SE/30 start up sequence

[T312-m68k]

Hi all, first post here.

I would love some help chasing down an issue with my new-to-me SE/30.

Problem:
- powers on, chimes, video is good, mouse moves. Stuck at flashing ? disk icon. Can't boot anything, and I can't figure out what to try next.

Work performed:
- got it from ebay, in the same stuck condition it's in now (same boot error)
- LB got an IPA bath for several hours, then few days to dry out.
- recapped the LB with tantalums and axial electrolytics. Removed the battery and bodged in a cr2032 on leads for now
- rebuilt the floppy drive before the eject gear grenaded. When booting, I can see the heads trying to read the disk and then it's ejected.
- it won't boot from the floppy, with several different physical disks and several different "known good" disk images written with dd / ubuntu.
- attempted boot with bluescsi (log file is written to SD card, states the image file is mounted) but no luck. Same flashing ? disk
- the above attempted with stock PROM. Next, attempted boot from ROMINATOR ii PROM, which detects memory ok (1 MB or 4MB depending on which set of four sticks I have installed, all properly seated) but same flashing ? disk when I select either the ROMINATOR ROM or RAM disk to boot from.
- visual inspection of the floppy and scsi cables look good (even if they are disconnected, I can't boot from ROMINATOR so that rules them out anyway)

Testing:
- replaced SCSI chip UI12 with new stock Zilog part from Digikey; no change (so the old one was probably ok)
- 5v on the SCSI cable termination pin 26 tests at 4.85 V; D3 is ok and F3 and F2 are ok.
- 5v on the floppy cable looks good (5v)


I'm not sure where to look next. I'm comfortable with a scope and following schematics, but decoding this for (my) first time is a bunch of work. I'm hoping someone who knows how the boot process works in detail can point me in the right direction of what to check next.

Thanks in advance, and I'm excited to be a part of this community!

 

[nxstudiomc]

Hello!
Did you boot the computer with the original HDD installed? If yes then unplug the SCSI cable on the drive side, that's how I got mine to work

Nick

 

[T312-m68k]

Did you boot the computer with the original HDD installed?

This SE/30 and the included hard disk was not booting when I got ahold of it. My scsi cable has only one drive connector, so the BlueSCSI replaced the original hard drive. I agree that it's good to rule out that I didn't mess up termination Thanks for your comment!

 

[halkyardo]

Congratulations on finding yourself a project! Sounds like you've done some solid troubleshooting so far. Take all of what I'm about to say with a grain of salt since I've never experienced this particular problem before, but being decently familiar with the SE/30 hardware and having poked around in the ROM a bit, I have a few thoughts:

As I understand it, the SE/30 should always attempt to boot from floppy before trying the hard disk. If it's possible to prove that the floppy drive is good (say, by testing it in another Mac), then I'd recommend taking SCSI out of the equation and just focusing on getting it to boot from a floppy disk. Also note that the V1 BlueSCSI hardware does not comply with SCSI's electrical specifications. Plenty of people (including myself) use them in old Macs just fine but issues aren't unheard of - if you've got one and are having SCSI issues, I'd recommend replacing it with a ZuluSCSI or V2 BlueSCSI just to rule out that possibility.

Either way, if it's getting as far as it is, we can be fairly sure that most of the system is healthy. It's reading from ROM, passing the RAM test, and the video hardware and GLUE chip are probably OK. Since the mouse moves, we can establish that the ADB chip and its connection to the processor are also healthy.

If it's spinning the motor up and moving the head when it attempts to read a floppy disk (and ejecting it afterwards), that suggests that the address lines to the SWIM are intact and doing the right things, since all of these functions are controlled by addressing only, with the data lines mostly disregarded (the SWIM and the Mac floppy interface are weird). However, reading the disk does involve the data lines, so if you've got, say, a broken trace on one or more of the SWIM's data pins, behavior like you're seeing would be a possible symptom.

The SCSI and SWIM chips (and most of the peripheral chips on that rear edge of the board) share a common 8 bit 'leg' of the data bus that runs right next to C8-C10. I'd suggest taking a close look at those traces, and testing the continuity from the data pins on the SWIM chip (and the data AND address pins on the SCSI chip), back to the 68030. Note that due to the way that 8-bit accesses work on the 68k, pins D7..0 on the SCSI and SWIM chips are connected to D31..24 on the 68030, NOT D7..D0. It's easy to forget this!

If you haven't come across them already, the 'remastered' SE/30 schematics will come in handy for this: https://github.com/mishimasensei/macse30mlb

Also, while not strictly necessary, Apple's internal documentation for the SWIM is interesting too: https://bitsavers.org/pdf/apple/disk/sony/SWIM_Chip_Users_Ref_198801.pdf

EDIT: it just occurred to me that I’ve got no idea how the “ROMdisc” boot mechanism fits into all of this, so I may be entirely barking up the wrong tree. Though are you sure your Rominator has a ROMdisc boot image in it? I don’t have one so I’m not very familiar with them, but I’m pretty sure the ROMdisc bit is optional.

 

[dochilli]

Perhaps a stupid question: Are you sure that you have a 1.4MB floppy drive? I own an SE/30 that had a 800KB drive, because it was an upgraded SE. What color has the cable of the drive? Grey/red or grey/yellow? What is the type number of the drive?
Did you clean the reading head of the floppy drive? Do you have a spare floppy drive, so you can change the drives?

 

[halkyardo]

^ yeah, good point. And while we're on the topic of floppy drives, make sure that the disk is properly seating on the spindle when it's inserted. When I cleaned the drive in my IIfx, I somehow managed to mess something up that caused the loading mechanism to hang up partway. It would accept and eject disks just fine, but wouldn't work reliably. Didn't figure it out until I was checking continuity of the sense switches and saw that the disk wasn't depressing them fully.

Also on that note, a quick way to tell the difference between an 800K and a 1.44MB drive is to look at the sense switches inside the disk opening - 800K drives have one on each side, 1.44MB drives have two on one side, one on the other.

 

[T312-m68k]

Hi folks,

First, thanks so much for your replies. I'm learning a lot from what you suggested.

@dochilli : I do have a 1.4 floppy; it's a Sony MP-F75-01G; the cable is gray with a red stripe, and 2 sense switches on the one side. I cleaned the reading head with IPA and a swab, when I had it the floppy drive apart. Visually it wasn't in bad shape beforehand, so hopefully this is sufficient. Unfortunately this is my only vintage apple computer at the moment, but I'm seriously considering pickup up a local SE this week, because it will have another superdrive, another keyboard, and mouse, RAM, etc that will all be useful in troubleshooting this SE/30 and others in the future (because believe it or not, troubleshooting this is fun!)

@halkyardo : I love the suggestion about leaving SCSI out of the picture for now, and focusing on the floppy. That's the kind of advice I needed to hear. I became distracted by the idea that if I couldn't boot off the floppy, perhaps I could off the hard drive. And that's also how I ended up trying to boot off the PROM (Rominator). None of these worked, but they're separate sub-systems, so I'll revert to one at a time. Floppy first.

Next, I buzzed out the traces between the SWIM chip and J8 to make sure that's all good (they are).
At your suggestion, I also checked the addressing and digital bus lines between the SWIM and SCSI chips and the 68030. Everything tests ok. Additionally, visual inspection of the traces near C8,9,10 looks ok.

So the hardware checks out so far.

...

Some weirdness that might be nothing, but might be something:

1. when booting a floppy, I found one image that will give me happy mac for ~2-3 seconds and then gets ejected. Everything else I have tried does not work this way, and is immediately ejected. Both conditions are repeatable.

2. separately, it appears my ABD keyboard is bad (!?) and here's why I suspect this: the ROMinator PROM wants you to press "R" to enter romdisk, and when you do, it accepts the keypress immediately (instead of waiting for a fixed time). I learned this from watching others do it on YouTube. On my system, when I hit "R" to enter romdisk, nothing happens immediately. Instead I get the flashing floppy question mark after the usual amount of time. I'm also unable to eject the floppy using apple-shift-1 or reset the PRAM with the keystroke. I can't prove that the keyboard works, so I have to assume it doesn't.
The weird part: the ADB mouse which is fed via the passthru ADB port on my M0487 keyboard works fine. So either the keyboard fault is on the circuit that goes to the keys themselves, or something is wrong decoding ADB. Now the question: Is there a failure mode of ADB that would cause a floppy boot (that would otherwise work) to fail at the stage where System is trying to load ADB keyboard drivers? Wondering if this is what causes my crash after 2-3 seconds when floppy booting.


As an aside, @halkyardo mentioned they have no idea how the PROM boot fits into the equation; I don't either. I assume it relies on the SCSI chip and not the SWIM chip, but perhaps both are used. I couldn't find any good documentation about how it actually works.

 

[T312-m68k]

I went back and forth on whether to post my trace notes, but ultimately decided they could be helpful in understanding what I have checked. So here they are.

 

[T312-m68k]

I have made some progress. It boiled down to checking assumptions.

First, the (only) ADB keyboard I have was not working. I took it apart and discovered that the keyboard ribbon cable was not connected. Maybe that's why the seller included it. Sigh. Once this was reconnected, the keyboard worked.

Once I had a working keyboard I could boot into the ROMinator ii image of SSW 7.1. Hooray!

[ As a side note, the ROMinator does not appear to use a SCSI id. It must boot as an extension of the PROM or some other magical way. ]

Second, I ran scsiprobe to see what the computer sees on the bus. Shows it's not terminated, even though bluescsi is connected and configured to terminate.

I started checking voltages. Discovered a 5v short to ground but only when the BlueSCSI is connected. Discovered the d@mn thing was assembled with the scsi notch upside down, which obviously won't work. Sigh #2.

I cut a new notch on the BlueSCSI connector, flipped the cable over, and now the mac detects that disk too. Hooray!

I'm further along, but not finished yet.

I still can't boot off the BlueSCSI; it hangs at "welcome to macintosh".

But it's progress, and I wanted to update the thread. I'll keep working and see how it goes. TechTool is complaining I don't have enough free memory to launch it (4mb installed), which gives me the excuse to spring for some new SIMMs

 

[Kouzui]

[ As a side note, the ROMinator does not appear to use a SCSI id. It must boot as an extension of the PROM or some other magical way. ]

The image is on the storage contained within the (extended) memory of the rom itself.

I still can't boot off the BlueSCSI; it hangs at "welcome to macintosh".

What images do you have on the BlueSCSI? If you only have 4mb installed, that might not be enough to actually load the OS.

 

[T312-m68k]

I'm now able to boot off the BlueSCSI. Resolved with software (different hda image). Hooray, I have a usable SE/30!

Next is physical restoration and figuring out why the floppy won't format successfully. I'll start a new thread for that if needed.

Sending my thanks to everyone who helped me with ideas and information!