Macintosh SE crashing at desktop

19 posts / 0 new
Last post
tony359's picture
Offline
Last seen: 4 months 3 days ago
Joined: Jun 16 2020 - 16:58
Posts: 291
Macintosh SE crashing at desktop

Can I please peek your brain as I have no experience with this platform.

 

I am currently repairing a Macintosh SE which shows a number of issues. Issue 1 was RAM oxidation on one module, I think I sorted that and yesterdat the Mac was booting up happily every time. I only have the mouse, I could browse the desktop.

 

Second issue was the gummed up floppy drive. I fully disassembled it and cleaned (did not disassemble the heads, I basically detached the loading mechanism, washed it in white spirit and then re-lubricated it with lithium grease).

 

When I went to test it

1. The Mac boots to desktop but as soon as I move the mouse the finder menu opens and that's the last thing the Mac does and I have to power it off.

2. I don't have a Mac formatted 800KB floppy unfortunately. But I just wanted to check if the loading mechanism worked. So I disconnected the HDD and let the Mac reach the "question mark" part. At that point if I load an MSDOS disk, the disk loads fine, it spins fine. Then I hear two buzzing noise coming from - I think - the heads. Then the drive ejects the floppy. Is this normal behaviour for "I cannot recognise this disk"? If I manually move the head away from track zero (with the Mac powered off), the head goes back to track zero when it tries to access the disk.

 

Well, the loading mechanism works. The disk loads and unloads perfectly.

 

As the Mac sometimes fails to boot altogether - with the grey screen - I might want to check those RAM modules again but I had a completelty different issue before - picture corruption and "sad mac face" pointing to the oxidised slot. 

 

Any help appreciated!

Thanks

Online
Last seen: 25 min 43 sec ago
Joined: Feb 27 2021 - 18:59
Posts: 608
some points

When the question mark is displayed, the code in ROM is waiting for bootable media to be inserted. If the FDD signals that a disk is inside, it seeks to track 0 and tries to read the boot block. A non-Macintosh-system-disk appears to it to be blank, so the search for a boot block will fail, at which point it ejects so that you may try with another disk. There should be a "Mac with an X" icon displayed to tell you the disk was bad.

I don't entirely understand what you mean by two buzzing noises. The head carriage is moved by its stepper motor but it relies on an optical sensor to tell it where track 0 is. It needs to seek in and then back out again until the photodetector is just barely blocked. This differs from e.g. a Disk II which uses its mechanical head stop to locate track 0: the Disk II always makes a loud buzzing sound when it boots. A Sony 3.5" drive is not silent, but it should be relatively quiet apart from inject/eject. The normal seek sound is "tick, tick, tick".

With the HDD disconnected, you should always get the question mark. That shows that the hardware self-test, including the memory test, was successful. With the HDD connected, you should see the "Welcome to Macintosh" / "Welcome to MacOS" (depending on whether the System is ≥ 7.5), then the INIT icons load across the bottom of the screen, and then see the Finder desktop with a menu bar, the hard disk's icon in the upper right, and possibly other icons or windows. If some of the time with the HDD connected you do not see this, it points to a failing hard disk. The small SCSI drives from that era are mostly all dead now.

I don't know what to make of your mouse issue. It sounds like there could be larger problems with the Apple Desktop Bus (ADB) microcontroller. Note that the keyboard and mouse are not supposed to be plugged or unplugged while power is on; the connector used doesn't prevent static discharge from zapping the electronics.

tony359's picture
Offline
Last seen: 4 months 3 days ago
Joined: Jun 16 2020 - 16:58
Posts: 291
Thank you I will check the

Thank you

 

I will check the "mac with an X" when I put it back together (working on cracked solder joints affecting deflection right now!).

 

I'll take a video to explain better. What happens is that I slide the diskette in, the disk starts spinning. The head does not move and I hear a half-a-second buzzing noise. If I put my finger on components, it seems to be coming from the head. It buzzes twice, then it ejects the disk. If iI try again, the same happens.

 I am doing this with the drive outside of the case, naked. So I can see the head is not moving. But as you say, it does go to track zero when it powers up. But does not seek when a diskette is inserted. 

 

I never hot-plugged the mouse, I suspected that. I'm trying to source another one hoping I can test it.

Does that mouse (ADB II) work on a Snowflake or a QuickSilver? I have those, if the interface is still there I could test it there.

 

When the Desktop appears, the mouse pointer is on top left as usual. As soon as I move the mouse (without clicking it), the menu below opens and at that point the mouse becomes irresponsive.

 

Online
Last seen: 25 min 43 sec ago
Joined: Feb 27 2021 - 18:59
Posts: 608
mouse

It does look like the mouse communication with the computer is somehow scrambled. ADB is a one-wire scheme, all the data in or out uses a single wire. The peripheral (mouse) is probed at boot time and after that it sends status packets with position and button changes. Somehow the packet which should contain only a position change is being treated as if the button was down.

ADB was discontinued shortly after USB was adopted in 1998 (there was one transitional model that supported both, the "Blue and White" G3).

Newer machines don't have ADB ports, but there was a 3rd party ADB-USB converter called the Griffin iMate.

tony359's picture
Offline
Last seen: 4 months 3 days ago
Joined: Jun 16 2020 - 16:58
Posts: 291
Thanks. It's £149 in the UK,

Thanks.

 

It's £149 in the UK, not sure I want to break the bank to test my mouse! But thanks for mentioning.

 

So, in order.

The mouse issue: I will source another one and test. 

 

The Floppy issue.

I've recovered another floppy (Superdrive, it should work as well?) from a battery-destroyed SE/30. It was gunky and horrible but I gave it a go: same issue. The buzzing noise comes from the head stepper. 

The behaviour is somehow erratic. 

 

At first I tested, the diskette stayed in the drive for a couple of seconds then it was ejected with a sad mac with an X on it. So good news. 

But on other tries, the diskette is immediately ejected and no X. When that happens the drive usually emits that "buzzing noise" constantly, whether the drive is in or not. The Superdrive did the same (it needed a little help as the loading mechanism is a bit jammed). Tried swapping the cable and same issue.

 

Please find a video here: https://youtu.be/YxCBSdotFLg

 

Shall I suspect a drive controller issue?

Online
Last seen: 25 min 43 sec ago
Joined: Feb 27 2021 - 18:59
Posts: 608
buzz

Buzzing could be from the stepper motor being stalled because the worm gear is caked with dirt and varnish. If you move the head carriage by hand does the "finger" on the carriage track smoothly in the worm gear?

Lucky that ejection is working, the (Omron) injection-ejection motor has a "sacrificial" gear made of a brittle plastic that usually breaks when the mech is dried up and stuck.

Have you checked any electrolytic capacitors on the FDD PCB? I find that those are a common issue on drives of this era. They may have leaked and distorted the drive signals to the stepper, causing it to buzz.

If it buzzes whether a disk is in or out, that definitely is abnormal. If you disconnect either the stepper or the eject motor it might isolate the location of the buzzing. It could also be one of the microswitches that should signal injection is complete to turn the motor off.

I expect you already know this, but you must be careful not to bend the arms carrying the heads themselves as they are on fairly weak copper spring arms. If overlifted they can get warped and no longer contact a disk.

The high-density "Superdrive 1.4MB" will also function at regular density only (for high-density disks you need to change the ROMs and the floppy controller), but I think there is some issue with the pinout of the cable being changed. There was a ribbon cable with a different colored stripe for it, I think.

OK, I see your video (problem #2). Those are normal sounds. The pitch of the "chirp" changes based on what it thinks is the current track (pseudo-CLV). It's just waiting for the addressed block (h0,t0,s1) to come in and after several rotations it gives up and ejects it.

The mouse problem is more puzzling. Maybe the +5V is being pulled down by the mouse causing erratic behavior?

tony359's picture
Offline
Last seen: 4 months 3 days ago
Joined: Jun 16 2020 - 16:58
Posts: 291
I washed the worm gear with

I washed the worm gear with white spirit (carefully, with a cotton bud and time!) to dissolve the caked grease and applied fresh lithium grease. I can slide it manually using a cotton bud. The head returns to track zero when it starts up.

 

I found that the drive(s) behaves "ok" when the mouse is disconnected... It still makes that weird buzzing noise but at least it loads the drive, it spins for two seconds while making the buzzing noise twice, then it ejects with an X on screen.

I've quickly restored the Superdrive enough to test and it does the same.

Yes, I did disconnect the stepper motor, the buzzing noise definitely comes from it. 

And thanks for mentioning about the heads - thankfully I was aware of that! :) 

 

So, I definitely need to try a new mouse. Unfortunately I do not have a proper disk to do an actual test and I understand I cannot make one with a PC. A Floppy emu is on its way but will take a while to arrive.

 

This is the situation without the mouse connected. I manually slid the head to its farther position so you can see the stepper is working ok

 

https://youtu.be/hSkDPlHk7FM

 

This is a 1.44MB MSDOS disk - so I am not expeting the system to recognise it. Just wondering if this is what is expected from this drive while I try and source another mouse and other things! :) 

 

Thanks for your help so far! It's fun working on this thing!

 

tony359's picture
Offline
Last seen: 4 months 3 days ago
Joined: Jun 16 2020 - 16:58
Posts: 291
Got a bit confused on the

Got a bit confused on the reply - I believe you edited your reply after I added mine, but the forum marked it as "new". Weird.

 

If those noises are ok, then good. I hope I can get the mouse back to desktop so I can try to format the drive. I've also purchased some proper 1MB diskettes.

Mouse, I've just had a look at the schematic. It's simple. I'll scope the signals later. It weird that I tested the Mac on an HDD I got from the destroyed SE/30, it belonged to a large corporation and it started doing an anti-virus check at boot. Took several minutes, no issues. I mean: the system seems to be stable and working, how come the mouse is now causing the floppy to behave erratically?

But without a replacement mouse I won't know. 

Thanks for now!

tony359's picture
Offline
Last seen: 4 months 3 days ago
Joined: Jun 16 2020 - 16:58
Posts: 291
Update. I fired up the scope

Update. I fired up the scope and there is some signal coming out of the mouse and it changes when I move it. 5V is ok. 

 

The Floppy started working as expected even with the mouse on (I removed the isolation transformer from the equation but honestly I don't see how that could be a problem).

 

Still "crashing" on desktop. I noticed that if I push a diskette in the drive BEFORE moving the mouse, I get the "this disk is unreadable" dialog. If I do that AFTER (when it stops working) then no dialog and, most importantly, the drive does not seek. So it does seem that something is crashing indeed. 

 

I need some software, a mouse and a keyboard :)

Online
Last seen: 25 min 43 sec ago
Joined: Feb 27 2021 - 18:59
Posts: 608
debugging

OK you might be getting far enough along to use the debugger. The real debugger (MacsBug) needs to be installed on the boot disk, but you can mess around in the "Microbug monitor" that's built-in to the ROM by pretting the "Interrupt" switch (that's SW1, closest to the rear panel) once you find a keyboard. Actually simply knowing that Microbug is able to draw its window tells you that the system is not completely hosed: it may have branched into hyperspace, but at least the drawing tools still work.

It might even be a corrupt software problem? It would be interesting to see if it behaves booted from a 6.0.8 Startup/Installation Tools floppy.

You can't write 800KB mac disks using standard PC floppy drives, but you can do it with a Central Point card, Catweasel, FluxEngine, etc.

tony359's picture
Offline
Last seen: 4 months 3 days ago
Joined: Jun 16 2020 - 16:58
Posts: 291
I don't remember if I

I don't remember if I mentioned that but I did try the debug button and yes that works when the computer is "frozen". Even though if I press it multiple time the OS mentions an error - which I cannot remember what it was.

I have a Floppy EMU coming soon. I'll wait for it, no point in investing in other hardware right now but thanks for mentioning, nice to know there are options.

 

I was looking into the Dead Mac Scroll yesterday and found some similar issues documented there such as "the mouse button is always pressed". I'll follow that lead, check some components and see if I find anything. I am 100% sure I did not hot-plug that mouse, I wasn't aware it was dangerous but it's also dangerous with a PS/2 mouse. 

 

A software issue is plausible but I think the mouse froze on a different HDD I tested so it really looks like either the mouse has gone or the logic board is unhappy. It's too bad as it took me some time to get to this point and suddenly I cannot input anything anymore! :)

 

Happy Christmas and thanks for your help so far!

 

tony359's picture
Offline
Last seen: 4 months 3 days ago
Joined: Jun 16 2020 - 16:58
Posts: 291
I played with the ADB

I played with the ADB protocol a bit and I really didn't find any smoking guns.

 

The connections are fine. The ADB controller is getting signal from the mouse, clock and it talks to the VIA. When I move the mouse, I see data coming through and the ADB is sending it to the VIA. All discrete components in the way measure ok.

 

I found a manual explaining the protocol so for fun I took a deeper dive in it. I foudn that when the computer boots up it tells device number 2 (default for keyboard) to use the bus. The moment I touch the mouse, the computer - which is the one in charge of the bus, so devices have to ask permission to use the bus) changes the address to device number 3 (default for mouse) and asks it to use the bus. 

 

So, from my limited and ignorant perspective, all works fine and the data is getting deeper into the system. But still no luck.

 

I'll continue when I have another mouse, a keyboard and the Floppy emu :)

Online
Last seen: 25 min 43 sec ago
Joined: Feb 27 2021 - 18:59
Posts: 608
It's £149 in the UK, not sure

It's £149 in the UK, not sure I want to break the bank to test my mouse!

There was also a project called the Waxbee that was aimed at converting ADB graphic tablets to USB hosts. With infrastructure for that, the ADB mouse protocol would be pretty trivial (but that doesn't mean I have done it). It might give you some ideas, anyway.

tony359's picture
Offline
Last seen: 4 months 3 days ago
Joined: Jun 16 2020 - 16:58
Posts: 291
Little update on this issue.I

Little update on this issue.

I've got a new keyboard and mouse, same issue. 

When I try to type on the debug window using the keyboard, each keystroke types "AA" twice (with no hypens). So AA when pushed, AA when released. All keys do the same but Return and "=" from the numpad (I cannot be 100% sure all keys work on this keyboard). 

After about 2/3 of "AAAAAAAAAAAAAAAAA" on the first line, then the typing stops. If I restart the debug window, I can type "AAAAAAAAAAAAAAAAAAAAAAA" again. 

 

The computer boots up and a random HDD I have from another machine runs a full AV scan at boot - that is telling me the buses are fine. Hence I suspect either the ADB chip or the VIA chip. 

Can you please confirm I am NOT supposed to see the time on the main OS desktop? Just thinking that the VIA also does RTC. I do not have a date but I think some pics online seem to confirm that.

 

Any ideas are welcome :)

 

 

 

Online
Last seen: 25 min 43 sec ago
Joined: Feb 27 2021 - 18:59
Posts: 608
etc

System 6 did not have any time or date in the menu bar, although there were extensions (at the time they were still called "INITs") to provide that. If you wanted to keep track of the time, you would open the Alarm Clock desk accessory and move it to wherever you liked on the screen. Pretty similar to using xclock on a unix workstation.

System 7 added a lot of chrome including the Application Menu (task switching menu on the right side of the screen with the current application's name and icon), the Control Strip, time and/or date in the menu bar, the Apple Menu Items folder, and the Help menu and Bubble Help.

Does your mouse still not work? If both the mouse and keyboard are wonky independently it's more likely to be a bad ADB microcontroller in the host. I seem to recall there are also some resistor packs that can go bad and cause mysterious symptoms. Not sure about the VIA—possible it could have a stuck I/O pin or something, but a totally dead VIA will prevent the machine from booting.

tony359's picture
Offline
Last seen: 4 months 3 days ago
Joined: Jun 16 2020 - 16:58
Posts: 291
Thanks. I've got a

Thanks.

 

I've got a replacement keyboard and mouse and they both don't work so definitely something with the logic board.

I've found someone with a similar issue online and it was the VIA. I need to look into that more thoroughly. I suppose that the VIA might work but just the data from the ADB is somehow corrupted. 

I've checked all the line continuity and all checks out ok. 

tony359's picture
Offline
Last seen: 4 months 3 days ago
Joined: Jun 16 2020 - 16:58
Posts: 291
To update: a replacement ADB

To update: a replacement ADB chip (made using a PIC and the code available online) fixed the issue. It looks like the ADB was not initialising, stopping the boot process. 

Online
Last seen: 25 min 43 sec ago
Joined: Feb 27 2021 - 18:59
Posts: 608
great

Great to hear, and useful to know that code exists for programmable PICs. I surmised (from the General Instrument logo on the chip) that it was a PIC but did not know the code was available.

tony359's picture
Offline
Last seen: 4 months 3 days ago
Joined: Jun 16 2020 - 16:58
Posts: 291
For posterity, Code: https:/

For posterity,

 

Code: https://68kmla.org/bb/index.php?threads/mac-se-adb-controller.38363/post-458217

I used a PIC16F87 but a PIC16F88 should work as well.

Programmed using my TL866II and did not have any issues.

The PIC can be sourced in DIP or PLCC so it works on a variety of machines.

 

There is also an adaptor (PLCC to DIP) if you happen to have a PLCC chip around

https://tinkerdifferent.com/threads/macintosh-se-plcc-adb-adapter-board.975/

Log in or register to post comments