Macintosh IIfx conundrum

[nathall]

Hey everyone. Been a reader here for a while. I want to get a "second set of eyes" so to speak on my issue as I've basically run out of ideas. I think I posted here about this right after it happened, but for some reason I couldn't find either my account or my old thread and had to sign up again.

Over 20 years ago, around the turn of the century, I acquired a truckload of Macs and Apples (35-45) from the University of Washington when they were "cleaning house" and purging a bunch of old machines. I knew the guy that worked there that was responsible for their disposal. I acquired this literal truckload for $20.

I repaired and flipped the vast majority of them in the intervening years, keeping only the most interesting and best units. 23 years on, I probably only have about 10, but I only keep 3-6 in use/set up at any given time.

One is a IIfx that has always been a workhorse for me. I modified the internals years ago to support 2 internal hard drives in tandem with the 2 1.4MB floppy drives at the expense of blocking a few NuBus slots. That was in 2005 and the configuration of the machine hasn't changed since.

One drive boots to A/UX and one to MacOS System 8.1. About a year and a half ago, while I was using the machine, the screen corrupted in front of my eyes and the unit froze: the bottom half of the display became static, the top half became the entire display compressed vertically into that space and only in green-scale. I reboot, and got a death chime. That was the last time the machine boot, and I'm out of ideas.

Here's what I've tried in those 18 months, that I can remember offhand at this moment:

* originally 32MB of RAM installed in 8 4MB modules, replaced with 1MB of ram in 256k modules I bought since to troubleshoot, and also all other configurations I could think of with this set of sticks
* removed the hard drives
* removed the floppy drives
* originally had an Ethernet card and a Radius video card, removed both
* tried booting without any drives or cards
* tried booting with known good drives pulled from an SE, SE/30, and Classic II respectively
* tried the above plus a video card
* tried booting with all the other NuBus video cards (known-good) that I own, which includes three Apple cards plus the Radius
* replaced the batteries 6 times over
* bought three ROMinator ROMs and tried each of those, as well as the ROM from my SE/30
* gave the MB a thorough alcohol bath (more on this later)
* obtained a brand NOS power supply and tried it
* tried the jump-start method of booting the machine vs the button
* tried pushing and pulling the RAM/ROM in various configurations while trying to boot
* put a terminator on the external SCSI port
* put an external HD with a terminator on the external SCSI port

In every scenario, the machine death chimes. The only things that varied when it death chimed or changed the chime:

* some of the Apple video cards chimed after the gray display popped up, before any disk icon, others chimed on a black screen before anything. The Radius card gets as far as the gray display with the Radius logo in the lower left, then death chimes
* the duration of time between the startup bong and death chime is directly proportional to the amount of RAM installed, the delay is longer with more, so I only assume it's failing at some point after the RAM check
* IIRC, the death chime was different with the SE/30 ROM, which I don't think is compatible anyway but it was before I bought the ROMinators
* death chime is different with invalid configurations of RAM installed (ie, 4MB/256k/4MB/256k etc), again leading me to believe the failure is after the RAM check
* also, there is no scenario in the Dead Mac Scrolls that describes what I am seeing

The logic board is and always has been immaculate. It's a late-run board that only has two electrolytics: C9 and C24, the rest are tantalums. I know C9 is part of the powerup circuit, which still works, so shouldn't be an issue. I haven't been able to determine what circuit C24 is part of, but it appears to have something to do with the RTC. I haven't re-capped either of these, but neither one seems like it could be the cause of a live-before-my-eyes crash and later death chime. As mentioned above, I am experienced at re-capping old Mac boards tens of times over. Either way, the board has been bathed in isopropyl so even if there was leakage from either of these it should have been nullified. I plan to re-cap these out of desperation despite the fact neither one seems like it could possibly be my issue but I wanted to reach out and see if anyone has any other ideas.

I'm heavily leaning towards an IC that randomly and spontaneously failed, but I have never done any kind of IC replacement on these, and wouldn't even know where to start looking, or how. That's always been above my pay grade.

Appreciate any ideas.

 

[nathall]

Oh yeah, also tried endless combinations of ADB keyboards/mice (also none at all) on both ADB ports, no change.

 

[nathall]

Ha, also just remembered that I connected to it from my Mac Plus at some point early on to access the diagnostic mode via serial port. I remember I got in but the only online documentation I could find for this mode was for an earlier (or later?) machine and the hexidecimal commands and responses I was working with did not correspond at all in any meaningful way to said documentation.

 

[volvo242gt]

I would go ahead and replace the electrolytics with a pair of tants, just to rule them out. 10 years ago, when I had my IIfx, before it was recapped, it wouldn't shut down on its own. Are the batteries in good condition, btw? That shouldn't be causing the Sad Mac chimes of death, but...

 

[jmacz]

I am guessing you saw @SophieRose 's thread? There was a link provided by @Phipli in that thread to a very detailed walk through of the boot sequence. From that it looks like it's dying in the low level initialization after the ROM and RAM checks (based on your point that it takes longer depending on the amount of memory you have in it) but before the disk check. Don't know if this helps further but:

Mac GUI :: Macintosh Startup Revealed: From Boot Blocks to Finder

Apple news, photo galleries, downloads, and discussion forums.

macgui.com

 

[Phipli]

Note that it is likely the Radius card is showing the splash screen of its own accord, rather than getting further than the Apple cards.

 

[ArmorAlley]

Post some high-resolutions of the motherboard and see if anyone in the fora notices anything untoward in them.

 

[nathall]

I am guessing you saw @SophieRose 's thread? There was a link provided by @Phipli in that thread to a very detailed walk through of the boot sequence. From that it looks like it's dying in the low level initialization after the ROM and RAM checks (based on your point that it takes longer depending on the amount of memory you have in it) but before the disk check. Don't know if this helps further but:

Mac GUI :: Macintosh Startup Revealed: From Boot Blocks to Finder

Apple news, photo galleries, downloads, and discussion forums.

macgui.com

Yeah I did, thank you. I also reviewed the linked page quite some time ago… I tried to correlate my situation with that page. Unfortunately, because that details the 512k’s boot sequence and not the IIfx, I have to believe there is more the IIfx is doing during boot since I can’t find an exact parallel.

My failure is happening around the “Boot Screen” phase mentioned there. As I indicated, some of my Apple video cards get to the gray screen, some do not. I never get the mouse pointer or the disk/happy/sad Mac icon.

So, what happens at this point in time on a IIfx?

According to that page on a 512:

Beginning at $400450, the following are initialized:

1. Memory Manager, to set the system heap zone
2. SCC, to allow the mouse cursor to move
3. Keyboard
4. I/O and device driver table for 6 ROM-based drivers
5. Interrupts are turned on

If we are assuming the 512 and IIfx boot sequence is identical, which I highly doubt, it must be happening near the end of the above process based on the fact it’s said the screen is still black as these steps occur and only 2 of my Apple video cards get to the gray screen before death. The other does not.

So again making assumptions, what does 4 and/or 5 mean there? What are the “6 ROM-based drivers?” Where do I find the “interrupts” that are turned on, on the logic board?

These are mostly rhetorical questions, primarily used to demonstrate why I’m out of ideas. I know interrupts are logical and not physical, but like I said in my original post, I don’t even know where to go from here. Also why I believe one of the ICs spontaneously failed, but have no idea how to figure out which one.

I’ll take a bunch of photos of the logic board and post them in the next few days.

 

[nathall]

 

[nathall]

So my thought process is, what could cause the screen to corrupt instantaneously in front my eyes like that and also cause the machine to fail boot up checks going forward? First thought would be RAM or VRAM, but it couldn’t be either of those because I’ve tried various RAM configs with sticks that weren’t in the machine at the time, and tried 3 other video cards with their own VRAM.

Bad CPU? FPU? Those are the most easily replaceable components on the board, but at some point just throwing parts at it, I’m going to exceed the cost of just buying a new one. I imagine I’m very close to that threshold already. And my instinct tells me it’s neither of those anyway, based on the fact it POSTs at all.

If you were in my shoes, what would YOU do?

 

[Phipli]

So my thought process is, what could cause the screen to corrupt instantaneously in front my eyes like that and also cause the machine to fail boot up checks going forward? First thought would be RAM or VRAM, but it couldn’t be either of those because I’ve tried various RAM configs with sticks that weren’t in the machine at the time, and tried 3 other video cards with their own VRAM.

Bad CPU? FPU? Those are the most easily replaceable components on the board, but at some point just throwing parts at it, I’m going to exceed the cost of just buying a new one. I imagine I’m very close to that threshold already.

If you were in my shoes, what would YOU do?

Sorry, but just about anything connected to the bus could cause what you saw to happen.

I'd start with an inspection of every chip, check voltages, replace the SMD electrolytics, then clean the board.

After that I'd review the information from the other thread about the IIfx startup sequence that joshc posted.

I'd stick to a basic Apple framebuffer video card to keep things simple.

 

[cheesestraws]

As you've gathered already, the IIfx startup sequence is considerably more complicated than the 512k one .

The grey raster means that the first-stage initialisation code in the video card is running. This is also the same stage that on other graphics cards draws the vendor splash screen, so don't read too much into that.

If you're getting death chimes, it's probably worth seeing whether anything is coming out of the Modem port. See this page for information on the diagnostic mode: https://web.archive.org/web/20140921020853/https://mac68k.info/wiki/display/mac68k/Diagnostic+Mode

If it's not, try powering it up and administering an NMI (using the debug switch) at the very start of startup.

 

[Byrd]

Bad CPU? FPU? Those are the most easily replaceable components on the board, but at some point just throwing parts at it, I’m going to exceed the cost of just buying a new one.

Looking at the pics, I'd also try this next, the 68030 and 68882 look crustier than they should be, possibly stored with a level of moisture nearby. I recently restored a IIfx with the same corroded CPU/FPU appearance, the 68030 when removed had two broken legs from corrosion and was not repairable. Thankfully you can get replacements easily enough. As a start, pull the FPU to see if this helps, but be careful in removing both these as you could pop pins off in the process if all a bit fragile.

I'd also go a bit further with cleaning, using electronic solvent spray the board doesn't look pristine even with an IPA clean (sorry I know I'm not your Mum, but had to be said)

 

[nathall]

As you've gathered already, the IIfx startup sequence is considerably more complicated than the 512k one .

The grey raster means that the first-stage initialisation code in the video card is running. This is also the same stage that on other graphics cards draws the vendor splash screen, so don't read too much into that.

If you're getting death chimes, it's probably worth seeing whether anything is coming out of the Modem port. See this page for information on the diagnostic mode: https://web.archive.org/web/20140921020853/https://mac68k.info/wiki/display/mac68k/Diagnostic+Mode

If it's not, try powering it up and administering an NMI (using the debug switch) at the very start of startup.

yeah, as I mentioned:

Ha, also just remembered that I connected to it from my Mac Plus at some point early on to access the diagnostic mode via serial port. I remember I got in but the only online documentation I could find for this mode was for an earlier (or later?) machine and the hexidecimal commands and responses I was working with did not correspond at all in any meaningful way to said documentation.

 

[nathall]

Sorry, but just about anything connected to the bus could cause what you saw to happen.

I'd start with an inspection of every chip, check voltages, replace the SMD electrolytics, then clean the board.

After that I'd review the information from the other thread about the IIfx startup sequence that joshc posted.

I'd stick to a basic Apple framebuffer video card to keep things simple.

Thank you,

Yes, I plan to replace the two electrolytics though I don't see how either could be causing my problem. If I do and it fixes the issue, I'll eat my words and allow myself to be publicly tarred and feathered

The board has been cleaned, cleaned, cleaned. Isopropyl bath, through the dishwasher, and isopropyl bathed again, aside from more isolated spot cleanings. There is nothing on the board left to clean at this point, frankly. Keep in mind, I am 18 months into this. I didn't pop in here with this because it didn't boot yesterday. I popped in here after 1.5 years and all the basic troubleshooting steps performed and all the more advanced troubleshooting steps performed, with the documentation I can find.

I don't have any sort of reference for what voltages to check, what they should be, or how to inspect every chip, other than visually. I've tried two power supplies, one brand NOS. Both yield identical results, I do not think this is a power supply issue. This is why I'm here-- if there is some documentation on how to "check" every chip, that's exactly what I'm looking for. Every chip looks like it rolled off the factory assembly line yesterday, with the exception of the FPU and CPU whose tops look like they were exposed to some corrosion at some point but honestly have looked like that since I obtained the machine 23 years ago. Their pins look great.

See my next reply for more on this...

(If I'm at the point of trying to determine what IC has gone bad, that's the guidance I'm looking for.... I've never tested individual ICs before to track down an issue like this. HOW do I check each chip, etc?)

 

[David Cook]

yeah, as I mentioned: .... to access the diagnostic mode via serial port

bought three ROMinator ROMs

Install the newer ROM and you'll get the newer diagnostic information and commands.

 

[nathall]

Looking at the pics, I'd also try this next, the 68030 and 68882 look crustier than they should be, possibly stored with a level of moisture nearby. I recently restored a IIfx with the same corroded CPU/FPU appearance, the 68030 when removed had two broken legs from corrosion and was not repairable. Thankfully you can get replacements easily enough. As a start, pull the FPU to see if this helps, but be careful in removing both these as you could pop pins off in the process if all a bit fragile.

I'd also go a bit further with cleaning, using electronic solvent spray the board doesn't look pristine even with an IPA clean (sorry I know I'm not your Mum, but had to be said)

I did pop both of those out and then put them back recently with no change. The pins and underside look considerably better than the tops, basically brand new.

I will try pulling the FPU and booting without it, that's a good idea. Thank you.

 

[nathall]

Install the newer ROM and you'll get the newer diagnostic information and commands.

Thank you, that's awesome. I will try that and report back. I did the diagnostic mode early on in the process, before I obtained the ROMinators.

 

[Phipli]

Be warned that ROMinators mask some issues because they skip the RAM test and skip the ROM checksum.

Secondly, you've asked for very specific advise and then dismissed or ignored everyone's suggestions. You're talking like you have decided what is wrong and want someone to tell you how to fix it - take care. Yes, you've tried most of what people are saying, but we're not as familiar with what you've done and are just casually dipping in and out of a thread, so it is inevitable.

You asked what we'd do. We're all getting up to speed.

Repeating some things means you might catch things you've missed. For example, cleaning.

What I don't understand is why you dismissed cheesestraws comment. As I understand, you tried that before and you didn't do it successfully? This can be improved upon surely?

Ha, also just remembered that I connected to it from my Mac Plus at some point early on to access the diagnostic mode via serial port. I remember I got in but the only online documentation I could find for this mode was for an earlier (or later?) machine and the hexidecimal commands and responses I was working with did not correspond at all in any meaningful way to said documentation.

 

[cheesestraws]

yeah, as I mentioned:

Oops, minus one observation point for me. Sorry!

Though I would personally be asking why it didn't work, and I think @David Cook might have answered that one for us