Mac OS 9 can't booting, grey screen only

[FBz]

Yesterday darthnvader suggested that Vitaly run debug on one of his installer CDs and the results were that he was finally capable of getting further into the actual OS 9.2 boot process than ever before. (Needless to say that after all this time, Vitaly was ecstatic!) Only, to be confronted with the following…


 

I've suggested that now, instead of booting with extensions off - he might remove / move “all but the essential” OS 9 extensions to the disabled extensions file folder in his OS 9 partition (via OS X) and then attempt an OS 9 boot. But with the time differences between our locations, it often takes a day to hear back from him.

So that’s where this all is, now.

He has provided a nine minute video of the entire debug process that can be made available to anyone wanting to review it.
Message me and I can provide a link.

[V.Yakob]

Yeah!! Yesterday I can't believe my eyes when I rebooted MDD several times, and every time I saw an unsuccessful attempt to start Mac OS 9. I've probably never been happy about errors before.

I don't know how, but turning on debug mode helped to boot Mac OS 9 in OpenFirmware:

Code: [Select]

dev /
13fff encode-int " AAPL,debug" property
boot cd:,\\:tbxi
A universal community disk was inserted into the CD-Drive.

Today, again, only a gray screen. I started boot from Debug mode ~7 times, but Mac OS 9 did not boot, only this damn gray screen.

GaryN, you're probably right, need to send this MDD to Romania, to the gypsies, so that they remove all the curses from him. 

[IIO]

if you get a "bus error" from a fresh OS9 install then something is wrong with some hardware.

if it is not the graphics card the next i would check is if the preowner maybe broke the power of the USB ports.

[GaryN]

CLUES???

1.  Way back when, you mentioned you were running PC3200 RAM. Some asshat told you "Don't worry, it will just run at PC2700 speed".

2.  From the beginning of Macintosh time, a "bus error" has (mostly) meant the system tried to access memory that wasn't there. THAT usually meant a defective RAM stick or bad RAM slot or something related to RAM.

It would seem like a good idea (to me at least) to track down and try some proper PC2700 RAM. I have no idea if that's easy or impossible for you at this time but if you can, I think it might be a good idea. Yes, on the surface it seems unlikely that this is the problem since OSX boots without issue, BUT:

OS9 may perform a memory test at startup. Perhaps it's hanging there because it's never before seen PC3200 RAM.

This "feature" can be turned on and off but unfortunately, you have to already be in OS9 to access the Memory Control Panel and, well……
If you had another OS9-capable machine you could check the test status in the System you're trying to boot and ensure that it's off but you don't and you probably ought to have PC-2700 in there anyway and, and……

Wait!! I have yet another idea!   Please don't ask why it took this long for anybody to think of this

1. Attempt to boot OS9.
2. Immediately after the initial chime, while holding your phone or other image-capture device in one hand, press and hold Command-V with the other to invoke Verbose Mode.
3. Text will scroll up the screen as the system starts and will stop when the boot hangs. Take a pic of the text and post it. It MAY offer a real clue as to what's going wrong.
4. If the text disappears first and the screen goes gray, get pics of all you can before the text goes away.
5. Post them it/them here and cross your fingers…

[V.Yakob]

The theory is unlikely, since the frequency from 400 to 333 should be lowered by the motherboard when primary initialization is performed. I checked your theory with the question "What if he's right?"
I have several PC2700 RAM sticks and installed 2x256 mb. Yes, you're right, finding such a memory, 4*512 PC2700 is not easy, and sometimes very expensive.

I started MDD and held ⌘+V. Most likely, in Mac OS 9 it is not a protocol mode, as in OS X, but a network boot... A flashing globe appears.



I opened a new topic on booting debugging.

[GaryN]

I hoped that the OS9 memory check was activated and hanging because of the 3300 RAM. Especially when the almost-boot showed "bus error".
Oh well…

As for verbose mode, much (all?) of it is a log of the UNIX kernel coming up but I hoped (there's that word again) possibly some little thing from firmware or ? might display first.
Oh well…

This is where (no, actually we were here long ago) I would be trying a different motherboard. Yes, I know they don't exactly grow on trees there. At this point however, if you can actually find another MDD, The purchase can be justified as acquiring some spare parts.

I am otherwise (and I really hate to admit it) totally out of ideas…

[vad12]

If you can burn cds and have access to macintosh repository, you can try the Apple diagnostics tools they have hosted there. One of them should be optimized for MDDs, I'm just not sure which. Those are supposed to be the "official" hardware diagnostics tools for Apple machines. macintosh garden probably has those CD images too.

I have a Sawtooth with a busted DVD drive, so I haven't had first-hand experience with those tools and can't say much more about them.

[IIO]

The theory is unlikely

if you get a bus error, it can be any bus, including RAM. (until we know better )

RAM, USB, IDE, PCI and AGP cards are in my opinion the most interesting things to look at (a malfunctioning speaker or bluetooth device should not interfere with booting)

[V.Yakob]

Hello friends! I haven't written on the topic in a long time, but I'm back.

So, I bought another exactly the same PowerMac G4 MDD (2003), Single 1.25 MHz. And I checked all previous attempts to run Mac OS 9 on it - they are successful! It works! That is, I did everything right from the very beginning.

Everything works on this PM without any problems.
After washing, I started comparing these two computers to understand how they differ.
I changed the components one at a time and had no positive result until I changed the processors.
After I swapped the processors, the new computer stopped starting Mac OS 9 (yes, gray screen), and the old computer began to start Mac OS 9 without problems.
I swapped them again and the situation was the opposite, the new computer starts, and the old one is not again.
Obviously, it's a problem with the processor.
Processors are really different. I added several photos to the archive.

Mac OS 9 won't boot -- XC7455B RX1250DF (84L64M, QEU0313C), Samsung 337 K7D403671B-HC30, 820-1497-A, 94V-0 343P
Mac OS 9 boot -- XC7455B RX1250EF (34L64M, QGG0340D), Samsung 440 K7D403671B-HC30, 820-1497-A, 94V-0 423T

How can this be? The XC7455B RX1250DF (84L64M) processor is working properly, it passes all mathematical tests, OS X (10.2-10.5) works properly on it. Bad revision?

[FBz]

Hey Vitaly! Hello.

(Not forgotten.)

Congratulations on finally getting the thing to boot with a different CPU. One would think that with what appears to be an earlier, alpha-numeric sequential manufacture date of the RX1250DF (J5338 vs. J5344 of the RX1250EF) that there would be no trouble booting OS 9 with the DF CPU unless of course it is just faulty. AND the DF chip also appears brand new (especially considering the white surround of the die itself). No heat trace from extended use?



I considered that it (the DF) might be a CPU from a later MDD FW800 (more than once) but even that should have booted OS 9 - IF it were placed on an earlier FW400 MOBO (as has been done here often).



I am afraid that we still have a mystery, but relieved that it will now boot OS 9 with the RX1250EF.

Should such a time ever occur that mail is once again permitted between our two countries, I shall send you another working MDD CPU. In the meantime, congrats on being able to boot and run OS 9 on your mysterious MDD!

*I’d still check, double-check and re-check ALL of those pins on the back of that daughtercard. That DF just doesn’t look as if it’s ever been used extensively / completely.

[V.Yakob]

Hey, FBz!!

No heat trace from extended use?

Nope, he looks great. I also tried to  cleaned up the contact group with compressed air to get rid of dust that could interfere with normal contact.

I considered that it (the DF) might be a CPU from a later MDD FW800 (more than once) but even that should have booted OS 9 - IF it were placed on an earlier FW400 MOBO (as has been done here often).

I checked -- FW800 was not released with a single 1.25GHz CPU.

*I’d still check, double-check and re-check ALL of those pins on the back of that daughtercard. That DF just doesn’t look as if it’s ever been used extensively / completely.

I've probably checked everything 5 times already, 100% The problem is repeated on each MDDs.
XC7455B RX1250DF -- Mac OS 9 won't boot (grey screen only)
XC7455B RX1250EF -- Mac OS 9 is boot

With the fact that I ran all the tests I could find, "RX1250DF" worked fine. If you remember, I tried OS X (10.2-10.5) and even Linux Debian built for 32-bit PPC, the processor has never failed.

I have only one consideration left - the revision problem.

I'll try to sell one of these MDDs, I'll buy another daughtercard: DP 867 MHz. At our flea market, I found only one for little money.

[FBz]

I checked -- FW800 was not released with a single 1.25GHz CPU.

Ahh, you are correct my friend.
My current, COVID-addled brain confused this with the 1.0 GHz SP FW800 MDD.

Perhaps someone else has a better explanation?

**One last thing…
- IF you are completely convinced that it will not work and have tried everything else, you might try lightly brushing all of those contacts with distilled white vinegar to remove any tarnish or oxidation that might not be visible to the naked eye. I wouldn’t leave the vinegar on there for very long and might try it a couple of times between rinses with clear distilled water. Then after ample drying time… try the CPU one last time.

[peeperpc]

I have only one consideration left - the revision problem.

Agreed, the first one might have the processor replaced and it's not fully compatible with the logic board.

As it is mentioned in the Powermac G4 repair manual that ,for some revisions of logic boards, when the processor is defective, you need to replace both the processor and the board with a whole new set from Apple, not just the processor. The reasons probably were compatibility issues and Apple no longer had the right processors for those revisions of the boards in stock.

PS: Well, if the logic boards in both of them turn out to have the same revision number (820-xxxx), this would go back to a mysterious case.

[V.Yakob]

you might try lightly brushing all of those contacts with distilled white vinegar to remove any tarnish or oxidation that might not be visible to the naked eye.

I used 90% alcohol (medical) and cotton pad.

[FBz]

The vinegar has acetic acid and will "cut" tarnish and oxidation, whereas the alcohol is less likely to do so.

Place a copper coin or copper wire in some vinegar overnight and see how shiny & clean they are next morning.

No scrubbing or friction necessary. Worth a shot on a CPU that has been so problematic thus far.

AND the following from ervus over on MacRumors, just now...

[refinery]

in 20 years since these machines came out i have -never- heard of a processor having a compatibility issue with OS9. We've gotten newer processors to work with OS9 boot, there's no reasonable explanation why some small subset of processors would have an issue... and that would also wreak havoc on Apple support having to maintain two distinct stocks of processors for warranty replacements back then, etc.,  It just makes no sense.
Neither does the RAM issue. 3200 Ram works fine. I have been using 1GB 3200 DDR sticks in my MDD machines for years, never had a problem.
i'd have to guess there is something really obscure that is wrong with that one processor. maybe a ROM version mismatch.

[robespierre]

Or... there is some processor feature(s) used by OS9 that are not used by OSX. Block address translation regs? I guess it's possible they could have developed a defect.

[V.Yakob]

I'll try to sell one of these MDDs, I'll buy another daughtercard: DP 867 MHz. At our flea market, I found only one for little money.

So, I managed to buy a DP 867 MHz daughtercard for only $20 with delivery, and expected Mac OS 9 is working fine.

XPC7455 RX933PC (91L96C, DQR0220), Samsung 226 K7D403671B-HC25, 820-1310-A, 94V-0 242P


To sum up this long topic:

Processors work properly on Mac OS 9 and Mac OS X

1250 MHz x1 -- XC7455B RX1250EF (34L64M, QGG0340D), Samsung 440 K7D403671B-HC30, 820-1497-A, 94V-0 423T
867 MHz x2  -- XPC7455 RX933PC (91L96C, DQR0220), Samsung 226 K7D403671B-HC25, 820-1310-A, 94V-0 242P


The processor works properly in Mac OS X, but Mac OS 9 cannot work properly. Sometimes Mac OS 9 boots (sometimes freezes, sometimes works fine), but in most cases only a gray screen.

1250 MHz x1 -- XC7455B RX1250DF (84L64M, QEU0313C), Samsung 337 K7D403671B-HC30, 820-1497-A, 94V-0 343P

The reasons for this behavior remain unclear, and it is most likely impossible to find out.

Thanks all for help in solving this strange problem! 


P.S.
A problem processor must be put in a frame and hang on the wall. 

[GaryN]

I still suspect the FSB… 

[FBz]

Given that today is Halloween (even if Russians do not celebrate it) perhaps rather than framing and wall-hanging this thing… I’d suggest the use of a propane or acetylene torch to burn it to a complete & total crisp… AND then bury it deep within the earth. (During a full moon?)

OR perhaps... instead:
After burning it sufficiently… then attach the remains to a wooden stick / tree branch, to then be waived (as a talisman) over any other troublesome Mac or particularly troublesome Mac part. This may serve to drive any malevolent (bad ju-ju) spirits from any future problematic machines or machine parts.

A few of us have invested many hours of thought into this thing, so it may now have some other, thus-accumulated symbolic psychic power?

I think the damned thing is cursed.

Boo, y’all.

Big congrats to Vitaly!
Long may his DP 867 MHz MDD serve him well.
(Nice price too!)