• Please review our updated Terms and Rules here

Sparcstation 5 on a VME card

Megatron-uk

Experienced Member
Joined
Oct 3, 2011
Messages
122
Location
Durham, Durham, United Kingdom
So this is the first time I've seen one of these - it's a Force SPARC-CPU-5V; according to the manual it's essentially a SS5 on a 1 to 2 slot 6U VME card.

I managed to pick it up cheap from a recent Ebay auction and hit lucky - it came with the optional SBUS frame, which stacks it from a single VME slot to twin, but also adding 2 SBUS slots, both of which where populated; one with a Sun HME fast ethernet/fast-wide SCSI2 card, as well as a TGX framebuffer.

Specs are a 110MHz Microsparc 2, 16MB onboard RAM + 64MB expansion. For all intents and purposes it's a Sparcstation 5.

I intend on bringing it back to life on a NOS Schroff 12 slot backplane I've also just sourced, then using it to drive and control a Transtech Transputer VME card that I've had sitting on my shelf for some 3-4 years. I may even add a 68060 Motorola MVME (running up to date NetBSD) as a slave to the same enclosure at some point.

OS will be Solaris 2.6, since it's the only one with a device driver for the Transputer VME board.
 
I need to find a serial adapter to fit with the micro DB26 connector on the front panel. Anyone seen anything like it before? I don't fancy sourcing an ancient Sun keyboard, sync-on-green monitor and cable just to power it up.

I can't even find a shell of the right type with which I could wire a simple 3 wire RS232 port.
 
I don't think you should need a SOG monitor, looks like you have what is probably a color framebuffer card in there, which should get along with modern displays with a 13W3-to-VGA adapter (note that you'll need a Sun-specific one; SGI used the same connector with a different wiring. But Sun adapters are cheaper and more common anyway.) You can also get PS/2-to-Sun keyboard/mouse adapters for fairly cheap.
 
It is indeed a colour framebuffer - in Sun world it's a cgsix adaptor (8bit, accelerated 2d drawing, up to 1280x1024):

IMG_20181125_165537.jpg

IMG_20181125_165625.jpg

I was hoping to be able to do the setup of both this and a MVME-177 sbc using a serial interface rather than have yet another monitor and keyboard hooked up... but that very high density serial port is proving to be a pain to track down... it actually be easier to use the pin headers on the board itself and build a short breakout cable myself.
 
So I tracked down the part number of the plug that connects to the weird VHDC serial port, it's an expensive TE/AMP part (because of course it would have to be), here are the numbers:

749608-2 - 26 pin shell housing
5750913-2 - 26 pin IO plug

At least with the shell and plug parts I can then connect them to a relatively standard 9 or 25 way RS232 on the other side and then use my standard USB/Serial dongle that I use for everything else.
 
Nice backplane, I helped design it.

The newest ones have switches in the connectors to take care of the IACK and BG daisy-chain signals.
The older ones have gates on the backplane to take care of the IACK and BG daisy-chain signals.
The oldest ones have jumper pins to take care of the IACK and BG daisy-chain signals.
 
Well I couldn't find a cheap VME chassis anywhere; not in the UK and not without paying an eye-watering amount for shipping from overseas.... so I decided to build one myself!

IMG_20181231_171827.jpg

IMG_20181231_174008.jpg

Power comes from a standard ATX PSU (I found one that had at least 1A on the -12v line), and I've cooked up a removable disk tray mechanism to allow the SCSI-to-IDE and SCSI-to-CF drives to be removed easily. Still have a bit of electrical work to do, but it's reasonably compact for a 12 slot chassis and hopefully my wife won't throw it out now :D
 
Nice backplane, I helped design it.

The newest ones have switches in the connectors to take care of the IACK and BG daisy-chain signals.
The older ones have gates on the backplane to take care of the IACK and BG daisy-chain signals.
The oldest ones have jumper pins to take care of the IACK and BG daisy-chain signals.

Hi, I'm hoping that since you have some knowledge of this backplane you'll be able to advise me :)

I'm just about finished the main power wiring for the backplane and now I'm on to the utility connector that is supposed to run to the front panel. In the documentation it is labelled as the following pinout:

SYSFAIL
SYSRESET
ACFAIL
+5v
GND

I'm not using any power monitoring so I'm ignoring ACFAIL. However, would I be correct to assume +5v and GND are actually outputs, rather than inputs? I'm assuming since the use of this connector would be to drive a monitoring device or status panel, they would be providing +5v and ground signals?

Also, I get the function of SYSRESET as a warm reset - though I'm finding it difficult to locate the documentation that indicates whether it needs to be connected to +5v/+12v or to ground to activate.

Lastly, SYSFAIL. I understand this signal can be generated by any of the boards in the system. However, is the signal that then appears on this pin suitable to drive a LED?
 
We make several devices that connect to the utility connector. The simplest device just includes several LEDs that are powered from the +5v and are activated SYSFAIL, SYSRESET, and ACFAIL signals. The more complicated devices monitor fan RPMs and temperatures, and under certain conditions will drive the SYSFAIL, SYSRESET, or ACFAIL signals.

The VME specification says that the SYSFAIL, SYSRESET, and ACFAIL signals all require 48mA driver chips. If you limit the current for the LEDs to less than 48mA then you could connect them directly to those signals. I believe that our display module has a SN7438 to buffer the backplane signals and drive the LEDs.

Switch 7 on your CPU board is determines if the CPU board will drive SYSFAIL, and SYSRESET backplane signals.
Switch 8 determines if your CPU board has the slot 1 functions enabled. It should.
 
Last edited:
Enclosure is nearly finished - some minor cosmetic issues to address (fill in a few blemished and then stain it), but it works and is now hosting the Force SPARC 5V card:

IMG_20190217_155620.jpg

... I'm also happy to see that the card is apparently working fine; I made up a simple 3 wire serial cable that mostly works (some text corruption, so that's one to revisit) and I can see the Openboot monitor start, grab the rarp request, feed it a tftp boot image etc. SCSI to connect up next and see if it's happy to talk to my SCSI to IDE bridge.

Next after that is testing the MVME 68k card and then move on to the additional Transputer VME cards once the Sparc half of the system is running.... as there is a driver to interface the Transputer VME cards with Solaris 2.6. I've got a couple of PPC Transputer cards (in addition to vanilla INMOS T400 and T800 chip cards) that won't fit in anything else that I'd like to try getting up and running.
 
Got things to a point where the Sparc-CPU-5V VME card now netboots. It's a bit of a pain; I got my Linux workstation handing out responses from rarpd, bootparamd and tftp, so it can load a Solaris kernel, acquire an IP address and even start the installer, but no matter what I do I just couldn't get the Solaris kernel to read the exported Solaris install image via the Linux nfs server.

All various incantations of nfsv3 and nfsv4, acl's on, acl's off, everything I tried just resulted in errors as follows:

Code:
ROM Rev. 2.15.1, 80 MB memory installed, Serial #9144285.
Ethernet address 0:80:42:b:17:dd, Host ID: 808b87dd.



Boot device: /iommu/sbus/espdma@5,8400000/esp@5,8800000/sd@1,0  File and args:

Can't open boot device

Type  help  for more information
ok boot net
Boot device: /iommu/sbus/ledma@5,8400010/le@5,8c00000  File and args:
2dc00
hostname: sparky
domainname: (none)
root server: 192.168.1.111
root directory: /export/tftp/solaris/Solaris_2.6/Tools/Boot
RPC: Timed out.
NFS read timed out. Retrying...

I eventually gave up and tried exporting the Solaris installer via a number of different BSD based systems; firstly NetBSD, but that didn't even get as far as Linux, then FreeBSD which worked first time, using the same configuration as I had used in Linux.

No idea what is up there. Anyway, it's now booting reliably over the network:

Code:
SPARC CPU-5V, No Keyboard
ROM Rev. 2.15.1, 80 MB memory installed, Serial #9144285.
Ethernet address 0:80:42:b:17:dd, Host ID: 808b87dd.



Boot device: /iommu/sbus/espdma@5,8400000/esp@5,8800000/sd@1,0  File and args:

Can't open boot device

Type  help  for more information
ok boot net -v - install
Boot device: /iommu/sbus/ledma@5,8400010/le@5,8c00000  File and args: -v - install
2dc00 
hostname: sparky
domainname: (none)
root server: 192.168.1.197
root directory: /export/solaris/Solaris_2.6/Tools/Boot
Size: 243560+176918+41926 Bytes
SunOS Release 5.6 Version Generic_105181-05 [UNIX(R) System V Release 4.0]
Copyright (c) 1983-1997, Sun Microsystems, Inc.
Using default device instance data
vac: enabled
cpu0: FMI,MB86904 (mid 0 impl 0x0 ver 0x4 clock 110 MHz)
mem = 81920K (0x5000000)
avail mem = 77926400
Ethernet address = 0:80:42:b:17:dd
root nexus = SPARC,CPU-5V
iommu0 at root: obio 0x10000000
sbus0 at iommu0: obio 0x10001000
ledma0 at sbus0: SBus slot 5 0x8400010
le0 at ledma0: SBus slot 5 0x8c00000 sparc ipl 6
le0 is /iommu@0,10000000/sbus@0,10001000/ledma@5,8400010/le@5,8c00000
Requesting Internet address for 0:80:42:b:17:dd
Found my IP address: c0a8010f (192.168.1.15)
hostname: sparky
domainname: (none)
root on 192.168.1.197:/export/solaris/Solaris_2.6/Tools/Boot fstype nfsdyn
obio0 at root
zs0 at obio0: obio 0x100000 sparc ipl 12
zs0 is /obio/zs@0,100000
zs1 at obio0: obio 0x0 sparc ipl 12
zs1 is /obio/zs@0,0

....
....
cgsix0 at sbus0: SBus slot 3 0x0 SBus level 5 sparc ipl 9
cgsix0 is /iommu@0,10000000/sbus@0,10001000/cgsix@3,0
cgsix0: screen 1152x900, double buffered, 4M mappable, rev 11
WARNING: /iommu@0,10000000/sbus@0,10001000/SUNW,fas@2,8800000 (fas0):
        Unrecoverable DMA error
WARNING: /iommu@0,10000000/sbus@0,10001000/SUNW,fas@2,8800000 (fas0):
        dma error: current fas state:
fas:    Latched stat=0x11<XZERO,IO> intr=0x10<BUS>
fas:    last msgout: <unknown msg>, last msgin: IDENTIFY
fas:    DMA csr=0xb2b70b13<TWOCYC,BST0,DSBLEDRN,ENBLE,WR,INTEN,ERRPEND,INTPND>
fas:    addr=fc004650 dmacnt=0 test=10 last=fc004650 last_cnt=30
fas:    fas state:
fas:            count(32)=0 cmd=90 stat=13 stat2=80 intr=10
fas:            step=c4 fifoflag=80 conf=17 test=1f conf2=90 conf3=81
fas:    Cmd dump for Target 15 Lun 0:
fas:     cdb=[ 0x12 0x0 0x0 0x0 0x30 0x0 ]
fas:    State=DATA_DONE Last State=DATA
fas:    pkt_state=0x7<CMD,SEL,ARB> pkt_flags=0x808 pkt_statistics=0x1
sbusmem0 at sbus0: SBus slot 0 0x0
sbusmem0 is /iommu@0,10000000/sbus@0,10001000/sbusmem@0,0
sbusmem1 at sbus0: SBus slot 1 0x0
sbusmem1 is /iommu@0,10000000/sbus@0,10001000/sbusmem@1,0
sbusmem2 at sbus0: SBus slot 2 0x0
sbusmem2 is /iommu@0,10000000/sbus@0,10001000/sbusmem@2,0
sbusmem3 at sbus0: SBus slot 3 0x0
sbusmem3 is /iommu@0,10000000/sbus@0,10001000/sbusmem@3,0
sbusmem4 at sbus0: SBus slot 4 0x0
sbusmem4 is /iommu@0,10000000/sbus@0,10001000/sbusmem@4,0
sbusmem5 at sbus0: SBus slot 5 0x0
sbusmem5 is /iommu@0,10000000/sbus@0,10001000/sbusmem@5,0
SUNW,bpp0 at sbus0: SBus slot 5 0xc800000 SBus level 2 sparc ipl 3
SUNW,bpp0 is /iommu@0,10000000/sbus@0,10001000/SUNW,bpp@5,c800000
pseudo-device: winlock0
winlock0 is /pseudo/winlock@0
pseudo-device: lockstat0
lockstat0 is /pseudo/lockstat@0
pseudo-device: llc10
llc10 is /pseudo/llc1@0
...
...
...
<installer starts at this point>

The problem I seem to have, as you can spot from the kernel messages above, is that my SCSI-IDE bridge doesn't appear to be playing nicely with the Sun Fast/Wide controller board. That's a pain in the rear, as that piece of kit (the SCSI-IDE bridge) cost me almost as much as the Sparc board itself.

It may be that I'm running a IDE-CF adapter from it (which are notorious for DMA problems), so I could swap that CF adapter with a spare 2.5" IDE disk relatively easy, and I think that may welll be my next course of action.
 
So a 2.5" IDE drive works much more happily than the CF adapter... however I'm having problems with the Solaris installer - everything proceeds as normal, I can partition the drive, select the software packages ("Complete including development tools") and the installer begins... but it skips over every single package really quickly, much faster than the onboard le0 10Base transceiver should allow :)

The installer then ends with "Could not install all Package contents".

Hmmm.... something going wrong somewhere, but I cannot break out of the isntaller at this point to examine the log files in /tmp/.
 
Starting to think that my Sun Fast Ethernet / Fast-Wide SCSI SBUS card has issues. I was still getting the odd DMA error with the IDE drive rather than the CF adapter, so I now swapped to a SCSI II AztecMonster board (native SCSI II to CF) and I'm still getting DMA errors:

Code:
# format
Searching for disks...done
No disks found!

# Type  'go' to resume
ok probe-scsi-all
/iommu@0,10000000/sbus@0,10001000/SUNW,fas@2,8800000
Target 0 
  Unit 0   Printer     Ap      SDCFH7G016G     H81018180426DC\.X16.
  Unit 3   Device type 7d  Qualifier  38 Memory error during DMA transfer.
DMA control register = b2930b03 , DMA address = ffed5000 
    ATA "DC SDCFH7G016G     H81018180426DC(DX7W.
  Unit 5   Device type 7d  Qualifier  38     A       SV#FHS-01      H81018180
                                             6DCHDX7W.
  Unit 7   Device type 7d  Qualifier  38     AT  "  fQS-FHSm014&     H81018180
                                             6DCHDX7

/iommu@0,10000000/sbus@0,10001000/espdma@5,8400000/esp@5,8800000

ok go

#

Next option is to back off the SBUS card and drop down to the basic SCSI II interface on the CPU board itself. If that works without error (though it will be limited to 10MB/s of course), then I may need to source an alternative SBUS card, or perhaps try cleaning and reseating the card on the SBUS headers.
 
Back
Top