SE/30 with non-working SCSI/no-boot

[nurrwick]

Hi all,

I'm pretty deep into refurbing an SE/30 that entered my orbit in need of both PSU and logic board recaps. It powers on, plays the chime, runs the memory test while the gray background shows, and gets to the point where it thinks about booting, but will just throw a flashing ? disk status.

I have two bluescsis, internal and external, both work to boot an SE. Internal bluescsi only with an HD0 image, it won't boot. Internal drive installed and external bluescsi connected with an HD2 image (in other words, non-conflicting ID), won't boot.

I've toned out every pin to other destinations on the board and they're all connected. I don't have any shorts between lines. I can tell you that the signaling voltage being delivered to the outputs appears to be 2.7V with dips for data transfers; the chip is receiving a stable 5V. Output lines do appear to be cycling some signaling activity, and that corresponds to the flashing of the Activity light on the internal bluescsi... and thanks to the INI file debug option, I know that activity light flashes once every approximately 1,525ms.

The debug log output starts at Initialization Complete, disables the status LED, reports the Pico's operating voltage, then says "DBG BUS RESET" After that, the following:

[7898ms] DBG -- BUS_BUSY [7898ms] DBG -- BUS_FREE [8151ms] DBG -- BUS_BUSY [8151ms] DBG -- BUS_FREE [8403ms] DBG -- BUS_BUSY [8404ms] DBG -- BUS_FREE [8656ms] DBG -- BUS_BUSY [8656ms] DBG -- BUS_FREE [8909ms] DBG -- BUS_BUSY [8909ms] DBG -- BUS_FREE [9162ms] DBG -- BUS_BUSY [9162ms] DBG -- BUS_FREE [9415ms] DBG -- BUS_BUSY [9415ms] DBG ---- SELECTION: 0 [9416ms] DBG ---- COMMAND: Read6 [9416ms] DBG ------ OUT: 0x08 0x00 0x00 0x00 0x01 0x00 [9416ms] DBG BlueSCSI Toolbox enabled = 1 [9416ms] DBG ------ Read 1x512 starting at 0 [9416ms] DBG ---- DATA_IN [9420ms] DBG ---- Total IN: 512 OUT: 0 CHECKSUM: 31910 [9420ms] DBG ---- STATUS: 0 GOOD [9422ms] DBG ---- MESSAGE_IN [9422ms] DBG ------ IN: 0x00 [9423ms] DBG -- BUS_FREE [9424ms] DBG -- BUS_BUSY [9424ms] DBG ---- SELECTION: 0 [9425ms] DBG ---- COMMAND: Read6 [9425ms] DBG ------ OUT: 0x08 0x00 0x00 0x00 0x01 0x00 [9425ms] DBG ------ Read 1x512 starting at 0 [9425ms] DBG ---- DATA_IN [9426ms] DBG ---- Total IN: 512 OUT: 0 CHECKSUM: 31910 [9426ms] DBG ---- STATUS: 0 GOOD [9428ms] DBG ---- MESSAGE_IN [9429ms] DBG ------ IN: 0x00 [9430ms] DBG -- BUS_FREE [9430ms] DBG -- BUS_BUSY [9430ms] DBG -- BUS_FREE [9682ms] DBG -- BUS_BUSY [9683ms] DBG -- BUS_FREE [9935ms] DBG -- BUS_BUSY [9935ms] DBG -- BUS_FREE [10188ms] DBG -- BUS_BUSY [10188ms] DBG -- BUS_FREE [10441ms] DBG -- BUS_BUSY [10441ms] DBG -- BUS_FREE [10694ms] DBG -- BUS_BUSY [10694ms] DBG -- BUS_FREE [10946ms] DBG -- BUS_BUSY [10946ms] DBG ---- SELECTION: 0 [10948ms] DBG ---- COMMAND: Read6 [10948ms] DBG ------ OUT: 0x08 0x00 0x00 0x00 0x01 0x00 [10948ms] DBG ------ Read 1x512 starting at 0 [10948ms] DBG ---- DATA_IN [10952ms] DBG ---- Total IN: 512 OUT: 0 CHECKSUM: 31910 [10952ms] DBG ---- STATUS: 0 GOOD [10954ms] DBG ---- MESSAGE_IN [10954ms] DBG ------ IN: 0x00

I note that between 9415ms and 9429ms, it goes through two cycles of the selection:0 command: read6 thing, then the block of activity from SELECTION:0 to IN: 0x00 repeats once every second and a half until I power off the machine.

I did have some green corrosion on the SCSI chip's pins that I got removed with some combination of IPA and heating flux on the pins. It's an AM53C80AJC, if that matters. I don't yet have a device I can use to try booting this machine via floppy, and we've been unable to locate any boot floppies in the hoard this machine came from, so I'm at a temporary impasse.

I'm new to this kind of trouble diagnosis so I also don't have a lot of experience with what the log output means. My gut is that the SCSI controller is at fault here, but until the floppy emulator shows up, I thought it might also be worth running the symptoms by the community. Any ideas are appreciated, thanks!

 

[nurrwick]

Hey folks, an update here:

I finally got a floppy emulator device. After confirming it works with my functioning SE, I connected it to the SE/30 and attempted to boot. No dice, with the internal BlueSCSI connected along with the TashTwenty on the external floppy port, it gives up the exact same speed it feels like as with just a SCSI device connected.

I also now have a Das Blinkenlights interposer, so after a couple tries getting it to seat correctly, I did finally get to see evidence that all of my address and data lines survived the cap leakage, and was further able to note that on the lines that are dark while the ? icon is showing, they're flashing on briefly at the same interval as the BlueSCSI log entries.

I did take a moment to run the machine with each of the two sets of 1MB SIMMS in and noted the issue doesn't seem to be affected by the memory installed.

At this point, my guesses are that I either somehow have a bad ROM, or there's something wrong with I guess the glue chip? If the processor had died from corrosion, it feels like nothing would be happening, so that doesn't seem like it.

Any ideas? Thanks!

 

[croissantking]

What floppy image are you trying to boot with? You know the SE/30 with stock ROM doesn’t support HD20, right?

The issue with SCSI almost sounds like a RTC or VIA problem.

 

[nurrwick]

Ah, no I did not know that the machine can't talk to an HD20, and I'm guessing the TashTwenty emulates that based on the name. So in a sense, the emulator not working *is* a ROM issue, at least I can be happy with that assessment.

So since further research indicates the other ROM simms to which I already have access are going to have that same problem (those from a IIcx and a IIci), for the time being I'm back to hunting for a boot floppy or ordering up a contemporary replacement ROM.

Thanks for the pointers to the RTC and VIAs for troubleshooting! It looks like there's at least one datasheet that should help figuring out what to look for with VIA activity, and figuring out if the RTC is misbehaving looks like it should be pretty straightforward based on the schematic labels.

 

[nurrwick]

Oh right, the boot image: I've been working off of the OpenRetroSCSI-6.0.8-500M image located via BlueSCSI setup guides. Normally I'd say "it boots the SE and the two Macs Plus" but I'm starting to figure out that may not mean a whole lot in the grand scheme of things...

 

[nurrwick]

Presumably it'll come as no surprise to more seasoned forum members: our IIcx and IIci, like the IIsi I've already started work on, have vacant ROM slots. It'll be on to checking RTC and VIA health and cleaning the drives in the working SE as next steps for this machine while I work to find floppies.

 

[jmacz]

Check out this thread and see if it helps.

https://68kmla.org/bb/index.php?threads/se-30-yet-another-scsi-issue.49949/

 

[nurrwick]

We found disks and after learning dd needs the bs=84 and skip=1 flags, and a servicing of the SE/30's drive, I was able to get the machine booted off a 6.0.8 disk!

As expected, the BlueSCSIs did not show up on the desktop. I shut down, installed a terminator (per the other thread's recommendation, and by serendipity, it turns out we actually had one of the active Apple terminators!), and booted again with the floppy and was still missing the internal BlueSCSI.

I may be able to try an actual hard drive I can try, though I don't know what software might be on it. If it was a boot drive for the IIsi it is in, should be fine, but it's in there without a sled-bracket so who knows.

I do think I want to take a shot at reflowing the connections on both of the VIAs; I found the connection between VIA1 and the RTC's 1Hz pin to be in the 4 megohm range. I threw a bodge wire on there and it did not fix the SCSI boot issue but then found a broken connection between VIA1 and 2, too.

 

[jmacz]

Did you try the 470 ohm resistor on the /req line like I mentioned in the other thread. I assume your active terminator was on the external scsi port, that didn‘t help me.

 

[nurrwick]

Rereading from your thread, I noted discussion of both internal and external termination, and it's a little unclear to me: did you just build a bridge connection with the 470 from REQ to +5V-TERM and with the caps between +5V-TERM and ground? Or did you modify an internal terminator? If the former, did you do that across the external port with an internal device connected, or on the internal port with an external device connected?

My preference would be to be able to have an internal device boot here but for diagnostics, I'm pretty sure I at least have the resistors in hand (if a 1/4 watt part will work for this) and should be able to bodge up some caps to about that value. I don't have any internal terminators I can modify, though, and only have the one external terminator, so it'd be a procurement exercise to try one of those.

 

[nurrwick]

Welp, turns out what I was witnessing was the result of pins 43 and 44 of the SCSI chip having a barely visible solder bridge. I was going to go in and scope to see if the connections between VIA2 and the SCSI controller were doing anything, and started by looking at the quality of the solder joints everywhere I did the flux-n-heat corrosion cleaning routine. Apparently the iron wicked some solder up and deposited it at the very bottom of the pins. I got that cleared up, re-cleaned a few other spots that weren't shorted together but kind of looked grody, and hooked it up to the slowest SCSI device I could find at work: an original Zip drive, and it sprang to life. Just tried it with the internal BlueSCSI and it's working as well.

Now I need to work up the courage to wash this board like I mean it...

Thanks for the effort, I definitely appreciate the wisdom and am glad this taught me a lot about what these components do and how they're related. I assume this will come in handy with the IIcx!

 

[jmacz]

Congrats!