PowerExpress - PowerMac 9700

[jimjamyahauk]

Hi,

I thought we could use this new thread to cover info about the PowerExpress 9700 prototype powermac and efforts to get them booting.

Past information to getting mine to boot is on this thread: viewtopic.php?f=12&t=9297 and over at applefritter: http://www.applefritter.com/node/23597

What would be interesting is trying to get the remaing ones out there booting. A place to start would be trying to get into OpenFirmware on ones that aren't booting at the moment. I found that this was only possible using a serial connection to a terminal programme on another mac, as the default output was via the serial/printer port, rather than onscreen.

A couple of other things are to take some pics of the motherboards and try and figure at which stages of the EVT process they were produced, and also dump some of the software in the other flash chips ( Trash80toHP_Mini - would you be willing to either do this, or send me your flash chips for this purpose?)

James.

 

[Trash80toHP_Mini]

That'd be a big maybe there, padnah!

Take a gander at: PExProjRedux to see what PEx shenanigans I've been up to lately!

It might be good to pick up on that thread and close out this one, or not, whatever you think! [] ]'>

 

[trag]

Okay, I've read all the threads linked to. And the threads linked to by the threads that you linked to. And so on.

I had never really read all that stuff before. And I went and updated the 68kmla Wiki entry for ROM DIMM pinouts to mention that the PEX is included in Apple's "Universal" ROM module.

Any idea if Mark Crutchfield actually lives in Austin, TX, where he got his PEX? If so, and if he's willing to help, that could greatly simplify testing of the ROM DIMM. Of course, if he'd let me dissect his DIMM, that would simplify matters even more..... I can put it back together when I'm done. I just tried PMing him, so with a little luck, he'll PM me back. If not, there's an email link on his site I can try.

How can we be confident that the PEX dump is an accurate reflection of what's actually on that DIMM?

Trash80, to answer your question from almost two years ago on the Applefritter thread, I ran off about 200 ROM DIMM module circuit boards several years ago. I still have a few around and that's what I'm using for Jame's ROM. However, I can't get the sockets to work, so we're going to have to solder the chips down, and I would like to be as certain as possible that we're putting the correct code into the chips before I solder them down.

If that works, I have a few more circuit boards on hand, but even if I run out of circuit boards, we can (if we can find them) use slightly modified ROM DIMMs from Power Computing machines.

From the photos, Mark Crutchfield also has a ROM DIMM, and I wonder if his has the same contents as the DigiBarn DIMM...

I used to have sockets built just to suck the contents of a ROM DIMM to a chip programmer, but I tossed them during my last two years of house cleaning figuring I'd never use them again. I could build new ones (given sufficient 7200 logic boards to cannibalize) but it's a nightmare of wire wrap to build them. And they don't work with my Needham Programmer any way, which had a lot to do with tossing them. They do work with a BP Micro 1600 which the folks at Pioneer Data Services had, but I don't know if that door is still open or not. They may not even exist any more. It's been most of a decade...

 

[Trash80toHP_Mini]

What chances do you guys think the low-level bootstrap routine & accelerator enablers on a ROM DIMM attempt has of succeeding?

Just in case I feel the urge to entrust some shipping service with my ROMs, where are you guys located? :?:

That'd be a very big decision making factor! :-/

edit: Never mind got the needed info from doing some homework! :I

 

[Trash80toHP_Mini]

It looks like all the appropriate PEx Links might have been posted in this thread, so I'll leave my other thread for the Hardware Hackage Shenanigans!

Since you have determined that both the PLCC ROMs and the ROM DIMM are required for PEXBoot to be achieved, maybe we can get this thing rolling without entrusting my PLCC Rev ROMs to the mails.

trag, if you'd be so kind as to make a "DigiBarn ROM set: DIMM & PLCC" available to me, I'll test them on my board. If that boots, then I'll see if I can get it to boot with my PLCCs installed in order to dump them for posting. If the attempt with the new combination fails, I can more confidently post my PLCCs without too much worry about another, possibly more permanent, recurrence of the UPS OOPSIE! [] ]'>

PM me with whatever arrangements we need to make to get mine properly configured for bootup. :approve:

Will my 9500 or 8500 CPU Cards be adequate or will I need to find another flavor for this attempt. :?:

Has anyone gotten a G3 or a G4 Accelerator's drivers to load and boot yet? }

 

[trag]

if you'd be so kind as to make a "DigiBarn ROM set: DIMM & PLCC"

available to me, I'll test them on my board.

I'd be happy to, but my progress is on hold until we discuss whether the ROM DIMM dumps are likely to be an accurate representation. I think James mainly needs to address that. The question is whether the dumps he made yielded the same file every time he dumped it. If so, that's pretty straight forward.

 

[Trash80toHP_Mini]

Sounds great! [] ]'>

LOL!! I'm not going to start being impatient after all these years! :lol:

 

[jimjamyahauk]

Hi,

I've tried the rom verification utitlity but no luck - it reports ROM checksum error and also unindentified rom (although that's not surprising, as the 9700 is not on the list.!)

I've also compared all the ROM dumps I have using FileBuddy. There are 3 for each utility. Each utility's ROM dumps are indetical, so I'm sure that I have a good copy taken with each one.

Apart from the utility that took 4mb - the other 2 which took 3mb dumps - are also identical.

I suggest that the ROM checksum in PowerMacs only validates the Mac OS Toolbox portion and skips the hardware-specific 1MB ROM code. This is because in order to get to to loading the Toolbox the hardware must already be initialised using the 1MB section for this.

Anyway, I guess the real issue is trying to find out on the original ROM simm how much information is contained, and if there is any padding.

From Tattletech (running on the booted PEX) the following information was obtained - which suggests that the 3mb file is the correct size - and also a clue to at what address this part of the rom starts:

• ROM :
  ◊ ROM Size = 3,145,728 Bytes
  ◊ ROM Version = $77D  (125)
     + ROM is 32 Bit Clean = Yes
  ◊ ROM Sub-Version = $39F1
  ◊ ROM CheckSum = $46001F1B
  ◊ ROM Start = $FFC00000
  ◊ MacOS Boot-ROM Name = [None]
     + MacOS Boot-ROM Version = [NA]
     + MacOS Boot-ROM Date = [NA]
     + MacOS Boot-ROM File Version = [NA]

I added the ROM checksum information to the code for the ROM Checker:

   // Source 3:  Tattletech running on PEX
       case 0x46001F1B: cout << " PowerMac 9700"; break;

and then tried it again on one of the 3mb files - which it correctly identified:

Macintosh-2:~ test$ /Volumes/1TB/Documents/PEX_ROMS/MacROMChecksum/build/Release/MacROMChecksum /Volumes/1TB/Documents/PEX_ROMS/Digibarn/SIMM/GetRom/ROM\ 01

ROM Length: 3 MiB

ROM Checksum FAILED.

This is a PowerMac 9700 ROM.

So, despite the ROM checksumming failing, I think that the files are good - as the checksum code reported by tattletech on the booted pex matches the information in the dumped rom file.

I also ran the checksumming on the 4mb file - which was also correctly idenitfied. Comparing (using TextWrangler) the 3mb and 4mb files the 4mb is identical to the 3mb file but has extra information at the end. I suggest that this could be the 1mb from the flash chips which it has managed to read - as perhaps all of this is loaded on the live system - much as was suggested that many rom utils only read a 3mb - apart from CopyRom (which wouldn't run on the PEX) and apparently rom grabber. This could also tie-in to the theory about in a normal mac rom 3mb being the toolbox and 1mb being hardware/model-specific.

All files are at: http://www.jkalittle.co.uk/jkalittle.co.uk/digibarn.zip

One final point could be to compare the digibarn flash chips and search for this within the 4mb rom dump - to see if this indeed apended. I'm not sure how to do this.

James.

 

[trag]

Okay, if all the files contain identical code, and the 4MB file is identical to the 3MB file but with stuff at the end, that pretty well answers what to do next. Take one of those files and program it onto chips and then solder the chips down.

I am actually happy that the 4MB file has 1MB of padding at the end. I was worried that the 3MB file might do something weird like have 1MB of padding at the beginning, which the dumper ignored, and then the 3MB of content in the top addresses. The 4MB dump having 1MB of padding at the end pretty strongly implies that the usable content starts at address 0 of the ROM DIMM.

The reason I had concerns about that was because of the bizarre arrangement of the PLCC chips providing hardware boot code and the toolbox on a separate DIMM. This leaves me wondering what the logic board is doing. I guess it must put the two types of storage at different addresses.

Anyway, there are about half a dozen ways/reasons I can think of that this still wouldn't work, but I only see one way in which to proceed at this point based on the information and resources we have available to us. So, as soon as I can clear some time, that's the direction I'll go.

My work bench is even more accessible than the last time I did any work on this project. I've actually spent close to two years cleaning up junk. Too much junk! Of course, with an eight-year-old under foot, the actual clock time spent on the neatening is much less than two years elapsed time implies....

 

[Trash80toHP_Mini]

Bumperoonies! )

More seriously, has any headway been made on getting this project on a roll? :?: