Exception when booting NeXTstation Color

Hi there, first time member and post!

I'm a vintage hobbyist with a particular affinity to Unix workstations from the 80's and 90's (Sun, HP, SGI, IBM RS/6000, ...), probably because I was a college student who could only dream about having enough money to actually have one at the time, but anyway...

After quite a bit of patience and $$ I've finally been able to piece together a NeXTstation Color, which has always been on my bucket list.  My system is pretty generic (see picture).

NeXTstation Color (bought from some guy in Oklahoma on eBay)
26MB RAM (6x4MB + 1x2MB)
ZuluSCSI emulating SCSI ID 0 2GB HD, or SCSI ID 6 for CDs
Non-ADB Soundbox (bought from Rob at Black Hole)
Non-ADB Keyboard (eBay from somebody in Canada)
Non-ADB Mouse (eBay from somebody in Florida)
Y-Cable (13w3, DB19, VGA) from Rob at Black Hole

After getting the system assembled and the NVRAM configured it booted great from the HD image I downloaded from winworldpc.  I was having a blast, installing various applications, marveling at the UI, etc etc.  It was working GREAT for about two weeks, then out of nowhere, without any hardware or configuration changes I'm now getting a variety of Exceptions when it tries to boot from SCSI.  The sequence:

1) Turn on power via keyboard
2) Display shows "Testing System"
3) Soundbox beeps (I turned that setting on in the config)
4) Display changes to "Loading OS" (I set boot device to sd in config), the graphic of the hard drive spins very very briefly...
5) Exception #29 (x74)
or
5) Exception #3 (0xB)
or
5) Exception #2 (0x8)
or
5) The screen just starts scrolling with non-stop "ipl5: Spurious Interrupt"

Things I've tried:

1) Removed 3v battery (that does indeed test at 3v) to clear config, re-inserted, re-configed (command-~) - SAME EXCEPTION
2) Tried all different combination of SIMMs to eliminate a bad SIMM, tried just 8M (one bank), 16M, 24M, etc.  SAME EXCEPTION so I don't think it's memory related.
3) Tried unplugging ZuluSCSI from SCSI bus and booting from NS 3.3 install floppy.  First I had to terminate the SCSI bus with an external terminator otherwise I would get a SCSI error during system test.  The floppy booted and got to the "Searching for CD-ROM drives" message but couldn't get further obviously because I had the ZuluSCSI detached.
4) Unplugged both floppy and ZuluSCSI and tried booting from network.  This worked as well, or at least got to the point of trying to find a BOOTP server.
5) Turned on extended self-test in config, all tests pass.
6) Booted from floppy with ZuluSCSI installed with NeXTstep 3.3 User CD emulated as SCSI 6.  The system booted, then message "Searching for CD-ROM drives", then the LED on the ZuluSCSI blinks for a bit, then the endless "spurious interrupts" message. (see picture)

So I think i've eliminated a bad config, bad SIMMs, bad floppy, and bad ethernet.  But the SCSI now seems suspect, or the system board.  But before I start replacing capacitors I wanted to ask the experts on this forum if they've seen these exceptions and words of wisdom at further tests I can perform.

Thank you!

---

Level 5 interrupt (ipl5) is connected to the serial port. So maybe this is the first place to look for an issue. But it might as well be some bad capacitor or something completely different. I also don't think this is caused by bad memory.

Hello: Welcome to the forums :

Is your SCSI Iso image clean as it may have been corrupted, if it ran check disk or fsck ... what size sd card and partition ?

A ram stick going bad or failing cap would cause the interupt and possibly corrupt the master image on your iso.

I do have known good 4MB simms for sale $2 a 1 Mb:) 32Mb = $64

The new caps usually clean up exception errors , they survived 30 plus years :)

From a Doc in the archives, my guess is a bad cap and or probably bad ram , it sounds like when the processor is pinging ie ram , if the timings are off then it would kick into a spurious interupt.

Have you tried a fresh image on the ZULU SCSI Card?

When yo have the boot floppy and NeXTstep 3.3 User ISO do you have a 3rd partition  on your ZULU sd HD0, HD1 for the install to go to?

QUESTION
What is the interrupt latency for the MC68000 family?
ANSWER
Interrupt latency is the time from when an interrupt occurs to when the
first instruction of the interrupt handler is started. The amount of time
required for this will depend the processor, what the processor is doing
when the interrupt occurs, and the speed with which the memory responds to
reads and writes.
Worst case interrupt latency is the absolute maximum amount of time it will
take for the processor to respond to the interrupt for a given set of
conditions. A given interrupt may respond sooner (and usually will), but it
will never take longer. The selection of the conditions can have a major
impact on the interrupt latency. For example, the number of wait states on
writes will effect the time it takes to stack the interrupt stack frame. An
infinite number of wait states on writes would result in an infinite
interrupt latency.

Quote from: thatsmanjear on March 11, 2024, 04:40:00 PMSo I think i've eliminated a bad config, bad SIMMs, bad floppy, and bad ethernet.  But the SCSI now seems suspect, or the system board.  But before I start replacing capacitors I wanted to ask the experts on this forum if they've seen these exceptions and words of wisdom at further tests I can perform.

Since you are using a ZuluSCSI, and the suspicion is that the issues might be related to that, I would recommend to check out the following video where somebody has documented the Process on how to set up a ZuluSCSI for use with NeXT systems:
https://www.youtube.com/watch?v=STaCv_yyF6Q
Based on another thread on the forum related to the BlueSCSIv2 (which is quite similar to the ZuluSCSI) the NeXT systems seem to require specific settings for the "SectorsPerTrack" and "HeadsPerCylinder" parameters in the .ini file, which is also mentioned in the video above: https://www.nextcomputers.org/forums/index.php?topic=5498

Quote from: don_apple on March 12, 2024, 07:07:19 AMSince you are using a ZuluSCSI, and the suspicion is that the issues might be related to that, I would recommend to check out the following video where somebody has documented the Process on how to set up a ZuluSCSI for use with NeXT systems:
https://www.youtube.com/watch?v=STaCv_yyF6Q
Based on another thread on the forum related to the BlueSCSIv2 (which is quite similar to the ZuluSCSI) the NeXT systems seem to require specific settings for the "SectorsPerTrack" and "HeadsPerCylinder" parameters in the .ini file, which is also mentioned in the video above: https://www.nextcomputers.org/forums/index.php?topic=5498

139 Sectors
4 Heads per cylinder

SCSI2SD XML

If your SD is formated Fat 32 with Master Boot Record

You should be able to drop 2Gb HD1.hda images into it then use it to boot and or install NeXTstep

Hi, thanks for the comments.  As I mentioned in my original post my ZuluSCSI based setup was working great for a few weeks, it booted fine, ran fine, I could install apps, etc, so my ZuluSCSI itself, SD card, and .img and .iso files worked just fine.  Since this exception error suddenly started (after making no changes whatsoever) I tried an alternate ZuluSCSI to rule that out, a different SD card to rule that out, used original fresh images on the card, heck, even used a different SCSI cable, but still same problem.  So I'm pretty convinced my ZuluSCSI isn't part of the problem.

Since my original post I also tried fresh 4MB SIMMs, which pass POST testing but I still get the exception, so I'm also convinced SIMMs aren't my problem.

I'm going to start replacing caps.  Anybody have insight into what caps are near the circuitry for SCSI and serial.  I think I'd like to start there, replace one at a time and test booting after each one.  I would say my desolder/solder skills are 7 out of 10 :)

Quote from: thatsmanjear on March 12, 2024, 05:28:10 PMI'm going to start replacing caps.  Anybody have insight into what caps are near the circuitry for SCSI and serial.  I think I'd like to start there, replace one at a time and test booting after each one.  I would say my desolder/solder skills are 7 out of 10 :)

Not sure the location on the Color Slab of caps near SCSI and Serial, but the number of caps on the board is at most 15 in total, mostly the surface mount 47µF sort.  The rest are 220µF radial caps.  I'd be of the thought, just be one and done.  Recap everything all in one go.  I can understand though the idea of wanting to isolate where the issue actually is, but sitting on all good caps *could* once and for all fix all the problems you're having. 

One of the folks in the forums put together a location map and parts list that you'd need to complete the job.  Parts total, I think were less than $10.  http://www.asterontech.com/Asterontech/nextstation_mono_retore.htm

Quote from: cooltr6 on March 12, 2024, 06:14:28 PMI'd be of the thought, just be one and done.  Recap everything all in one go.  I can understand though the idea of wanting to isolate where the issue actually is, but sitting on all good caps *could* once and for all fix all the problems you're having. 

Thanks for the link, very helpful.  Yeah, I agree on the "one and done" approach, I just don't have the confidence in my solder skills not to irreversibly mess up my board.  The radial caps, easy.  SMDs, not my strength.  Wish me luck!

Quote from: thatsmanjear on March 12, 2024, 10:43:45 PMThanks for the link, very helpful.  Yeah, I agree on the "one and done" approach, I just don't have the confidence in my solder skills not to irreversibly mess up my board.  The radial caps, easy.  SMDs, not my strength.  Wish me luck!

I recapped a couple of 25MHz cube motherboards a while back and used an inexpensive tweezer style tool to desolder the SMD caps. That plus a good flux made things a lot easier for me.

https://www.nextcomputers.org/forums/index.php?topic=5047 (https://www.nextcomputers.org/forums/index.php?topic=5047)

Good luck and let us know how it goes.

Quote from: thatsmanjear on March 12, 2024, 10:43:45 PMThanks for the link, very helpful.  Yeah, I agree on the "one and done" approach, I just don't have the confidence in my solder skills not to irreversibly mess up my board.  The radial caps, easy.  SMDs, not my strength.  Wish me luck!

You're most welcome, my friend.  Trust me, I completely understand your feelings.  I was in exactly the same boat as you - with what could only be called sub-par soldering skills.  First board I did, was a NeXTstation Mono.  Two traces rotted...managed to get those fixed.  Got all the SMD's off...got the radials out...reassembled.  The most delightful thing...was it sprang to life when put back together.

Echoing the words of Nitro...flux...and in the words of Brian, who put together the cap guide, "Take your time and use LOTS of flux.".  It was recommended you use a tweezer-style desoldering tool, wish I had one when I did my job...I used a nice chisel-tipped iron at 350°C...one side at a time, carefully working things away from the board and using tweezers and a magnifying light to help guide the rest.

Wishing you the best of luck in this endeavor!  Trust me, when it's all said and done...the stresses of it all will melt faster than snow in summer.

Well, darn.  After sending $20 to my friends at Mouser, and then spending a stressful hour desolder-deoxit-flux-soldering six new radial and seven new smd caps onto my system board, it booted, passed all system tests, then gave me the same exact "spurious interrupt" messages when trying to boot...

In the meantime I also double checked the voltage levels from the power supply just to cross that off the list.

Any further suggestions, or is my system board just needing replaced altogether?

Just a shot in the dark, but try the parity setting both ways in the PROM.