Thread
An SE/30 that has a socketed motherboard is on eBay
Well actually, the amazing things these so-called "nuts" create are like this -- which, sadly, are wonderful modern add-ons that I've not seen for the compact Mac. True, the Mac is a more complex beast. Even so, there are just so many things one could do with the SE/30's PDS slot! I've mentioned this in another thread, but I would just love to see a flash disk that works in the SE/30 via PDS, theoretically bypassing the rather slow SCSI implementation.
It would have no effect whatsoever.
The CPU speed is set by an oscillator on the motherboard. That is part of the reason all CPU accelerators have an extra electronics board between the CPU and the motherboard. There's another oscillator (and other stuff) on that board that sets the upgrade CPU's speed.
And before you ask - no - swapping out the motherboard oscillator for a faster one won't get you up to 50MHz either. Every other component on the motherboard is specified for the 16MHz clock as well. At best, you might be able to get to 20MHz. Output Enablers have oscillator kits.
[mistake redacted]
All of the PowerCache cards that I have seen use the MC68030RC50B chip, which may well have been the latest revision then available. None of the 030s wears a heatsink. By 2005 there was a revision C. While all things are possible, I have to doubt that Motorola, or Freescale, would go to the trouble of producing a revised processor with 0.8 micron technology only so that it could be clocked up 25%, rather than be designed for 50MHz from the outset.
We know/suspect that many components are classified after manufacture into less than, spot on and more than before they are labelled, but rev. C 50MHz 68030s would not fit convincingly into that paradigm. It would be interesting to know the unit price of the rev. C chip.
de
Ahem. Clearly there is bloodlust associated with this socketed motherboard and related accelerator card. So much so that my previous post was overlooked. :beige: So again I ask: did Apple actually manufacture the SE/30 with a socketed CPU, or was this a this party modification specifically for the aforementioned card? If Apple produced the board, why? To accommodate the third party companies, or some nefarious internal project of their own?
I believe the earliest of these boards were manufactured this way. The one I have has not been modified, and I seem to recall reading this to be the case somewhere.
UPDATE: According Larry Pina's book "Mac Classic & SE Repair and Upgrade Secrets" there were two SE/30 motherboards:
P/N 820-0260-10, 1988 (this is the one I have)
P/N 820-0260-A, 1989
This link has pictures of the 1989 board, top and bottom views: http://www.kevinomura.com/macs/se30/index.html
Notice the little wire (the "engineering change") on the underside is not present on this one....
Allow me to muddy the waters further. In five SE/30 MLBs, only one of which is in use, I have just found the following:
1) 820-0260-10 dated 1988 on the MLB, socketted for a CPU, soldered-in MC68882FN16A, assembled in Singapore. A white sticker next to the PDS connector shows SG9200TA198A. On the underside of the board, C70 and C71 are positioned directly under the socket.
2) 820-0260-A dated 1989 on the MLB, socketted for a CPU, soldered-in MC68882FN16A, labelled as assembled in Singapore in 1989 (SG9...). A white sticker next to the PDS connector shows SG940J0E380A. On the underside of the board, C70 and C71 are positioned directly under the socket.
3) 820-0260-A and 640-4198, dated 1989 on the MLB, soldered-in MC68030RP16B and MC68882FN16A, labelled as assembled in Singapore in 1989 (SG9...). A ? sticker next to the PDS connector shows ? (The Mac is reassembled at the moment.) On the underside of the board, C70 and C71 are positioned between the socket and the PDS. This MLB is in use in a 32MB/4GB/MacCon 30si machine.
4) 820-0260-A, dated 1989 on the MLB, soldered-in MC68030RP16B and MC68882FN16A, labelled as assembled in Ireland in 1991 (CK1...). A buff sticker next to the PDS connector shows CK1071S900YA. On the underside of the board, C70 and C71 are positioned between the socket and the PDS.
5) 820-0260-10 and 640-4198-V, dated 1988 on the MLB, socketted for a CPU, soldered-in MC68882FN16A, assembled in (?). An orange sticker next to the PDS connector shows BK9042591. On the underside of the board, C70 and C71 are positioned directly under the socket. This is the MLB that I mentioned in another post as having the broken Bank B RAM clips. This may be the oldest of the five boards, although the manuscript 'V' after the inventory/part number may contradict that.
Seemingly, there are socketted and non-socketted MLBs in both revisions. Nothing indicated that the sockets are after-market additions by a third party. The traces around C70 and C71 are part of the board's original layout in all cases.
de
Supplemented on 23 Feb. 09
1) 820-0260-10 dated 1988 on the MLB, socketted for a CPU, soldered-in MC68882FN16A, assembled in Singapore. A white sticker next to the PDS connector shows SG9200TA198A. On the underside of the board, C70 and C71 are positioned directly under the socket.
2) 820-0260-A dated 1989 on the MLB, socketted for a CPU, soldered-in MC68882FN16A, labelled as assembled in Singapore in 1989 (SG9...). A white sticker next to the PDS connector shows SG940J0E380A. On the underside of the board, C70 and C71 are positioned directly under the socket.
3) 820-0260-A and 640-4198, dated 1989 on the MLB, soldered-in MC68030RP16B and MC68882FN16A, labelled as assembled in Singapore in 1989 (SG9...). A ? sticker next to the PDS connector shows ? (The Mac is reassembled at the moment.) On the underside of the board, C70 and C71 are positioned between the socket and the PDS. This MLB is in use in a 32MB/4GB/MacCon 30si machine.
4) 820-0260-A, dated 1989 on the MLB, soldered-in MC68030RP16B and MC68882FN16A, labelled as assembled in Ireland in 1991 (CK1...). A buff sticker next to the PDS connector shows CK1071S900YA. On the underside of the board, C70 and C71 are positioned between the socket and the PDS.
5) 820-0260-10 and 640-4198-V, dated 1988 on the MLB, socketted for a CPU, soldered-in MC68882FN16A, assembled in (?). An orange sticker next to the PDS connector shows BK9042591. On the underside of the board, C70 and C71 are positioned directly under the socket. This is the MLB that I mentioned in another post as having the broken Bank B RAM clips. This may be the oldest of the five boards, although the manuscript 'V' after the inventory/part number may contradict that.
Seemingly, there are socketted and non-socketted MLBs in both revisions. Nothing indicated that the sockets are after-market additions by a third party. The traces around C70 and C71 are part of the board's original layout in all cases.
de
Supplemented on 23 Feb. 09
Gah. That's the second time I've confused my '040s and '030s on this issue. I shall [redact] so as not to mislead others in the future.
Amigas seem to have cornered the market on real 68030/50's. By the mid 90's when accelerators for Amigas were coming out with fast 030's the 50 Mhz speed existed and was cheap.
A socketed CPU in one of Apple's flagship products!? And from the beginning to boot ... what were they thinking? Surely they didn't do that for just third party manufacturers to take advantage of ... clearly Apple had plans for it that they abandoned, probably after the SE/30 began to erode sales from their high-end Macs.
Speculations could run rife this long after the event, but some may have germs of truth in them. Recalling the reported difficulty that Apple engineers had in getting some capacities incorporated into MLBs, even to the point of having to resort sometimes to subterfuge (Classic II, perhaps?), it is just possible that the SE/30, as the 'wickedest' of the closed-box AIOs, did have some grand things conceived for it as possible aftermarket upgrades. If that was canned, it still left the way open for a company such as DayStar to be invited, overtly or covertly, to consider a plug-in CPU replacement. DayStar did offer to replace soldered-in CPUs with socketted CPUs at one stage, but they and others settled for the PDS in the end.
I have seen the SE/30 referred to as a 'condensed IIcx', which makes it clear that if the IIcx can be made into a zinger (mine has a DayStar 33MHz PowerCache), the SE/30 might have done at least as well with a socketted adapter and a suitable accelerator-card, and thereby saved the PDS for other expansions. The saddest part, however, is the apparent disappearance of all things to do with CPU-oriented plug-ins, with the PDS-oriented not far behind.
de
I have seen the SE/30 referred to as a 'condensed IIcx', which makes it clear that if the IIcx can be made into a zinger (mine has a DayStar 33MHz PowerCache), the SE/30 might have done at least as well with a socketted adapter and a suitable accelerator-card, and thereby saved the PDS for other expansions. The saddest part, however, is the apparent disappearance of all things to do with CPU-oriented plug-ins, with the PDS-oriented not far behind.
de
At the same time that Apple produced SE/30 variants with socketed/soldered 68030s, they did the same thing with the IIcx. Hence the existence of the Daystar IIcx PowerCache adapter that provides a IIci cache slot from a 68030 socket. I own one of the IIcx adapters, as does at least one other member here.
Don't ascribe the existence of socketed/soldered boards to commercial conspiracy -- it is more likely that Apple adjusted their production lines according to what Motorola could ship. The 68030 was a very desirable chip, so whilst Apple may have been the best customer, they may have needed to be flexible for a few years.
Daystar became very cosy with Apple in late 1992 and 1993, in order to create the PPC 601 accelerators. Before that time, there is evidence of cosiness in the way that IIci PowerCache accelerators are so well supported, particularly in A/UX. The manual for my SE/30 68030 socket accelerator PowerCache is dated 1992.
As Equill notes, owners of a soldered processor board could return it *for free* for a socket transplant. If the SE/30 production lines were shifting predominantly to soldered processors, this would provide an incentive to Daystar to produce an accelerator for the PDS slot.
Don't ascribe the existence of socketed/soldered boards to commercial conspiracy -- it is more likely that Apple adjusted their production lines according to what Motorola could ship. The 68030 was a very desirable chip, so whilst Apple may have been the best customer, they may have needed to be flexible for a few years.
Daystar became very cosy with Apple in late 1992 and 1993, in order to create the PPC 601 accelerators. Before that time, there is evidence of cosiness in the way that IIci PowerCache accelerators are so well supported, particularly in A/UX. The manual for my SE/30 68030 socket accelerator PowerCache is dated 1992.
As Equill notes, owners of a soldered processor board could return it *for free* for a socket transplant. If the SE/30 production lines were shifting predominantly to soldered processors, this would provide an incentive to Daystar to produce an accelerator for the PDS slot.
What Motorola could supply in the way of 68030/16MHz processors at any given moment throughout the 2-3 years of the SE/30 and IIcx is indeed a plausible reason for Apple to have given itself the ability to switch at will from socketted to soldered-in assembly. I have fewer IIcx socketted MLBs (to wit, 1) than I have socketted SE/30 MLBs, but I have the impression that that for the IIcx is much less commonly found.
Physical as well as electronic considerations impelled DayStar's dedicated adapter for the IIcx, but the SE/30 and IIsi share a PDS electrically, if not the geometric orientation of an attached adapter card. I have two (differing) IIsi/IIci PDS adapters, one of which is more suited to the SE/30 in its shape. Was there a dedicated IIsi/IIci adapter for the SE/30? Or have I misremembered what JDW's monochrome pic. is about?
de
Physical as well as electronic considerations impelled DayStar's dedicated adapter for the IIcx, but the SE/30 and IIsi share a PDS electrically, if not the geometric orientation of an attached adapter card. I have two (differing) IIsi/IIci PDS adapters, one of which is more suited to the SE/30 in its shape. Was there a dedicated IIsi/IIci adapter for the SE/30? Or have I misremembered what JDW's monochrome pic. is about?
de
So, it isn't as simple as swapping out a single oscillator as in the IIsi ?
It's exactly as simple * - to go up a small amount, say to 18 or 20MHz, tops. No way are you getting to 50MHz without something asplodicating.
Think about it - would you expect a 1966 Volkswagen beetle to do 250MPH, even with a Ferrari engine in it? That's what you're asking the poor little SE/30 to do, to run every component inside more than three times as fast as it was built for - and it's a twenty year old machine.
No matter how many times you ask, the answer will be the same - not gonna happen without a proper upgrade card with a separate clock - and no, you can't build one.
BTW, we bottom post and edit quotes here >
* /edit/ in fact the Master says Not Possible!
Think about it - would you expect a 1966 Volkswagen beetle to do 250MPH, even with a Ferrari engine in it? That's what you're asking the poor little SE/30 to do, to run every component inside more than three times as fast as it was built for - and it's a twenty year old machine.
No matter how many times you ask, the answer will be the same - not gonna happen without a proper upgrade card with a separate clock - and no, you can't build one.
BTW, we bottom post and edit quotes here >
* /edit/ in fact the Master says Not Possible!
It's the same on the Newton. You can't accelerator it and retain serial port functionality at the same time. That's why many Newton acceleration solutions have a switch. But an SE/30 would be a different case. I would think that you would really need to have the acceleration and access to your I/O ports at the same time, making a switch an unrealistic solution here.