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.)

 

[YMK]

The LEM article says the first 1MB is used for video, but the SE/30 has dedicated VRAM chips. Maybe he's thinking of the SE?

The last three pages of this PDF might explain what's happening:

https://www.tech-insider.org/mac/research/acrobat/9204.pdf

 

[JDW]

@YMK
You are correct about the error in that LEM article! Wow. I never noticed that. BTW, this discussion explains why adding more VRAM to an SE/30 won't do anything.

Thank you for the PDF. I read most of it, including the last 3 pages, but that closing text pertains mostly to the Macintosh II, and to a lesser extent the IIx. After reading it a couple times, it's still not clear how that discussion impacts the SE/30. In any case, it says nothing clearly about the use of 16MB SIMMs in the SE/30 other that to indirectly say that "SIMMs larger than 1MB... should be considered theoretical." I can understand that back in the day, but that was 30 years ago. Unfortunately, official documentation on the subject never went beyond the "theoretical stage."

Basically, when I wrote my opening post, what I was (and still am) hoping for was fellow SE/30 owners to chime in about their own 16MB SIMM experience. I certainly cannot be the long man with 16MB SIMMs (30 pin), and I'm curious if others have seen what I've seen, or if it is just a fluke of the particular motherboards I am using. I doubt it's a motherboard specific issue, but by asking the question to others, perhaps I can learn if others are seeing the same issue.

To repeat, if I put four 16MB SIMMs into Bank A and four 4MB SIMMs into Bank B, I get death chimes. Swapping Banks works, however. That is odd.

I should add that filling all 8 banks in an SE/30 with 16MB SIMMs yields 128MB and a perfectly booting machine. It's just that when you only want to fill a single bank with 16MB SIMMs and the second bank with smaller sized SIMMS (in my case 4MB), the way I am forced to bank the SIMMs goes against the general rule of putting larger RAM SIMMs into Bank A. I'd like to hear from fellow SE/30 owners who also have 16MB SIMMs to see if they are finding the same thing I am.

 

[YMK]

Which ROM does this happen with?

 

[JDW]

Which ROM does this happen with?

Stock SE/30 ROM. Nothing unique about the ROM.

 

[This Does Not Compute]

There's also an interesting note on page 7 of that PDF:

Due to an undocumented feature in the ROM firmware shipped with the original Macintosh II, a
Macintosh II with original ROMs is limited to using SIMMs no larger than 1 MB in Bank A. Large
SIMMs can only be put in Bank B (that is, 4 and 16 MB SIMMs). Remember that if Bank B is to
be used at all, Bank A must be populated first.

According to this Tech Note, the SE/30 is based on the Mac IIx and uses its ROM. While the Mac II's SIMM size limitations were fixed in the IIx's ROM, I wonder if the SIMM installation order still applies -- that is, smaller SIMMs in Bank A and larger ones in Bank B. This would certainly explain the behavior you're seeing, @JDW.

 

[JDW]

According to this Tech Note, the SE/30 is based on the Mac IIx and uses its ROM. While the Mac II's SIMM size limitations were fixed in the IIx's ROM, I wonder if the SIMM installation order still applies -- that is, smaller SIMMs in Bank A and larger ones in Bank B. This would certainly explain the behavior you're seeing, @JDW.

Thank you for making time to share your thoughts.

Through the years, I've not read any stories about the SE/30 having issues with 16MB SIMMs, but I will admit that so few people seem to have them that the data pool is rather small. Below is a graphic I extracted from an old book on Mac upgrades, and interestingly, it only shows 16MB SIMMs being used in all 8 banks; however, that could have been because they simply lacked space to show a 64MB or 80MB configuration.

 

[Kai Robinson]

I wonder if you'd have more success using PAL SIMM's for this - that was quirk on the IIx's

 

[JDW]

I wonder if you'd have more success using PAL SIMM's for this - that was quirk on the IIx's

Thanks, Kai. Well, I don't have that kind to test. But if anyone should know if a PAL SIMM can work at all in an SE/30 it would be @Zane Kaminski, seeing that Garrett's Workshop created them!

GW4194A -- 4x4MB "PAL SIMM" RAM for Mac II and IIx by Garrett's Workshop on Tindie

4x4MB "PAL SIMM" 16 MB RAM kit for Mac II and IIx

www.tindie.com

However theirs is a 4MB type, not 16MB, so even if they confirm their 4MB PAL SIMMs do work in an SE/30, that test wouldn't confirm if a 16MB sized PAL SIMM would work. I do know that regular 4MB SIMMs (30 pin, not PAL) work fine with 4pcs in Bank A alone.

 

[MacKilRoy]

You do realize that when using 16MB SIMMs in the SE/30, you have to populate all 8 slots, right?

If you only have 4x 16MB SIMMs, populate the other 4 slots with any 4 matching SIMMs. It's an SE/30 quirk.

 

[JDW]

You do realize that when using 16MB SIMMs in the SE/30, you have to populate all 8 slots, right?

If you only have 4x 16MB SIMMs, populate the other 4 slots with any 4 matching SIMMs. It's an SE/30 quirk.

No, I absolutely do NOT realize that, as per my previous posts on the matter!

You seem to suggest in your second sentence that only 8pcs of 16MB SIMMs can be used in the SE/30. Not sure about the meaning of "If you only have 4x 16MB SIMMs" because in the second half of that second sentence you then say "populate the other 4 slots with any 4 matching SIMMs." That seems to imply this rephrase: "If you only have 4x 16MB SIMMs, buy 4 more because the other 4 slots much have matching SIMMs." At least, that is the way I am led to interpret your second sentence. If I am wrong, please clarify.

You seem rather emphatic about my findings being a quirk, but can you point us to some online info about this being a quirk? Or are you simply speaking from your own testing experience which merely happens to match my experience? (Not a criticism, mind you. I'm just curious what info is out there about 16MB SIMMs in the SE/30 that I may have overlooked.)

So few people have 16MB SIMMs in their SE/30s, it's hard to find good testing info on the subject. Basically, i am curious if this affects all SE/30 motherboards or just some.

The issue though is this. If one strongly argues that all 8 slots of an SE/30 motherboard MUST have 16MB SIMMs — 8 pcs of 16MB SIMMs — then how does one explain my finding that 4pcs of 16MB SIMMs can be made to work fine without a second 4pc matching set, ONLY if the 16MB SIMMs are placed in Bank B, and then place your other RAM SIMMs (from 16MB and smaller) into Bank A? I haven't tested all compatible RAM sizes in Bank A while 4pcs of 16MB SIMMs are in Bank B, but I can tell you what my previous posts have said, that 4pcs of 4MB SIMMs in Bank A and 4pcs of 16MB SIMMs in Bank B works (shows 82MB in the Finder).

Thanks!

 

[MacKilRoy]

No, I absolutely do NOT realize that, as per my previous posts on the matter!

You seem to suggest in your second sentence that only 8pcs of 16MB SIMMs can be used in the SE/30. Not sure about the meaning of "If you only have 4x 16MB SIMMs" because in the second half of that second sentence you then say "populate the other 4 slots with any 4 matching SIMMs." That seems to imply this rephrase: "If you only have 4x 16MB SIMMs, buy 4 more because the other 4 slots much have matching SIMMs." At least, that is the way I am led to interpret your second sentence. If I am wrong, please clarify.

You seem rather emphatic about my findings being a quirk, but can you point us to some online info about this being a quirk? Or are you simply speaking from your own testing experience which merely happens to match my experience? (Not a criticism, mind you. I'm just curious what info is out there about 16MB SIMMs in the SE/30 that I may have overlooked.)

So few people have 16MB SIMMs in their SE/30s, it's hard to find good testing info on the subject. Basically, i am curious if this affects all SE/30 motherboards or just some.

The issue though is this. If one strongly argues that all 8 slots of an SE/30 motherboard MUST have 16MB SIMMs — 8 pcs of 16MB SIMMs — then how does one explain my finding that 4pcs of 16MB SIMMs can be made to work fine without a second 4pc matching set, ONLY if the 16MB SIMMs are placed in Bank B, and then place your other RAM SIMMs (from 16MB and smaller) into Bank A? I haven't tested all compatible RAM sizes in Bank A while 4pcs of 16MB SIMMs are in Bank B, but I can tell you what my previous posts have said, that 4pcs of 4MB SIMMs in Bank A and 4pcs of 16MB SIMMs in Bank B works (shows 82MB in the Finder).

Thanks!

If using only 4x 16MB SIMMs, install any capacity SIMMs in bank A and your 4x 16MB SIMMs in Bank B.

4x 256Kb in A, giving 65MB, or 4x 1MB for 68MB or 4x 4MB for 80MB

 

[JDW]

If using only 4x 16MB SIMMs, install any capacity SIMMs in bank A and your 4x 16MB SIMMs in Bank B.

4x 256Kb in A, giving 65MB, or 4x 1MB for 68MB or 4x 4MB for 80MB

Yes, that parallels the finding I put forth in my opening post. However, as I asked you in my previous post, does that information come from your own personal experience, or did you read about it somewhere? If you did read about it somewhere, could you please provide a link to that information source?

The reason I am persistent in asking about the “information source,“ is because I had a hard time finding such information myself. So I am quite eager to learn about your sources of info on the subject.

Thank you very much!

 

[MacKilRoy]

Yes, that parallels the finding I put forth in my opening post. However, as I asked you in my previous post, does that information come from your own personal experience, or did you read about it somewhere? If you did read about it somewhere, could you please provide a link to that information source?

The reason I am persistent in asking about the “information source,“ is because I had a hard time finding such information myself. So I am quite eager to learn about your sources of info on the subject.

Thank you very much!

A Green Jade Upgrade: RAM Upgrades, Custom ROMs, and running Mac OS 8 on an SE/30

I recently posted about restoring an SE/30 – but the story didn’t end there. Since then, I’ve made some additional modifications including more RAM (more than double the Apple-sup…

metalbabble.wordpress.com

Lots of reading here:

My SE/30 hates 16mb ram sticks

I’m trying to upgrade my SE/30 with 16mb ram sims, but no matter what I try it refuses to boot as soon as I add them (jailbars and chime of death). I have two different sets of 16mb ram sticks (a total of 8). One set from Siliconinsider here on the forums, and a second set from Keystron which...

68kmla.org

It likely has to do with the memory being multiplexed in 12 bit chunks, as outlined on that 68kmla post.

All I know is that this issue has been known for many years and why I lead with “you do know…?” line. There are many blogs posts and pages with people having the exact issue of needing to put 16MB SIMMS into bank B if only 4 are being used.

 

[JDW]

A Green Jade Upgrade: RAM Upgrades, Custom ROMs, and running Mac OS 8 on an SE/30

I recently posted about restoring an SE/30 – but the story didn’t end there. Since then, I’ve made some additional modifications including more RAM (more than double the Apple-sup…

metalbabble.wordpress.com

Lots of reading here:

My SE/30 hates 16mb ram sticks

I’m trying to upgrade my SE/30 with 16mb ram sims, but no matter what I try it refuses to boot as soon as I add them (jailbars and chime of death). I have two different sets of 16mb ram sticks (a total of 8). One set from Siliconinsider here on the forums, and a second set from Keystron which...

68kmla.org

It likely has to do with the memory being multiplexed in 12 bit chunks, as outlined on that 68kmla post.

All I know is that this issue has been known for many years and why I lead with “you do know…?” line. There are many blogs posts and pages with people having the exact issue of needing to put 16MB SIMMS into bank B if only 4 are being used.

Thank you very much for the two links! I've not had time yet to read through the MLA thread, but I did read through the first URL you gave me, and the specific part you are referring to is this...

Unfortunately, he links to this Low End Mac page, which is otherwise superb info except for the error pertaining to "video memory." While "the first 1MB for video memory" applies to the Mac SE because the SE does not have dedicated video memory and therefore must use RAM instead, that statement does NOT apply to the SE/30 which has dedicated video memory and therefore should not be using RAM for video display. I actually posted about this error on the LEM FaceBook group some days ago, tagging Daniel Knight and the new owner of the LEM website, but neither of them have replied back yet.

As to his use of rubber bands to secure to a ROM-inator II, that works only until they ultimately break, which means you really do need to use the DIY metal bar gizmo for any long-term fix for that PCB thinness problem. If only it was slightly thicker, none of that would be needed.

I will read through that MLA link when I have more time. For now, I will say that I did not notice it there because I stopped all posting and reading on the 68kMLA in July 2021 due to a variety of things which ultimately led us to create TinkerDifferent. I see that the date stamp on the opening post is Dec. 2021, which explains why I never knew about it. And I guess it isn't indexed well because it never appeared in my Googling. Even so, I humbly appreciate your having pointed it out for me because I would not have found it otherwise. Again, I will go through that a bit later.

Many thanks!

P.S. Here's a nice chart showing RAM prices back in the day. Yowza!

 

[retr01]

As to his use of rubber bands to secure to a ROM-inator II, that works only until they ultimately break, which means you really do need to use the DIY metal bar gizmo for any long-term fix for that PCB thinness problem. If only it was slightly thicker, none of that would be needed.

Yo @JDW!

@PotatoFi designed 3D printed clips to keep the ROM SIMM down in the SE/30.

I plan to print that. I need to build my 3D printer and do that for my SE/30. I have it maxed out to 128 MB, as I discussed my woes in this post several months ago, getting the SE/30 to use all of those 16 MB RAM SIMMs in both banks A and B.

 

[JDW]

I paid Kay Koba to replace all the RAM and ROM sockets in one of my motherboards. All have metal clips. I can now use a ROM-inator II SIMM without issue, whereas before I could not. Part of the issue is most SE/30 motherboards are slightly warped, bulging ever-so-slightly downward due to gravity. The SIMM warp the same amount. So while warping isn't the only issue with the ROM-inator II SIMM, it's a big contributor to it. I never considered that until Kay Koba enlightening me! So the 3D printed clips are a "fix" for the warped motherboard, we can say.

 

[retr01]

I paid Kay Koba to replace all the RAM and ROM sockets in one of my motherboards. All have metal clips. I can now use a ROM-inator II SIMM without issue, whereas before I could not.

I did the same thing - I had Tom at Amiga of Rochester replace my SE/30 logic board plastic clip SIMMs with new SIMMs with metal clips for ROM and RAM.

Yet, the Rominator II SIMM still didn't work. Steve at BMOW told me to do one of those things to seat the SIMM in more. It's a brand new SIMM slot which baffled me.

Then, Drake suggested running solder across on ROM SIMM and maybe the RAM SIMM contacts. I admit I am still not confident doing that myself.

Part of the issue is most SE/30 motherboards are slightly warped, bulging ever-so-slightly downward due to gravity. The SIMM warp the same amount. So while warping isn't the only issue with the ROM-inator II SIMM, it's a big contributor to it. I never considered that until Kay Koba enlightening me!

Aha! That might be due to the slight imperfections of the rail grooves in the metal chassis when sliding and swinging in/out of the logic boards. Manufacturing quality control wasn't that great in the 1980s, which may have contributed to issues like this, especially if part of the rail grooves had rusted over the years.

Amazing! Thank you, @Kay K.M.Mods and @JDW, for helping enlighten me!

So the 3D printed clips are a "fix" for the warped motherboard, we can say.

Sweet. I will find out for myself in a few weeks. Hopefully, I will get the 3D printer to build it and use it.

 

[JDW]

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.

 

[retr01]

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.

Hmmm, good to know. Whew!

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.

Hmmm. Should we compare and see?