SE/30 RAM Banking QUESTION

[JDW]

According to this LEM page, unless you are installing 17MB, you should be putting the larger sized SIMMs into Bank A, with the smaller sized SIMMs into Bank B.



However, when I put four 16MB SIMMs into Bank A, and nothing in Bank B, I get death chimes and a crazy display. But if then put 4 sticks of 4MB SIMMs into Bank A and my 4pcs of 16MB SIMMs into Bank B, then the Mac boots just fine and shows the correct 82MB of RAM in the Finder.

No doubt at least one of you have tried this with 16MB SIMMs before, so what are your thoughts? Is this another except to the rule like when installed 17MB of RAM? Please read the LEM article for more info.)

 

[JDW]

Should we compare and see?

NEW sockets are CREAM colored...



My old socket was black, and it also had metal tabs. Even so, only the new cream colored socket works with the ROM-inator II...





Lastly, please read the following excellent post by @Kay K.M.Mods regarding motherboard warping...

https://tinkerdifferent.com/threads/smc-rom-simm-prototype.2198/#post-19023

 

[retr01]

For me, the old cream color with plastic clips for ROM SIMM and black RAM SIMMs with plastic clips on my SE/30 logic board, then replaced with new ones with metal tabs:

Thank you, @JDW. I am reading @Kay K.M.Mods' post about the warped board endemic and the new SMC ROM SIMM that I am already drooling over.

 

[JDW]

Not sure where Kay sourced the ROM socket he installed in my SE/30 motherboard, but I purchased the RAM sockets from RS Components Japan and had them directly shipped to Kay. In the first batch, which was foolishly shipped in a mere bubble-wrap envelope, 4 of the sockets were chipped, but we deemed them okay at the time. Sadly, later when Kay shipped the board to me, I was installing RAM and SNAP! One edge broke clean off!


All came from Japan inventory, so Kay thought the weakened plastic was induced by high humidity. I complained to RS and they then dispatched free replacements from inventory from the UK, and those sockets were good. But those replacements also came in a stupidly thin bubble wrap envelope offering next to no protection...


Well, I hope the sockets prove to be good over time. Not sure when they were made though or what plastic was used.

I believe Amiga of Rochester sourced his RAM sockets from a US based supplier because the shape is a tad different. Comparing your photos to mine, I see the ROM sockets used are also different.

 

[retr01]

We will see how the 3D-printed clips work out for me, then.

 

[croissantking]

Not sure where Kay sourced the ROM socket he installed in my SE/30 motherboard, but I purchased the RAM sockets from RS Components Japan and had them directly shipped to Kay. In the first batch, which was foolishly shipped in a mere bubble-wrap envelope, 4 of the sockets were chipped, but we deemed them okay at the time. Sadly, later when Kay shipped the board to me, I was installing RAM and SNAP! One edge broke clean off!
View attachment 10848

All came from Japan inventory, so Kay thought the weakened plastic was induced by high humidity. I complained to RS and they then dispatched free replacements from inventory from the UK, and those sockets were good. But those replacements also came in a stupidly thin bubble wrap envelope offering next to no protection...

I'ved use exactly the same type of NOS sockets for my Bolle SE/30 builds that I got from a UK supplier, and they are just as shockingly brittle, more fragile than the sockets on my original boards.

One of my new build boards isn't finished yet and I've only soldered in four of the RAM sockets. I'm able to run 4x 16MB SIMMs without any trouble.

 

[JDW]

I'ved use exactly the same type of NOS sockets for my Bolle SE/30 builds that I got from a UK supplier, and they are just as shockingly brittle, more fragile than the sockets on my original boards.

It's a travesty, to be honest. Even if we argue that the sockets which broke in my case broke due to age and humidity, the fact remains that many vintage sockets withstand even longer spans of time, and all the humidity during those years, better than these "new" sockets do. These are RAM sockets, after all. They have metal tabs on them because we will swap out RAM rather often. But if the underlying plastic is brittle, the metal tabs have no meaning, and the manufacturer of the sockets is basically putting out a product that isn't worth selling. And yet, what other choices do we have? It's all very troubling, to be sure.

... I've only soldered in four of the RAM sockets. I'm able to run 4x 16MB SIMMs without any trouble.

I assuming that means the four sockets of Bank A, which means you were somehow able to get only a set of four 16MB SIMMs to work in your SE/30, which is interesting because I am unable to do that. As mentioned in my earlier post, when I put four 16MB SIMMs into Bank A, and nothing in Bank B, I get death chimes and a crazy display. But if then put 4 sticks of 4MB SIMMs into Bank A and put my 4 sticks of 16MB SIMMs into Bank B, then the Mac boots just fine and shows the correct 82MB of RAM in the Finder. I'm still at a loss to explain this, especially now that you say you can use 4 sticks of 16MB SIMMs in Bank A just fine. I'm baffled!

 

[croissantking]

And yet, what other choices do we have? It's all very troubling, to be sure.

Exactly, those were the only ones I could get hold of at the time (and they weren’t exactly cheap). I had one module that wasn’t lining up properly and I pushed just a bit too hard and the whole corner came off the socket.

I did get a free replacement from the seller on a one-off basis as he was reluctant to admit fault. I am just being very, very delicate when installing and removing SIMMs but I am quite nervous about using them tbh. I noticed another socket has chipped a bit just through normal handling while I test and troubleshoot my build.

I assuming that means the four sockets of Bank A, which means you were somehow able to get only a set of four 16MB SIMMs to work in your SE/30, which is interesting because I am unable to do that.

Yep, the four sockets of bank A. I am going to try this again just to be sure, because now you have me questioning my memory!

 

[JDW]

@Croissanting
Seeing your photo upsets me greatly because no expensive sockets like that should break so easily! And these are expensive! So what I am forced to do is use both hands and press outward on the metal tabs of the socket while I carefully and slowing insert a SIMM and pull it into an upright position, then let the metal tabs move inward and latch onto it.

Perhaps it can be said it isn't the fault of the sellers of these sockets, but it absolutely is the fault of the manufacturers. Even if one argues these are "new old stock" (who knows when they were originally made!), the plastic used (and/or the design) is flawed in being so brittle and sensitive to breakage from only "normal usage." It's outrageous.

Maybe other people who have build Bolle's SE/30 board or who have swapped out their RAM sockets could chime in too with their experiences.

Thanks for re-testing the 4 sticks of 16MB in Bank A (the bank nearest the edge of the board). It really baffles me what it doesn't work in my SE/30.

 

[retr01]

Maybe @Kay K.M.Mods can look into this and make more robust plastic housings of SIMM sockets with 3D printing. Then, how do we move the metal parts from the "cheap" and "lousy" SIMM sockets to the better ones?

It was the gears for the floppy drives. Now it's the SIMM sockets that need help.

 

[croissantking]

Maybe @Kay K.M.Mods can look into this and make more robust plastic housings of SIMM sockets with 3D printing. Then, how do we move the metal parts from the "cheap" and "lousy" SIMM sockets to the better ones?

It was the gears for the floppy drives. Now it's the SIMM sockets that need help.

I think those metal pins pull out fairly cleanly, although there is probably a large variation of shapes out there.

Perhaps someone can make up some new ones? It seems like there will be enough demand.

 

[retr01]

Yes. I agree with you, @Croissanting, that perhaps the demand is there because of manufacturing quality problems of the SIMM sockets through various suppliers. And maybe humidity?

 

[croissantking]

Thanks for re-testing the 4 sticks of 16MB in Bank A (the bank nearest the edge of the board). It really baffles me what it doesn't work in my SE/30.

Sorry for taking so long to follow up on my findings on this. I can now clarify what’s going on here. Basically, I can have just 4 sticks on 16MB installed when I’m using my Rominator II (which is basically a IIsi ROM with some extra features). With the stock ROM, it won’t boot: I need to have those other four sockets populated with something, though not necessarily 16MB SIMMs.

I assume you’re using the stock ROM, in which case I think this clears up the confusion.

 

[JDW]

Sorry for taking so long to follow up on my findings on this. I can now clarify what’s going on here. Basically, I can have just 4 sticks on 16MB installed when I’m using my Rominator II (which is basically a IIsi ROM with some extra features). With the stock ROM, it won’t boot: I need to have those other four sockets populated with something, though not necessarily 16MB SIMMs.

I assume you’re using the stock ROM, in which case I think this clears up the confusion.

Yes, I am using a stock SE/30 ROM because most people use a stock ROM rather than the BMOW version. I have a ROM-inator II MEGA (8MB edition), but I didn't test my RAM configs with that, for the reason that most people probably wouldn't have that ROM. But it would be an interesting test to see if a stock IIsi ROM (not the modified version used on the ROM-inator, would have the same issues as the stock SE/30 ROM.

I'm big on non-stock ROMs though. I have the IIsi, IIfx, ROM-inator II Mega, 2MB Kay Koba ROM, and 2MB CayMac vintage ROM, as well as ROM SIMM Programmers from BMOW (no longer sold) and CayMac (currently sold). Indeed, I am doing the prep work for a video on the CayMac programmer now, which will have a discussion on the ROM SIMMs it can program.

 

[trag]

I do NOT advise the solder across the ROM SIMM pad suggestion at all because I tried it and had trouble. Why? Because solder flakes apart over time, and the flakes short out adjacent pins! No damage occurred on mine, thankfully, but it was pretty scary discovering that.

As to why my Kay Koba ROM socket worked with the ROM-inator when your new socket did not, I can only guess it was either a difference in the socket model number, or maybe Kay installed it differently than Amiga of Rochester? Not sure.

I wonder whether one of these copper brush electroplating kits might be a reasonable solution.

They also have gold, but the price goes up tenfold...

Nickel as well, which should be more durable and for such thin layers the difference in conductivity probably doesn't matter.

Plug N' Plate Acid Copper Kit

caswellplating.com

 

[JDW]

I wonder whether one of these copper brush electroplating kits might be a reasonable solution.

I believe you are talking about the solution to bad ROM SIMM connections, not RAM Banking. The RAM Banking problem exists UNLESS you use a non-stock ROM, and then it magically vanishes.

Bad ROM SIMM connectivity is caused by most non-stock ROMs being only a nominal 1.20mm thickness, which can vary by +/- 10%! I have a stock SE/30 ROM that measures only 1.24mm across the pads and it won't allow one of my SE/30 motherboards to boot. I have another stock SE/30 ROM that measures 1.29mm at the pads, and it works perfectly.

Kay Koba found that most SE/30 motherboards are slightly warped in that, over time, they tend to bow slightly downward in the middle, due to gravity, humidity, heat and age. That can negatively impact the ROM SIMM socket. Even so, thick ROM SIMMs tend to overcome any troubles caused by motherboard warping.

I suspect another reason is that if you have a fat 1.30mm stock ROM sitting in the stock socket for decades and then suddenly put in a 1.20mm ROM SIMM, the pins probably were shaped in such a way they expect a fatter SIMM. In other words, those metal pins lost their ability to press hard on both sides of the SIMM pads, creating trouble for thinner SIMMs.

I spoke to GGLabs only a day ago and confirm their current SIMMs, black in color, are 1.27 to 1.30 measured at the pads. BMOW thickened their ATOM SIMM as well. They are the only two suppliers that offer such thick PCBs at this time.

ROM Clips by Potatofi present a fix in theory, but all my SE/30 boards require the Type-A ROM Clips, and for the life of me, I cannot make them fit on ROM SIMMs like that from Kay Koba, which have SMD capacitors at the left/right PCB edges. That's because the ROM Clips slide straight down onto the SIMM's left/right sides. The SMD caps get in the way. This indicates that ROM Clips are not a fix for every ROM.

The best solution is the hardest and most expensive. I did that on one board. I don't have a good Hakko Desoldering Station, so I paid Kay Koba a little money to swap out all my RAM and ROM sockets. I can put any ROM SIMM into that new ROM socket, which has metal tabs, and it works perfectly.

With that said, nothing beats gold-plated pads. Nothing. Only SiliconInsider offers that on their ROM SIMMs. But here's the caveat. Their ROM SIMMs are nominal 1.20mm, which means if the PCB isn't thick enough, the plating won't matter because the pins won't touch the pads well, if at all.

 

[trag]

I believe you are talking about the solution to bad ROM SIMM connections, not RAM Banking.

I should have stated it explicitly. I was suggesting electroplating as a method to thicken the connector region of the ROM SIMMs. Make the pins thicker. I don't know if electroplating is a practical way to add enough material to make the required difference....

If I ever find the top of my workbench, I'll probably just replace my sockets. I have plenty of the 64 pin SIMM sockets from "The Connector People" and I bought a bunch of dual-30-pin SIMM sockets with metal clips a long time ago, maybe from Jameco. Their spacing(socket to socket) is wrong, but they're not difficult to cut into singletons with a Dremel cutting wheel.

 

[MacOfAllTrades]

I just got 4x16MB simms for my se/30 last week. I put them into B bank IIRC. I left the 4x1MB SIMMs in Bank A which must have saved me because
I had no problems with the install and it worked fine -- Really glad I didn't get caught up in the situations well documented above that could have tripped me up!

Given the possible compatibility problems with non-32-bit-clean software when running in 32 bit addressing mode, I have chosen to just run in 24-bit mode anyway. Makes me feel a little silly for even buying the 64MB upgrade but it's fun to have it available. I mainly got it thinking I wanted to have plenty of cushion running Netscape Navigator with my blueSCSI v2 though 8 mb was plenty.

 

[JDW]

You really only need to run in 24-bit mode when using System 6. But for System 7, I always use 32-bit mode and never have issues.

Interesting that the RAM config you mentioned works. I can tell you that putting the reverse in also works (4pcs of 16MB in Bank A, then 4pcs of 1MB or larger in Bank B). It's only when you try to use 4pcs of 16MB in Bank A with nothing in Bank B that trouble begins, if you have the stock SE/30 ROM SIMM.

 

[croissantking]

You really only need to run in 24-bit mode when using System 6. But for System 7, I always use 32-bit mode and never have issues.

24-bit dirty software will crash if set to 32-bit mode, regardless of which OS you have booted. One example is Shufflepuck Café.

 

[JDW]

24-bit dirty software will crash if set to 32-bit mode, regardless of which OS you have booted. One example is Shufflepuck Café.

Very interesting. I've obviously not tried that app.

This raises the importance of have a comprehensive list of vintage software that works fine in 24-bit mode but which crashes in 32-bit. Would be interesting to see how many apps beyond Shufflepuck are affected.