those, or you can find originals on bitsavers, eg: https://bitsavers.org/pdf/apple/lisa/hardware/Lisa_Memory/050-4010-B_Mem.pdf

AFAIK, the bitsavers' Lisa Hardware directory has all of the Lisa schematics now. ( Thanks Al !)