• Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.

IIfx Restoration Issues

KGLlewellyn

Well-known member
Hello there,
I recently picked up a battery damaged Macintosh IIfx and I'm experiencing some strange behavior with it. Here's basically a summation of where things are at:

Current Situation:
  • Will not recognize any SCSI devices except for a BlueSCSI.
  • IIfx will consistently sound the death chimes if the IIfx ROM is installed, but not the ROMinator.
  • When the ROMinator is installed the IIfx will only boot to the ‘?’ floppy if it has not been powered up in the last two minutes.
  • If the IIfx has been powered up again within the last two minutes or is rebooted, it will either chime to a black screen, or sound the death chimes again with a black screen.
  • It can only be booted from the ROM disk in the ROMinator. When this is done it appears to operate normally with no crashes, or other system instability.
  • The IIfx can also mount BlueSCSI drives and browse files and run applications normally when it is booted up from the ROM disk.
  • It will attempt to boot from the BlueSCSI but it will consistently bomb out with a ‘Bus error’ or the ‘Welcome to Macintosh’ splash will repeatedly flash over and over.
What is working:
  • Soft power is working, however the power button will not power down the IIfx. If shutting down within the Mac OS, the IIfx powers off normally.
  • NMI Switch works
  • Sound works.
  • ADB works.
  • PRAM works and settings preserved after power has been disconnected.
  • 128MB RAM installed, all recognized correctly by the machine.
What is not working:
  • Serial is untested
  • Reset switch not working
  • All SCSI drives don’t work except for a BlueSCSI. Cannot boot from BlueSCSI.
  • SCSIProbe reports ‘Bus not terminated’ unless the BlueSCSI is installed. Even if the SCSI filter block is used.
  • RAM chip (UD16 M5M5256BFP-10L) supporting the Serial ASIC (UC16 Apple 344S1021-B) is damaged. However, this should not be impacting the SCSI bus based on schematics that I've gone through.
  • I don’t have a black SCSI terminator, however external SCSI devices aren’t being used.
Work Done.

  • Replaced all SMD electrolytics (only two present), rest are tantalums.
  • Used Deoxit to clean up all dulled contacts on ASICs, RAM contacts, resistors, SCSI connectors etc.
  • Restored three broken PCB traces need the right hand battery.
  • Removed damaged batteries and holders, installed new holders and batteries.
  • Used baking soda to neutralize and remove all battery acid.
I'm thinking of the next steps is to look at replacing the axial and tantalum caps, but I understand these don't typically fail but the odd behavior does appear to be very 'cap-related' in nature. Was also thinking of loading MacBug onto the System 7 install and seeing if I can get an output for the bus error. I've also got a replacement RAM IC for the serial ASIC on the way, but I'm not expecting that to resolve the issue. Failing that, buzzing out the connections from the SCSI bus to the related IC's might also be a next step to take. It definitely seems that SCSI is largely responsible for these headaches right now.

Not sure if anyone here has experienced such behavior with a IIfx, or known the axial's/tants to fail?
 

Johnnya101

Well-known member
I don't think those are capacitor related issues. I have had two IIfxs (I know, not a lot), neither had any tantalum caps go bad, and all the Quadra 700s I've had never had issues either. When those go bad, as I'm sure you know, they explode.

From memory, I think those two caps you replaced really only affect the power circuit, which probably isn't working correctly because of cap goo/bad traces.

Reset switch could be a bad switch or could be cap goo related, I think that's kind of close to one of the two caps? Can you test the reset switch for continuity to verify? Sometimes the switches just go bad.

SCSI issues definitely could be some bad traces. As youve already planned, I'd start buzzing them out.

Do you have a photo of the damage/before photo? How bad are we talking here, a little drop next to a battery holder or Maxell Armageddon?
 

olePigeon

Well-known member
I'd recommend replacing the capacitor at C1 with a higher rated capacitor. C1 really likes to blow up. I'm convinced the default capacitor rating is too low. This is evidenced by all the IIfx boards with blown up C1 capacitors.

Also, I've heard that early-revision IIfxes need a special terminator (either internal or the "black" terminator external) for the SCSI to work reliably. It could be that the BlueSCSI is more tolerant. But I don't know what revision qualifies. Hopefully someone can comment on that.
 

KGLlewellyn

Well-known member
Thanks for all your responses! I've done some further testing. Ran Techtool on the CPU, FPU, SCSI, RAM and VIA all of those pass. I did the drive test and the following issues come up.

  • Self Diagnostic: Fail
  • Test Unit Ready: Not Ready
  • Seek: Fail
  • Rezero Unit: Fail
I ran the same tests on my 7100 which also has a BlueSCSI and only the Self Diagnostic Fails. The BlueSCSI works fine in that machine.

I installed MacsBug on my SSW 7.1 install and when it crashed out when booting it returned the following

"Bus Error at 00082B8C while reading byte from FFFFFFFF in Supervisor Data space"

Entering ‘ES’ will dismiss the crash and the computer will drop immediately to Finder and it’ll work as normal.

I haven’t checked Cap C1 yet, I’ll take a look at that one shortly.

I don’t have a third party terminator to play around with, but I did hook up a RaSCSI to the external connector with no devices mapped and ensured termination was enabled. Didn’t seem to make any difference. The termination power seems to work just fine as the BlueSCSI doesn’t need power over USB or the Berg connector.

I’ve got the schematic printed out and I’ll be buzzing out the traces but if any of you know which ones might be linked to the above errors, that’d be useful.

I’ve attached in some before photos. The batteries blew ‘up’ rather than around the whole board so the damage was isolated near the Serial ASIC and it’s 256k RAM IC. The original RAM modules took a hit but I’ve put new modules in which are working just fine. The traces/contacts around the SCSI ASIC and the OSS ASIC look really good, so I’m quite surprised that this is what’s causing the main issues right now.
 

Attachments

  • Before Damage.jpg
    Before Damage.jpg
    294.6 KB · Views: 60
  • Before Damage 2.jpg
    Before Damage 2.jpg
    442.1 KB · Views: 62
  • MacsBug.jpg
    MacsBug.jpg
    195.5 KB · Views: 56

joshc

Well-known member
How much clean up did you do around the batteries? You mentioned using Deoxit to cleanup corroded legs but did you remove and resolder any components and did you replace any ICs? Some high res post clean-up photos would be good to see.

As for the power button not powering the machine off, have you checked continuity there and referred to the IIfx startup circuit schematics?

 

Crutch

Well-known member
I installed MacsBug on my SSW 7.1 install and when it crashed out when booting it returned the following

"Bus Error at 00082B8C while reading byte from FFFFFFFF in Supervisor Data space"

Entering ‘ES’ will dismiss the crash and the computer will drop immediately to Finder and it’ll work as normal.

This sounds like a software issue that might be unrelated. Like someone is dereferencing a bad pointer. As you can see on the left A1 is indeed all ones which is not good if you are looking at (A1). Is that a clean 7.1 install? Does it boot clean with extensions off?

if you want, type IP into Macsbug and post a screenshot here to see what’s happening.
 

KGLlewellyn

Well-known member
I don’t think it’s a bad install as it booted happily on another machine. Additionally booting up SSW 7.5.3 or 6.0.5 causes a crash earlier in the boot process while it is loading the System file. I haven’t screencapped MacsBug but on 7.5.3 it said dsMemFullError which was odd as it’s got 128MB RAM and is set to 32 bit mode. I can’t ES out of the 7.5.3 crash.

On the 7.1 crash, running stdlog also crashes out when displaying resource information.

I haven’t desoldered the surface mount ASICs or anything like that, and to be honest, I’m a bit afraid to, so I’d need to be convinced there’s an actual necessity of that given the risks of doing so.

I’ve been going through the IIfx startup schematics, this is a work in progress and I haven’t nailed down where the issue is with that. The reset button is a bit low priority for me right now, the bigger issues that I need to fix is SCSI.

I desoldered the SCSI ribbon connector and verified traces underneath. I have also traced all the SCSI Pins back to the ASIC, all are looking good. However, there are some grounding pins (according to the 50 pin SCSI spec) on the top row that don’t seem to be connected, although that appears by design on the IIfx board as I didn’t see any traces coming out of them. Though if you think that a grounding issue could be causing these problems I sure could manually bodge those pins to ground.

Added in some pictures of the post-cleanup. If there’s any specific areas of the board you want closeups from please let me know.
 

Attachments

  • 7.5.3 Crash.jpg
    7.5.3 Crash.jpg
    1.8 MB · Views: 50
  • board.jpg
    board.jpg
    2.5 MB · Views: 42
  • board2.jpg
    board2.jpg
    2.7 MB · Views: 42
  • ipoutput.jpg
    ipoutput.jpg
    1.7 MB · Views: 38
  • scsiheader.jpg
    scsiheader.jpg
    2.2 MB · Views: 39
  • stdlogcrash.jpg
    stdlogcrash.jpg
    1.9 MB · Views: 42

joshc

Well-known member
It looks pretty grotty around the batteries still - noticably this:

1641202962390.png

This is why I'm recommending some surface mount component removal, the larger chips like those are quite easy to work with though you may want to practice on something less valuable than a IIfx.

I can see at least one pin there that isn't connected anymore, in fact it looks like the pin is just gone completely. That chip needs replacing with a new one IMO.

  • IIfx will consistently sound the death chimes if the IIfx ROM is installed, but not the ROMinator.

The ROM SIMM sockets on these are troublesome, the ROM is probably not tight enough in the socket - though I'm surprised the ROMinator works as that's a thinner PCB.

With the amount of issues you are having, I still think I would start with the startup circuit first, check continuity of everything there, and do as much cleanup as possible around the batteries, removing ICs where possible to clean under them and replacing any that can be replaced. The corroded SOIC-28 next to the battery that I've pointed out looks like a part that can still be obtained new, if you check the part number on it (sorry, can't make it out from the photo).

The legs on these look a bit corroded/dull as well?

1641203357434.png

I'm less sure about the SCSI issues to be honest, have you tried it with a normal 50 pin 3.5" SCSI drive instead of a BlueSCSI?
 

KGLlewellyn

Well-known member
Yeah, that chip is a 256k RAM chip that supports the serial pic ASIC. It shouldn't be needed for the rest of the computer to function, though it could be causing a POST to fail. Naturally if you try and do anything that interacts with the serial ports (i.e. select a printer in the Chooser) the computer will just hang, again, that'd be expected. I haven't taken it off in its entirety just in the outside case that it being there is allowing the computer to have some modicum of functionality, probably unlikely though. Anyways, I've got a new IC ordered, it's just a matter of waiting.

Regarding the ROMinator working, all I know is that it bypasses the RAM test and the original ROM does not. I don't know if they removed any other tests that are failing. I.e. if there's a test of the serial ports on the original ROM, but not on the ROMinator then that'd account for the differing behavior.

I'll take a look at those IC's and clean them up.

Re. SCSI, I've tried with an ordinary 50-pin SCSI drive, the one that it came with actually works and I booted my IIsi up from it, but when plugged into the IIfx it will just flash the "?" floppy and appears to not be detected at all.
 

KGLlewellyn

Well-known member
Further update, I desoldered/resoldered UE13 and UD13 and while it didn't resolve the SCSI issues, somehow it does seem to have resolved the button power-off issues. The machine now powers on and off normally without any issues at all. Progress!!

Did some further testing, my running theory about how the BlueSCSI drawing on termination power and thus causing the power to go lower thus allowing it to work seemed to been further supported by when the BlueSCSI is powered via an external source, it no longer works with the IIfx. It does get as far as starting the boot process but it just hangs, adding power to the BlueSCSI after the OS is booted will cause a non-recoverable hang. Also the bus error crash for SSW 7.1 is consistent, keeps happening regardless of extensions being enabled or disabled.

It's giving me cause to think that maybe the termination power on the IIfx is too high and causing some issues there? Definitely seems a bit strange. I should add that even with a BlueSCSI connected internally and a device connected externally, only the BlueSCSI is detected.

Here's UE13 and UD13 removed.
 

Attachments

  • UE13 UD13.jpg
    UE13 UD13.jpg
    2.1 MB · Views: 39

MacKilRoy

Well-known member
The IIfx was the first Mac to handle termination differently. So it wouldn’t surprise me if something that gets power from the termination on the Mac would operate strangely on the IIfx.

Additionally, the IIfx sometimes required special terminators that weren’t needed on other Macs. If your IIfx is one of these models, installing only the BlueSCSI and enabling its termination coils cause you issues I would surmise.

I do have two IIfx machines and both work just fine with standard terminators, but it was such an issue back in the day that Apple made black “IIfx only” terminators.
 

KGLlewellyn

Well-known member
The IIfx was the first Mac to handle termination differently. So it wouldn’t surprise me if something that gets power from the termination on the Mac would operate strangely on the IIfx.

Additionally, the IIfx sometimes required special terminators that weren’t needed on other Macs. If your IIfx is one of these models, installing only the BlueSCSI and enabling its termination coils cause you issues I would surmise.

I do have two IIfx machines and both work just fine with standard terminators, but it was such an issue back in the day that Apple made black “IIfx only” terminators.
Actually, something that gets power from termination is the only thing that works. It's anything that *doesn't* get power from termination which doesn't work.

If termination is disabled on the BlueSCSI it doesn't work at all. I am using the filter block on the 50-pin ribbon and have nothing connected to the external SCSI connector.
 

MacKilRoy

Well-known member
Actually, something that gets power from termination is the only thing that works. It's anything that *doesn't* get power from termination.

If termination is disabled on the BlueSCSI it doesn't work at all. I am using the filter block on the 50-pin ribbon and have nothing connected to the external SCSI connector.
Ok. Interesting.
 

KGLlewellyn

Well-known member
I have found the SSW 7.1 crashes at boot were indeed caused by Serial. I had AppleTalk disabled on a boot and it had no crashes during the startup. Apparently other versions such as SSW 6 and 7.5 do some testing on the serial interfaces whereby 7.1 does not, which also explains why I'm getting different behavior on those.

I would also take bets that the IIfx stock ROM tests the serial ports while the Rominator does not. Anyways, I'm gonna focus my efforts on getting serial working again and I'm hoping that might help clear up some of the wider issues with this machine.
 

KGLlewellyn

Well-known member
Well, the last 24 hours have been a bit grueling. Been cleaning up the board more this week, trying to remove all remaining oxidization. After finishing that the board would immediately produce the chimes of death...no regular chime at all. Anyways, much troubleshooting proved fruitless and I could not get it going at all, so I left it overnight and boom, it's...exactly how it was before. Works if powered off for 2 mins, only works with BlueSCSI.

Regarding serial, I'm pretty close to having it fixed. I've attached a pinout of bad pads/no continuity for the UD16 IC (please excuse the rough drawing). I'm thinking of bodgewiring these straight to the Serial PIC per the schematics. I'm gonna be doing that after my vacation.

I'm dubious that the serial issues would be impacting the SCSI bus, but hey, it's possible, the IIfx is an odd beast. I recently picked up a Granite Active SCSI terminator with diagnostic LEDs. When connecting it via an adapter straight to the IIfx, all LEDs are flashing in the line of drive usage. But if I have an external device connected, the computer will not detect anything on the bus and only the TERM, SEL, REQ LEDs are lit, the ACK LED stays off, and it just hangs in this fashion. Additionally, when hooking up a PowerBook in SCSI disk mode, it gets a bit of activity and again just hangs, so it seems that something is getting lost down the line. Now I'm thinking about it, the ACK LED was pretty faint when it was just the BlueSCSI. If any of you familiar with SCSI and think that this looks to be a faulty ACK line, please let me know!

Another note, I may be able to account for the 7.5 crashing. That might be an unrelated matter due to the fact that I'm using the Rominator, I've seen that the System file for 7.5+ needs to be patched for it, though I'm not sure quite how important that is to do. SSW 6 has been confirmed to me by someone else that it will poll serial no-matter-what during boot, and thus will crash when it can't access it.

Finally, I want to get to the bottom of the random death chiming. It's extremely grating that these happen before video is initialized because all you know is something is wrong, but not *what* is wrong. The solution is Diagnostic Mode and for some reason the documentation is extremely scarce, and actually isn't actively online anymore. It involves hooking a serial cable up to the Modem Port and then using 9600 baud, 8 data bits, 2 stop bits, no parity, should give you some kind of console access to the ROM which will be able to output why it produced the error chords. Naturally with a broken serial bus, I can't use this yet, but I'm hoping it'll help me down the line. I've saved down these instructions, but really we should have this saved elsewhere as I think this info being lost to time is a bit of a risk given that it only exists on the Internet Archive now. https://web.archive.org/web/20210310114138/https://mac68k.info/wiki/display/mac68k/Diagnostic+Mode - Has anyone here used diagnostic mode before? What was your experience?
 

Attachments

  • UD16 Faults.jpg
    UD16 Faults.jpg
    166.9 KB · Views: 37
  • SCSI Disk Mode Hang.jpg
    SCSI Disk Mode Hang.jpg
    477.9 KB · Views: 26
  • Granite - LEDs.jpg
    Granite - LEDs.jpg
    285.3 KB · Views: 41
Last edited:

9166188

Well-known member
Hello!

I also have SCSI issues w/ my IIfx which have baffled me for a long time:- it cannot see any SCSI drives connected internally. Otherwise the machine works fine and I can boot from its external SCSI. This is true to SCSI HD as well as SCSI2SD. Jumper setting SCSI termination does not make any difference. I don't have a BlueSCSI and I wonder whether BlueSCSI can help in my situation or the only solution is to get a IIfx-specific SCSI filter which is very expensive.
 

KGLlewellyn

Well-known member
I think it's definitely worth trying the BlueSCSI as it seems to operate in a way that the IIfx prefers. The kits can be obtained for $25 and fully assembled is $50, so not a huge gamble to see if it improves things, plus they're handy to have anyways.

I also heard from another poster that any active SCSI terminator should suffice for the IIfx. I used a Granite SCSI terminator which I picked up for about $25. Don't waste your money on the black terminator, they ain't worth it imo. This one might also be fine and it's more handy given that it has a DB25 connector on it. https://www.tindie.com/products/gglabs/gterm25-db25-active-scsi-terminator/

I'm gonna be away for two weeks so I've taken a break from working on the IIfx for now. But I'm gonna revisit it with bodging out the wires for the 256k RAM chip for serial when I get back. :)
 

9166188

Well-known member
Got hold of a blueSCSI. But no luck here. It works externally like my HD/SCSI2SD. Nothing happens when connecting thru a 50pin ribbon internally whether the device has termination jumpers set or not. That said, BlueSCSI works nicely externally and I like it. Thanks!

My 2fx has a 1990 board. I guess early versions probably need the SCSI filter to work. Or maybe last attempt would be plugging an external active SCSI terminator -- not sure yet.
 
Top