72-pin SIMM injection for updated SE/30 Schematic?

[Trash80toHP_Mini]

In the back when, I spent a lot of time and a bit of the filthy lucre on toys in the attempt to replace the two banks of memory in my IIsi with pairs of 32MB 72-pin SIMMs. I finally found and dug out the box playtoys last night.  Documented the more promising stuff today. We'll call this the tale of two SIMMexpanders, a remarkably incompatible SIMMPlus and a trio of true oddballs I've yet to see IRL that make the project ring true.



The 72-pin SIMM to 30-pin Socket converters appear to be fairly high grade unobtanium ore. Searching Minden or the Model Numbers didn't find them back in the day or today for that matter. :/

I hit a brick wall with what I thought was the most promising SIMMextender, the ME72472, /When it arrived, I spied a GAL on board and there went that notion  .  .  .  at that time. Bolle and Co. have come a long way in cracking GALs over the last few years, so this one may pose no problem.

I fell back on the ME72272, a pair of which would work just great  .  .  .  if only they supported 32MB SIMMs. But no, it does no support them, nor does the aforementioned SIMMPlus. For some reason the decoding on the pair supports only, 1 4, 16 and 64MB SIMMs. Dunno, maybe going for a pair of 64MB a a single 128MB SIMM makes more sense? I stayed wedded to the notion of using 32MB SIMMs because someone we know is sitting upon a dragon's hoard of the things.

At any rate, now it's a free for all, no set parameters.

The SIMMextender ME72472 is pure magic according to the specs on the packaging:



It uses single OR double sided modules! Is that the same as double banked, composite or whatever? Dunno, my brain's offline ATM, so you tell me.  ;-D







ME7227 in all its 32MB hostility:





Pesky SIMMplus:



Hey @trag you want I should send you some toys?

In the off chance someone thought I broke down with a case of sanity, I've got another design case in mind if nobody wants to make replacement boards for the pretty clear cases. It too would benefit from a massive reduction in PCB real estate from its highly customized SE/30 logic board. But that's for another story.  [}]





Portable MemCard is placeholder for tiny PDS card in Modem Slot, a NIC with RJ-45 pointing out the back would be perfect, room for another to the right..

 

[olePigeon]

I used to have the 72-pin to 30-pin expanders, but I gave them away (don't remember who.)

However, I have a bunch of the 30-pin to 30-pin expanders so you can use lower capacity SIMMs.  I also have a set of SIMM doublers for the IIfx that lets you use 2 SIMMs per slot.

 

[Franklinstein]

The problem with the SIMM expanders is that you're increasing the number of power-sucking chips by a quite a bit in addition to increasing electrical loads on the memory circuits. I wouldn't likely use them outside of a PC since they often have more power to spare and are more tolerant of weird memory configurations. Plus they're really tall so wouldn't fit well in a compact Mac or most of the II series or similar machines with limited RAM clearance. That said, I got a set of 72-pin SIMM-to-168-pin DIMM expanders in a box of random stuff once. Never tried to use them but they may work in old PCs that can use unbuffered EDO DRAM DIMMS.

Since some people are working on remanufacturing SE/30 boards anyway it shouldn't be too difficult to reshuffle some wires to accommodate 72-pin SIMM sockets in place of the 30s. Maybe flip the ROM SIMM 90 degrees and scooch the FPU over a little? Or maybe just use 4 72-pin sockets instead of 8 to keep things mostly in their original orientations? 4 sockets will still yield 128MB. 

 

[Trash80toHP_Mini]

The problem with the SIMM expanders is that you're increasing the number of power-sucking chips by a quite a bit in addition to increasing electrical loads on the memory circuits.

Yeah, there is that along with capacitance. I've started looking at using a single 128MB SIMM. The RAM ion that would be much more power efficient than a hectare of 30-pin SIMMs I'd think.

.  .  .  they're really tall so wouldn't fit well in a compact Mac or most of the II series or similar machines with limited RAM clearance.

Using the angled 72-pin socket I have, I don't see that as a problem. I'll have to snag the tallest SIMM I can get my hands on using a clear plexi logic board stand in. Better yet, what's the tallest 72-pin SIMM anyone happens to have on hand? Mocking that up on a cardboard or taping and extension to any single sided SIMM I've got would be a piece of cake.

Since some people are working on remanufacturing SE/30 boards anyway it shouldn't be too difficult to reshuffle some wires to accommodate 72-pin SIMM sockets in place of the 30s.

Yep, in Alex' schematic development thread, that was my motion, seconded by @trag

I've got a lot more flexibility in terms of height for RAM in my theoretical version of the SE/30 board. Decided this morning to put a pair or maybe a single vertical 72-pin connector in place of the SIMMexpander based riser in the pics above to further reduce PCB real estate requirements and capacitance issues. That also frees up the Memory Card position for a PDS slot. DIN-120 doesn't fit, so I'm switching to much shorter edgecard connector for vertical cards and the mini version of same for horizontally mounting PDS expansion devices.

 

[Trash80toHP_Mini]

Sorry, Alex, ran out of time to edit your name back into the title. Props for some really nice work! :approve: Macintosh SE/30 Schematics (modernization effort)

Sorry again, just found this popping up in the edit window. I'd meant to hit send as the first post after the IP. You've inspired all kinds of fun!  [)]

 

[Trash80toHP_Mini]

Back with pics of the glass bottom chassis.



Loads of cubic down there for a quartet of the 72-pin SIMMs I have on hand plus the thickness of the SIMMexpander PCB. I shoved it as far back as I could to where the top of the rear SIMM makes contact with the Plexi MoBo stand in.



At 100mm we'll call mine "standard height" SIMMs. The most rearward would probably be about even with the fourth 30-pin SIMM back to allow extra headspace required for the bayonet board mounting process. At the very front there's room for a SIMM at least 120mm tall and probably a second one given the thickness taken up by the PCB.

Then again, there's also that single angled socket in my bag of toys, just pulled it out for one last pic:



I'd dearly love to see a SIMM tall enough that it won't fit right up there in the front of an SE/30.

And now I've got visions dancing around in my head of of an inverted design mobo, its component side visible through the picture frame bottom of my clear case. . Thanks much, @maceffects []

 

[olePigeon]

The SIMMs in my IIci are about as tall as your SIMM expander.   I have a full stack (I call it "the brick") of 8 huge SIMMs.  I'm guessing they were some of the earliest 16MB modules and just used a crap ton of lower density chips, utilizing both sides of the board.

I love the extra big SIMMs.  It's like a hot rod. :lol:

 

[Trash80toHP_Mini]

HOT Rod is right! I wonder what kind of a chunk that many early tech DRAM ICs take out of the IIci's Power/Cooling budget? A pair of 72-pin SIMMs with later tech, higher density DRAM ICs is probably a lot more power efficient than a set of eight run of the mill 30-pin SIMMs?

 

[Franklinstein]

SIMMs 4-6 can probably be replaced on the SE/30's logic board with 72-pin sockets with minimal effort: delete SIMMs 1, 2, 7, and 8, then scooch some of the discrete chips between the RAM and the PDS around a little, either closer together or reposition in the empty space above/below the new RAM sockets.

Alternatively, perhaps rotate the new 72-pin sockets 90 degrees instead? I don't have measurements handy but that may work without reorganizing any of the other board components.

It appears the slanted sockets need about 50% more board space than the vertical type sockets though. If you're going to be spending a bunch of money printing a custom circuit board, you may as well buy newer low chip count/shorter SIMMs rather than try to squeeze in big old SIMMs. I know there were some 4-chip (TSOP style I think) 32MB 72-pin SIMMs but I think they were EDO, so as long as the SE/30's memory subsystem can use them (I know FPM and EDO features are harmlessly ignored on most Macs) they would be a good choice.

 

[Trash80toHP_Mini]

It appears the slanted sockets need about 50% more board space than the vertical type sockets though. If you're going to be spending a bunch of money printing a custom circuit board, you may as well buy newer low chip count/shorter SIMMs rather than try to squeeze in big old SIMMs. I know there were some 4-chip (TSOP style I think) 32MB 72-pin SIMMs but I think they were EDO, so as long as the SE/30's memory subsystem can use them (I know FPM and EDO features are harmlessly ignored on most Macs) they would be a good choice.

Good point, I play with the building blocks at hand.

I was under the impression that much of the interest in bringing the SE/30 design back to life was to condense a large part of its known logic into FPGA, CPLD or other magical little black boxes. In that scenario there's PCB real estate available to build a lot more function into the board, or stick a smaller board into curious spaces. [}]

Why not slather a full 128MB of TSOP or the like on the bottom/top of the PCB eschewing memory connectors of any kind. Why limit the dream to the thruhole state of mind? :huh:

 

[Franklinstein]

Careful not to get too far into feature creep. I think the motivator behind FPGAing some of the logic functions is because the original chips are NLA, which would allow people with battery-damaged boards to return their machine to life even with the loss of those original chips. My current take so far has been assuming original logic placed onto a new board. In this route the only real improvement available would be to add 72-pin SIMMs or possibly EDO SODIMMs (those could be had in 64MB quantities, so you'd only need two slots for 128MB). I don't know enough about the memory subsystem used here to say whether it would support high-density EDO chips or not. Previously these used 8 16-bit banks of basic or possibly FPM DRAM for its maximum of 128MB; I don't think EDO was introduced until the 72-pin SIMMs. Maybe with the higher-density memory it can use more than 128MB? I don't know what the hard limit is here. Not that I'd probably have a use for more than that in an SE/30. That being said it's probably less expensive to fix an SE/30 board by moving its components to a new board with four 32MB 72-pin SIMMs than it would be to get a replacement original board and find 128MB worth of 30-pin SIMMs.

 

[Trash80toHP_Mini]

128MB is a hard ceiling for all of the 030 Macs right up through the IIfx. Apple hobbled every Quadra to 136MB with with a useless bank of 8M on the system board excepting the 900 and 950, which had the full four acres of 32pin SIMMs on their boards.

Cost and availability of 16MB 30 pin SIMMs would be what got me onto the 72-pin adaptation several years ago. I'd imagine, newer, higher density RAM ICs for 128MB might be surface mounted above and below the two acres of thruhole 30-pin SIMM sockets? Dunno, @trag is our memory boffin, he could figure out if we may as well shoot right past the 72-pin SIMM upgrade option?

< TLDR TANGENT >

I guess it all depends on what people really want, as in terms of spending money on it? Purists may want the proprietary chips harvested from their dead board to populate a new design logic board that looks the same as the old in a sexy new color that you can't see anyway? I'd rather see some new possibilities opened up while we're at it. I'm not talking about featuritus as in add-ons, just about taking a more modular approach in the board design process. Think in terms of a layout blocked out for relatively easy rearrangement and adaptation to something like an SE/30 replacement board for the plastics of a dead Mac Portable, that's my notion, not Eleanor, but close.

The necessary parts and technology are in hand for a fully reversible, non-destructive SE30 Portable mod with a minimum of 640x400 on a perfectly fitting IPS panel.

One add-on though, a new format PDS card in this case, so to speak. Given @maceffects amazing developments based upon the Xceed 306-48, I don't see much difficulty in running the SE30 Portable's replacement panel at its native 1280x800 at a minimum of 16bit. A standard 306-48 build, its resolution trimmed down from 640x480 to 400 lines in 8bit color should work amazingly well as I currently plan. It's far better than 512x324 even in grayscale. Purists from either camp could knock it back to single bit and maybe even be turn down the backlight? :lol:

Maybe drop a RaspBerry Pi 6+ or better into the HDD Bay one day along with the SCSI2SD and browse the web at 1280x800@24bit at the press of a macro. :approve:

< /TLDR TANGENT >

 

[johnklos]

128MB is a hard ceiling for all of the 030 Macs right up through the IIfx. Apple hobbled every Quadra to 136MB with with a useless bank of 8M on the system board excepting the 900 and 950, which had the full four acres of 32pin SIMMs on their boards.

Not exactly. The Quadra 630 could take 196 megs (one 128 meg SIMM, one 64 meg SIMM, plus four on the board), and if you're running NetBSD or can write your own ROMs, you can get 260 megabytes in a Quadra 610 or 520 megabytes in a Quadra 650 / 800.

 

[Trash80toHP_Mini]

Thanks, two hours ago I was half asleep and fact-checking on LEM. :   Mac II series stuff is correct, no? Nothing there addresses the use of 128MB SIMMs anyway AFAIK? Shoulda' remembered that a whole bunch of late Quadras could be bumped over the 136MB mark when those hit the streets.

Quadra 630 shipped with only in single SIMM Slot configurations. I've been trying to get my grubby little paws on an LC or Performa DOS compatible board or 640 with two slots for twenty years. I have an itching desire to desolder some RAM and jumper RAS lines from those pads to a 128MB SIMM in the "64MB Slot." Heck, if I could find the thing, my original Q630 I would have had that wired up by now.

 

[trag]

Actually, the Quadra 610 and 650 did have a pretty hard 136 MB limit on RAM, even when larger modules were installed, until bbraun (or was it Doug, darn, now it's been too long) dug into the firmware a bit and changed some boot parameters so that the memory controller was allowed to "see" more memory.   There was a thread on it around here before the system change.  I hope its still around.   IIRC, he took the 650/800 up to 512MB or thereabouts.  Part of that may have been forced into a RAM Disk. 

On the SE/30, each bank of four 30 pin SIMMs is the equivalent of one, single bank, 72 pin SIMM.    If you built the machine with two 72 pin SIMM sockets, they'd be limited to 16MB or 64MB SIMMs.     32 MB 72 pin SIMMs are not going to work in such an adaptation, because almost all the 32 MB SIMMs in existence are 2 bank SIMMs.  They're made out of two separate banks of 16 MB.  Banks are controlled by having independently controlled RAS lines.  There's no reason to believe that the there are any more than two sets of RAS lines in SE/30.  

One possibility would be to modify the board for a single 72 pin socket.  Combine the signals for the two SE/30 memory banks onto a single 72 pin Socket, which would support one or two banks.    That way, one could install a single 16, 32, 64 or 128 MB SIMM and it would work.

If you go with two sockets, then the only way to get 128MB would be with two separate 64MB SIMMs and again, 32 MB SIMMs would not work, or would only be seen as 16 MB.

 

[trag]

Thanks, two hours ago I was half asleep and fact-checking on LEM. :   Mac II series stuff is correct, no? Nothing there addresses the use of 128MB SIMMs anyway AFAIK? Shoulda' remembered that a whole bunch of late Quadras could be bumped over the 136MB mark when those hit the streets.

AFAIK, the only Quadras that could go over 136 were the Q605 variants and the Q63X version that had two SIMM sockets (636?).  

The C/Q 610/650/800 all have, as shipped, limits of 136 MB.  As it turned, as mentioned in my previous post, that was a limitation of the firmware and how memory was detected at boot time, not a limit of the memory controller hardware.   Apparently, the firmware didn't check the higher Row (or column?) address lines that are needed when larger capacity memory is installed.

Let's see, a 16MB bank in 72 pin memory is 4M addresses X 32 bits (4 bytes).   4M addresses is represented in 22 bits.

A 64MB bank in 72 pin memory is 16M addresses X 32 bits (4 bytes).   16M addresses is represented in 24 bits.

The address is "multiplexed" (split up) as ROW and Column addresses.

The 30 pin and the 72 pin SIMMs support a maximum of 12 address pins/signals.    So, as you can see, 16M addresses is the maximum possible (12 + 12 = 24 bits => 16M addresses), and a single bank of memory, which is 4 bytes wide, cannot be larger than 64MB (16M X 4bytes).  

Anyway, a 16M bank can be 10 X 12, 11 X 11 or 12 X 10, in theory.   A 64 MB bank must be 12 X 12.  

The Quadras were seeing 64MB SIMMs as 16MB, because they were ignoring a couple of the address bits. 

 

[johnklos]

AFAIK, the only Quadras that could go over 136 were the Q605 variants and the Q63X version that had two SIMM sockets (636?).  

Quadra 630 shipped with only in single SIMM Slot configurations

Yes, my mistake. I just remember it's a Quadra 630 form factor. I have one of these with 196 megs.

The Quadra 610 / 650 / 800 can use larger SIMMs with a custom ROM, yes, or when running NetBSD. I tried looking in to ways to warm reboot the system so that Mac OS would see the larger SIMMs after a reconfiguration of the memory controller, but I couldn't figure it out, and if Rob Braun also couldn't, then it's probably not doable without lots of work.

His page about this is still available here:

http://www.synack.net/~bbraun/djmemcrom.html

Doug's ROM SIMM project can be used to get the updated ROM image on to a system.

 

[johnklos]

Amiga 3000 machines and the i386SX Bridgeboards used ZIP memory, which was dense for its time and allowed for static column RAM for faster access in m68030 machines. Otherwise, they were a huge pain - hard to find, difficult to install without bending pins and usually expensive.

They were handy in one respect - adapters for SIMMs were easy to install because you just needed to line up pins and push. While a new SE/30 motherboard would of course be easy to make with a 72 pin SIMM socket, adapting current 30 pin sockets to a 72 pin socket would require flexible ribbon cables. It's a shame it isn't easier to get a 72 pin SIMM on an SE/30.

I'm looking forward to seeing where this goes

 

[NJRoadfan]

Keep in mind that 30-pin SIMMs are 8-bits wide, hence why you need 4 of them for one bank of 32-bit wide memory. A SE/30 would technically only support 2 72-pin SIMM slots (up to 64MB max in each) without some additional work. Depending on chip availability/compatibility, it might actually be cheaper to just solder 128MB of compatible memory onto the board directly.