My First Quadra

[Themk]

Hello,

So, I got my first 040 powered system recently! I really like it!

It started life as an ordinary Performa 638CD. In its factory configuration, the machine was pretty fast, with its 33MHz LC040.

I have been busy though with upgrades.

• 128MB RAM SIMM in the SIMM slot, for a total of 136MB
• Full 68040 to replace LC040
• Ethernet
• SCSI mod for internal SCSI disk (or external SCSI enclosure) In an attempt to try to get A/UX up and running.
• 40MHz, which will make it as fast as Quadra 840av!

Picture time:


I decided to make this thread as I am in the middle of some upgrades for it!

 

[omidimo]

Welcome to Qlub! The 040s are pretty damn awesome. You might get addicted and start getting more Quadras. 

 

[Themk]

Ahh thank you omidimo, the Qlub is certinally fantastic. I love the 040 performance, it is very responsive!

I've got some more pictures.

Ethernet has been installed!

 

[Themk]

As a bonus, to show the 040s FPU at work, I wanted to show Graphing Calculator (NuCalc).


 
 
Now, it doesn't seem like a ton of people know this, but this is the 68k version of Graphing Calculator 1.0. I think a lot of people see Graphing calculator is PPC only. As distributed by Apple, it was PPC only, but there was a single 68k version, which is the version I am running right now.
 
Executable title is "NuCalc 1.0 - FPU"

 

[Gorgonops]

[*]SCSI mod for internal SCSI disk (or external SCSI enclosure) In an attempt to try to get A/UX up and running.

Unfortunately A/UX isn't going to happen, unless you can miraculously get your hands on the source code and documentation needed to build new kernel drivers. This isn't some trivial case of needing to ResEdit a resource file or something; unlike regular Mac OS A/UX doesn't use the low-level hardware drivers in ROM so the *significant* underlying hardware differences between the "LC class" machines (which the 630/640 basically is, with IDE tacked on) and the "real" Quadras is a deal breaker.

For A/UX the newest/cheapest boxes that will work are the 610 and 650. (The AV machines, 660 and 840, don't work either.)

 

[Themk]

Yeah, I know about A/UX architecture, and how the Macintosh Toolbox is not really used.

A/UX was kind of an afterthought, the real prize is NetBSD, which does indeed work with the 630-series machines.

 

[Gorgonops]

Okay, then... although if you already knew that I'm curious what you plan to try to get around it.

 

[Johnnya101]

Woo, cool! Glad to see something neat happen with that Performa!

Gotta get AUX running on my 700...

 

[Themk]

Okay, then... although if you already knew that I'm curious what you plan to try to get around it.

We'll burn that bridge when we get there. I still have to get my SCSI cabling all sorted out!

Dev Note keeps talking about how similar the hardware is to the 610 series, which gives me some hope.

 

[Gorgonops]

Dev Note keeps talking about how similar the hardware is to the 610 series, which gives me some hope.

I'm looking at the Dev note and it says it's like the 610 in two places (well, okay, three if you count the statement about it being in a low profile case "similar to the 610"). One of them is about the ROM being similar "with necessary changes to support machine-specific hardware", followed by an enumeration of all the things that are different which includes a new memory map and a different video controller chip, IE, A/UX game breakers. (Note specifically how it says that the Valkarie custom IC is "similar to the DAFB IC used in the Macintosh Quadra models". DAFB was used in the 605, IE, an LC-class machine, not the 610.) And the other is an obvious misprint; there's a sentence in the section about the I/O expansion slot that says:

"Like the expansion slot in the Macintosh Quadra 610, the I/O expansion slot the Macintosh LC 630 and Macintosh Quadra 630 computers supports many PDS cards designed to operate with the MC68030 bus.."

but the slot in the 610 *is not compatible* with LC expansion cards; it's a different 68040 PDS slot shared with the 650 and 800. The document is obviously referring to the Quadra 605, for which these statements regarding the expansion slot are true.

Since I don't know anything about how the A/UX device drivers are structured or *exactly* how different the chips involved are (they are both fairly dumb framebuffers) I'll grant some metaphysical possibility exists that there might be some "resource" section of the kernel binaries that could be hacked to initialize the hardware enough that A/UX could run in some capacity on it, but if you manage it it'll be a feat never before accomplished.

 

[Trash80toHP_Mini]

"Like the expansion slot in the Macintosh Quadra 610, the I/O expansion slot the Macintosh LC 630 and Macintosh Quadra 630 computers supports many PDS cards designed to operate with the MC68030 bus.."

but the slot in the 610 *is not compatible* with LC expansion cards; it's a different 68040 PDS slot shared with the 650 and 800. The document is obviously referring to the Quadra 605, for which these statements regarding the expansion slot are true.

Maybe, maybe not, the text seems equivocal to me if it's within the context of describing differences between the 630 and the 610. If it's part of the discussion about which ROM the 630 is based it's likely not a misprint. One way I'm reading that is that the 630 PDS is architecturally similar to that of the 610 PDS in terms of the Slot Manager, not that the physical slots are the same in terms of the connectors.

The 630 is similar to the LC class machines in terms of the implementation of its expansion interface and use of lower cost, less capable expansion cards. But it sounds like its architectural forbears may place it squarely in the NuBus Quadra camp. Implementation of the CS Slot puts it far ahead of the Quadra 605 in terms of its I/O mapping. How do the Block Diagrams compare within that trio?

 

[Gorgonops]

Maybe, maybe not, the text seems equivocal to me if it's within the context of describing differences between the 630 and the 610.

No, it's not equivocal; the text as it appears in the I/O section is factually *wrong* because a 610 cannot accommodate an LC card. It may well be abstracted similarly in the "Slot Manager" but, remember, a bare-metal OS doesn't give a hoot about the Slot Manager, it cares about the physical hardware addressing.

As for it being part of the "NuBus Quadra Camp", the devnote for the 605 describes it as "electrically similar to the Quadra 650" and points out the similarities of the architecture, which are "real"; it *is* conceptually a dumbed-down Q650/800 with fewer RAM/VRAM slots, the NuBus controller hacked off (it's also hacked off in the 610 minus the Nubus adapter), and an LC bus adapter tacked on instead, but the emphasis is on *conceptually*. They use Memory/Framebuffer controllers and I/O bridges which are different enough to make them incompatible at a bare hardware level. But, if you want to talk about block diagrams the Q605 and 650 resemble each other more than *either* resembles the 630, which shuffles the discrete SCSI and SCC chips from the 605/650 into the F108 memory controller, divorces the Framebuffer to a separate chip (it's in the MEMC/MEMCjr memory controller in the previous devices), and otherwise shuffles the cards significantly. Notably, though, both the "PDS" slot (which still accommodates LC cards, ala the 605) and the "comm slot" are hanging off the "PrimeTime II" chip, which appears to be an extended version of the "Primetime" chip in the 605, which contains the LC slot transceiver. So, again, no, the 630 has more in common with the 605 than the 650.

 

[Trash80toHP_Mini]

".  .  .  supports many PDS cards designed to operate with the MC68030 bus.."
 
It says with the MC68030 bus, not in the LC Card Slot. Could be a typo, I'm not arguing, but there's a big difference between bus and slot. Lots of cards were designed to run on the MC68030 bus that found themselves in LCPDS, NuBus and 030PDS implementations like the Radius Pivots, Spigots and several NICs. That's where the Slot Manager shines.
 
I don't think there's a snowball's chance in hell the 630 comes close to supporting A/UX either. I don't know about bare metal B.S. but A/UX was obsolescent at best at the 630's intro. I've never understood the fascination with this aberrant offshoot of Mac OS development. Does something make it more interesting that real UNIX implementations other tha the fact that it'll run on a few Macs?

 

[Gorgonops]

".  .  .  supports many PDS cards designed to operate with the MC68030 bus.."
 
It says with the MC68030 bus, not in the LC Card Slot.

It's sort of splitting hairs. The original LC card slot *is* a 68020/030 PDS slot, or *almost* one, anyway, it's missing a few address lines and signals. The LC III added an extension (from 96 to 114 pins) that made it a "True" PDS slot, or at least a complete one. It's the LC III one that the bus converter in the PrimeTime IC emulates.

Again, the PDS slot in the 610 (and 650/800) is a *TRUE* "PDS" slot because it's basically the raw 68040 bus brought to a connector. The bus converter in the LC 475/Quadra 605 may be less complex than a NuBus bridge but it's still a *bridge*, and it's a bridge the 610 doesn't have.

To use an imperfect analogy, the Quadra 605 is sort of like an ISA-bus equipped 486 machine; ISA is essentially a 286 PDS, and requires more and more complex glue to use on later Intel CPUs, just like running a card designed to talk to a 68030 needs a little glue to talk to a 68040. The 650/800 is like a dual EISA/VL-Bus 486 (EISA being roughly equivalent to NuBus if you forget about its backwards ISA compatibility, and VL-Bus is basically a 486 PDS), and the 610 is like a 486 with *just* a VL-Bus slot that can be converted to an EISA slot by plugging a bridge in. The BIOS (IE, slot manager), mostly papers over the difference for DOS, but these differences are going to matter a lot if you try loading Linux on any of them.

... and, again, by the analogy above, the 630 is a 486 that has both an ISA slot and some *other* kind of slot. What's a Comm slot most like, S-100? [ xx( ]

 

[Themk]

I don't think there's a snowball's chance in hell the 630 comes close to supporting A/UX either. I don't know about bare metal B.S. but A/UX was obsolescent at best at the 630's intro. I've never understood the fascination with this aberrant offshoot of Mac OS development. Does something make it more interesting that real UNIX implementations other tha the fact that it'll run on a few Macs?

The appeal of A/UX is that it will run your Mac programs without requiring a reboot. That's about it. A real UNIX implementation is better, and, as I mentioned, NetBSD is the real prize here! I will probably spend 2-3 hours at best muddling my way through the A/UX installer, just trying to see how far I can get; beyond that, SSW is for running Mac programs, and NetBSD is for using your 68040 as a UNIX workstation.

Do I think that A/UX will run? I'm not banking on it. In fact, I'm with everyone else and expecting it not to run. As mentioned earlier, I want to as an exercise just see how far I can get through the install process. Beyond that, A/UX doesn't matter to me. If I really wanted to run A/UX I always have the SE/30.

I have some more thoughts on DevNote, and I have some things I want to say in response to you jt and Gorgonops, but that will come in a later post.

 

[Gorgonops]

I'm pretty sure the installer is going to loudly tell you to get bent *very* early in the process, but good luck.

 

[Themk]

Well, I already got past the first hurdle, which is to get the installer floppy to boot.

I deleted the system enabler already on the floppy, and copied the system enabler from my 7.1.2P installation to the disk, that enabled the A/UX installer disk to boot.

It also recognizes my A/UX CD in my CD-ROM drive. I just don't have anyway to put a SCSI HDD on my SCSI chain ATM :-/

How well would running wires from the SCSI port to a breadboard, and then connecting the wires from the bread board to a 50-pin SCSI cable... I have a scrap ATX PSU with molex connector on it, for HDD>

 

[Themk]

Annnnnnddd that's where it stops.

The installer floppy will then take the A/UX install disk, and then it boots to a black screen....

That's okay, I didn't expect it to "just work". Looks like my journey with A/UX is over on the 630 [ ]

Wait, maybe it needs the new system enabler in the CD-install disk......

This is what hacking is all about

 

[Themk]

After that I can see the A/UX startup, but "autorecovery" fails. How do I bypass autorecovery?

It's not about bypassing autorecovery, it's about my broken install mediea. In the process of modifying the *.toast to have the new system enabler on the disk, that also managed to screw up the A/UX install disk.

A/UX doesn't use system enablers, so, this is the end of the line for A/UX.  I'm not too depressed though.

NetBSD here I come! If Mini vMac or Basilisk II runs okay under NetBSD, then I won't even be missing A/UX's killer feature!