• Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.

SE/30 doesn't like new 4MB SIMMs: Any thoughts?

Stephen_Usher

Well-known member
I've just bought a set of four 4MB SIMMs from eBay and I'm getting the following symptoms:
  • All four SIMMs in either RAM bank (if in bank B then with known working SIMMs in bank A): Fails POST: corrupt display.
  • Replace one known good 1MB SIMM with any one of the new 4MB ones: Happy chime.
  • Replace more than one known good SIMM with a new SIMM: fails POST: corrupt display.
Seeing as the second test works with any of the new SIMMs suggests that the SIMMs are operational, at least the lower 1MB is anyway, I'm wondering what the problem might be.

Any thoughts?
 

Phipli

Well-known member
Replace one known good 1MB SIMM with any one of the new 4MB ones: Happy chime.
They must be in groups of 4 SIMMs.
Replace more than one known good SIMM with a new SIMM: fails POST: corrupt display.
Same with this. You have to use sets of 4, in each bank.

Option 2 is a false positive, the Mac is confused and thinks you have 4x 1MB SIMMs or something.

So I'd start by spraying the RAM slots with contact cleaner. The 4MB SIMMs use more of the contacts. If you're lucky, one of the additional pins is dirty. If you're unlucky, the RAM isn't compatible for some reason, or the bigger SIMMs have exposed an existing fault on your bord.

People tend to have better luck with RAM SIMMs with 8 chips, if you have SIMMs with only two chips, it might be worth borrowing some SIMMs with 8 chips to see if they are recognised. If they are, return your SIMMs and get some others.
 

Stephen_Usher

Well-known member
Indeed, the option 2 was merely a test to see if they worked at all, which they do (at least the lower 1MB), so it's not the contacts in the slots. I was surprised that it did actually POST to be honest.

Electrolube EML already used on the slots.

I have no other 4MB SIMMs to test with.

The new SIMMs are new and specifically designed (as marked on the PCB) to work with the SE/30.
 

Skate323k137

Well-known member
The official recommendation is the 4x 4m simms in one bank and 4x 1m in the other. I did this config recently, pictured below in the correct placement. Ignore the system reporting so much ram usage, I had to enable 32 bit addressing still.
 

Attachments

  • 20230414_231827.jpg
    20230414_231827.jpg
    3.4 MB · Views: 10
  • 20230414_234522.jpg
    20230414_234522.jpg
    1.4 MB · Views: 10

Stephen_Usher

Well-known member
OK. I decided to look at A9 (pin 19) on the SIMMs, just to see what it looked like on the 'scope, as this is the address line the 4MB uses and the 1MB ones don't. Fitted the SIMMs into bank B again, put the probe on the pin, switched on... And happy chime. *sigh*

It's having a laugh. It saw me get the 'scope ready didn't it? :)

Just wait. It'll stop working again when I turn the 'scope off.
 

Stephen_Usher

Well-known member
Well, I guessed that it was probably a connection issue, so reflowed the pads on the SIMMs, and the problem has gone away entirely. The PCB is probably slightly thinner than the original SIMMs and so the contact between the pad and the socket wasn't good enough.
 

Berenod

Well-known member
Well, I guessed that it was probably a connection issue, so reflowed the pads on the SIMMs, and the problem has gone away entirely. The PCB is probably slightly thinner than the original SIMMs and so the contact between the pad and the socket wasn't good enough.
I think it's a typical issue, the pcb's being just slightly thinner.
I've seen people stick small pieces of paper between the simm's and the retainer clips just to increase pressure on the pads/contacts.
 

Melkhior

Well-known member
The PCB is probably slightly thinner than the original SIMMs and so the contact between the pad and the socket wasn't good enough.
It's highly likely to be that if the SIMM are of modern manufacture.

The relevant JEDEC standard for 30-pins SIMM is "30 Pin Circuit Pluggable Single-In-Line Package (SIP) TABs on .100 inch Centers. Addition of Variation AE. Item 11.14-009.", MO-064-C. The thickness is between 0.047" and 0.053", nominal 0.050". That translates to 1.1938mm, 1.3462mm, and 1.2700mm in SI. This should include plating and/or metallization (note 4). Modern equivalent PCB are 1.200mm nominal, JLCPCB says "The thickness of finished board.", so presumably it includes "plating and/or metallization". The tolerance is 10%, so it's really between 1.0800mm and 1.3200mm. An over nominal JLCPCB 1.2mm board is in the low range of the JEDEC standard, but nominal is nearly the minimum and anything JLCPCB below nominal is outside of JEDEC specifications.

It's possible manufacturing the modern SIMM with thicker copper on the outside layer would help a bit, but that adds cost and may not solve the problem entirely.
 
Top