Macintosh SE Sad Mac: Bad RAM chip identification possible?

[JDW]

I've been testing my inventory of 1MB RAM SIMMs using my Mac SE Reloaded machine, and with one particular set of SIMMs, I get a Sad Mac at cold boot showing these 2 lines:

00000004 <— "Bank A"
00004000 <— Failed bits...

AABBCCDD​

​

AA = bits 31-24
BB = bits 23-16
CC = bits 15-8
DD = bits 7-0​

Failed SIMM (8 Chips marked SANYO LC321000J-10):





There's another SIMM just like it which tests fine when used with my known good RAM. So it's not a matter of that particular RAM type not working in an SE.

QUESTION: As per the above, can it be determined which IC or ICs are in need of replacement?

 

[Kai Robinson]

30-pin SIMM's are 8-bit wide, so, the problem is in one of the right most SIMMs on the board, not the leftmost. so - Slots 2 & 4.

Left SIMM is D0 to D7, Right SIMM is D8 to D15.

On the SE, there is only one bank, Bank A.

Each SIMM uses 1-bit wide chips - from left to right, so in a single row you can see bits 0 to 15.

 

[JDW]

30-pin SIMM's are 8-bit wide, so, the problem is in one of the right most SIMMs on the board, not the leftmost. so - Slots 2 & 4.

Left SIMM is D0 to D7, Right SIMM is D8 to D15.

On the SE, there is only one bank, Bank A.

Each SIMM uses 1-bit wide chips - from left to right, so in a single row you can see bits 0 to 15.

Kai, thank you for your reply.

I gleaned the Rear = Bank A, Front = Bank B info from this Low End Mac article. (6th paragraph from the top)

I know for a fact which SIMM is bad because I put it into different slots with 3 other known-good SIMMs and the Sad Mac codes changed in accordance to the slot location.

So basically, I asking if there is a way to know which IC (or ICs, if more than one) on that bad SIMM is/are bad.

Why do I ask this question? Because although it's easy to fix a bad SIMM by swapping out all the chips, it's faster, easier and potentially less costly to swap out 1 or 2 bad chips, while leaving the other good chips as they are.

 

[Kai Robinson]

So, if you re-read what i was saying - the faulty one will be in SIMM SLOT 2. You could get a SIMCHECK unit, or something similar - however, the only way you can really test without one, is 'in-situ' using an oscilloscope. If you probe the data and address lines you should be able to see which line one is acting weird, but that only narrows down WHAT is going wrong. You'd need to test each chip on the SIMM individually to really test it further.

I did come across this though: https://hackaday.com/2014/04/09/using-simms-to-add-some-extra-ram-on-your-arduino-uno/

That might be a basis for constructing a tester. Alternatively, I'd lend out my Innoventions SIMCHECK to someone like @Zane Kaminski if there was interest in producing a replica. Innoventions recently folded, so there's no longer any copyright issues surrounding cloning the unit.

 

[retr01]

@JDW, I uploaded a copy of Apple's TA (Technical Article) that provides more about SAD Mac codes. It's in the resources section here on TD.

Now I want to find a paper copy from Apple that lists even more error codes ranging from SAD Mac to System to other codes. That is one of my hunts.

 

[YMK]

So basically, I asking if there is a way to know which IC (or ICs, if more than one) on that bad SIMM is/are bad.

You should have continuity between D6 on the SIMM edge contacts and the bad IC.

 

[Crutch]

Now I want to find a paper copy from Apple that lists even more error codes ranging from SAD Mac to System to other codes. That is one of my hunts.

I am not aware of a comprehensive paper copy from Apple (the various editions of Inside Macintosh often listed error and result codes pertaining to functions in that volume only, e.g. Appendix A in Volume VI), however the best/most convenient single source I know of for error and return codes (not sad Macs) is the “Error/Return Codes” section in THINK Reference, which is available on the Garden.

 

[retr01]

I am sure Apple did have a paper copy or at least listed on one of the Mac Technicians Reference CDROMs. Paperless was unheard of back in the day.

Authorized Apple Dealers with a service center gets a lot of documentation from Apple, including reference sheets.

 

[JDW]

...the best/most convenient single source I know of for error and return codes (not sad Macs) is the “Error/Return Codes” section in THINK Reference, which is available on the Garden.

@Crutch , could you please provide an exact link to the "THINK Reference" you are referring to? The best I can find on Macintosh Garden is Symantec THINK C 6.0.1, but when I open "THINK Reference" on "Development.img" in Mini vMac, I only see this...

You should have continuity between D6 on the SIMM edge contacts and the bad IC.

Now THAT is the speedy testing tip I was hoping for! I'll give that a try tonight after work. Thank you very much!

 

[Crutch]

@JDW try this https://www.macintoshrepository.org/9613-think-reference-2-0

THINK C came with a very stripped down version of THINK Reference that only included documentation on the C standard library. The Toolbox documentation was a separate, paid product, linked above.

 

[JDW]

@JDW try this https://www.macintoshrepository.org/9613-think-reference-2-0

Thank you for the specific guidance! I appreciate it.
I've extracted the content you spoke of here...



It is basically the same content as what @retr01 posted here, and what is shown on this website. It was that same info I consulted when putting together my opening post.

Simply put, that info helped me know it was a RAM related problem, and ultimately I found the bad SIMM. There is absolutely no question in my mind about which of my SIMMs is bad.

Later tonight I will try the excellent advice of @YMK to see if I can isolate the problem to an exact chip. Once that is done, it would simply be a matter of removing the bad RAM chip(s) and replacing them, rather than trashing the SIMM or replacing all the chips.

 

[JDW]

You should have continuity between D6 on the SIMM edge contacts and the bad IC.

@YMK
To understand how to test "D6" I was looking for a schematic of 8-chip 30-pin SIMMs. I only found this one, which doesn't have 8 chips...

SOURCE: https://alexandrugroza.ro/microelectronics/system-design/4mx9-simm/index.html

The only useful thing the above schematic tells me is that pin-23 on the SIMM is D6, which in the case of my SIMM is here...

So if I take my DMM set to Continuity (beeper) mode, I need to check between pin-23 on the SIMM and what pin on each IC ?

If I test between SIMM pad-23 and the actual D6 pin on each IC, then of course it should beep, so I assume you mean test between pad-23 and a different IC pin?

(Again, we are doing this test to determine which IC or ICs are bad.)

 

[YMK]

The chips on your diagram are four bits wide each. Your SIMM has eight chips with a single data line each:

Datasheet Archive: LC321000-10 datasheet by Sanyo Semiconductor

Preview and download the LC321000-10 datasheet by Sanyo Semiconductor on Datasheet Archive.

www.datasheetarchive.com

The data pins on each IC are 1 and 17, and should be tied together.

 

[JDW]

@YMK
I apologize. I was busy and hadn't looked at the LC321000-10 datasheet (SMD chip version) as I should have. But extracting the relevant chip diagram from the correct SMD version datasheet shows this...

Din = pin 1
Dout = pin 25

So the 2 pieces of advice you gave me were:

1. You should have continuity between D6 on the SIMM edge contacts and the bad IC.
2. The data pins on each IC are 1 and [25], and should be tied together.

I am trying to harmonize that advice, so please forgive me but I wish to clarify. Do you mean I need to confirm that Pins 1 & 25 are connected together on each of the 8 chips, AND THEN check to see if pad-23 (D6) on the SIMM (as per the schematic in my earlier post) is also connected to Pins 1 & 25 (in order to confirm that one of the 8 chips is bad)?

 

[YMK]

Sorry, I was looking at the wrong package, so yes, 25 is the right one.

Continuity between D6 and either pins 1 or 25 will identify the IC.

Continuity between D6 and both pins 1 and 25 will additionally rule out a bad SIMM PCB (since 1 and 25 need to be tied together through the PCB).

 

[retr01]

This is amazing. My brain is churning more. The basics help me to learn more about the software engineering side.

 

[JDW]

Sorry, I was looking at the wrong package, so yes, 25 is the right one.

Continuity between D6 and either pins 1 or 25 will identify the IC.

Continuity between D6 and both pins 1 and 25 will additionally rule out a bad SIMM PCB (since 1 and 25 need to be tied together through the PCB).

I confirmed that pins 1 & 25 are connected to each other on all 8 chips.

I then tested between SIMM pad 23 and pin 1 of each chip and found that only the second IC from the right, as shown below, is connected to pad 23.

 

[YMK]

The second IC is the likely culprit then.

If you remove that chip from the SIMM, put it back in the SE and get the same error code, you can be fairly certain that you've found the problem.

 

[JDW]

The second IC is the likely culprit then.

If you remove that chip from the SIMM, put it back in the SE and get the same error code, you can be fairly certain that you've found the problem.

Thank you very much! I will do that test when I have time, but for now, I humbly wish to thank you for educating me on the proper way to diagnose 30-pin RAM SIMMs down to the chip level. Amazing!

 

[retr01]

Hey @JDW, do you think you could replace that particular IC rather than the whole SIMM? I wonder which is more cost-effective - the failed IC or the entire SIMM to replace?