Hacker News new | past | comments | ask | show | jobs | submit login
Restoring a Sun SPARCstation IPX Part 3: SCSI2SD, Solaris Install and Expansion (rs-online.com)
75 points by rbanffy 14 days ago | hide | past | favorite | 46 comments

The SPARCstaton SLC (Silly Little Computer) was integrated with the monitor, and was meant to be an "X-Terminal Eliminator".



And like so many things with Sun in the 90's, it was way too late and aimed at the wrong market. X terminals became popular at the very end of the spinning rust era, when maintaining storage was a big part of the IT budget. Drives would die suddenly, they were hard to replace due to all the OS muckery required post-install, they remained very expensive relative to the rest of the system, they were much slower than network in general, and in practice half their capacity was dedicated to storing the same bytes of software anyway. Putting all that complexity into a single system with redundancy and central maintenance was a win.

But it was ephemeral. By the mid 90's, innovation in the PC world had pushed storage prices way down and reliability WAY up. No one wanted an X terminal by 1996, they'd just grab a 486 and put Linux on it. This was the market Sun should have been worried about, and they were oblivious.

I know this, because my very first job in 1996 handed me an X terminal, and I grabbed a 486 out of a closet and put Linux on it instead. The terminal did have a bigger display, though.

It's interesting. Sun didn't have anything to compete at the ultra low-end. They tried to sell Amiga 3000's as entry-level workstations, but Commodore, which was managed by complete imbeciles, wasn't interested.

They had a x86 port of SunOS back in 1988 (to go with the 386I workstation). They could have sold that separately for clone PC's, but alas, history turned out differently.

In fact the 386i was a nice piece of hardware at the time, priced a little under where the Sun-3 68020 machines were placed, with somewhat better performance. But they were launching SPARC at the same moment and this thing just got lost in the shuffle. It was a technically good product, but it was a marketing puzzle that just couldn't be solved.

Only now, many decades later, I realized the Sun x86 based workstations use more or less the same design language and proportions of the 386i.



At one site, the "tech doc" people got the SLC. I felt bad for them because they were running Interleaf and FrameMaker (two nice solid publishing tools with page layout aspects), plus our CAD-ish products... on a smaller screen than all the Software Engineers and EEs had on their adult-sized SPARCstations.

We did experimentally get an IPC or IPX for one engineer. Most all Sun workstations up to that point were pizza boxes, but I recall the first IPC/IPX had arrived one afternoon. And the weird non-pizzabox form factor, and general curiosity about whatever Sun was doing next, was why three of us went back inside from a company social event, to watch while one person set it up (installed more SIMMs, etc.).

Based on my recollection of who got the IPC/IPX, he was a great engineer, who then left software development, and now owns cool pubs in Portland, so maybe he really wanted a SPARCstation 2 instead. (Just a joke; I like the IPX, and when I wanted to personally buy a SPARC for home, I chose an IPX.)

Industrial-design-wise, Sun liked pizza boxes for workstations (they fit exactly under the bases of the huge CRTs, and fairly easy to open and get access to all the components), plus putting peripherals in a few form factors of what we called "shoeboxes" (but no longer shoebox-sized). (The 386i models had shoeboxes that attached atop the PC-ish minitowers.) The various peripherals you see stacked atop the IPX in that post are of the incarnation of shoeboxes that went with with latest pizzabox workstation chassis industrial design (3/80, 4/60, 4/65, 4/75, and maybe others, until the SPARCstation 10 (?) evolved the design a little). So the IPC and IPX looked like Sun was maybe thinking "Hey, we could also put the workstation in this peripheral form factor, and it could open in half". And so it seemed the lunchbox workstation was born.

I wonder whether the IPC and IPX moving the floppy drive (which I think was hardly ever used, when we had networks, CD-ROM, and QIC/Exabyte/DDS) to a more prominent position-- was an accident of fitting things into the internals, or because the peripherals all had their openings on front, or to signal something about the market positioning/differentiation of these workstations.

I'd love having one of these so much... At one time I bought one off eBay, only for me and the seller to discover that shipping it to Brazil, where I was living, was ludicrously expensive.

The seller promised to keep it for me, but that was 16 years ago and we lost contact. I hope it found a nice place.

I still love the aesthetics of early 90s (Desktop) Workstations. Sparc Station 10, Amiga 3000/4000 etc. and this one.

The period on the end of your wikipedia article link is being stripped off by HN's URL parser doodah. Sticking a url encoded period on the end does the trick:


The Apple IIc here somehow looks more futuristic than any present day kit


Hrm. Sorry. OTOH, it's sufficient to ask you "Did you mean: Frog Design Inc.?" and get you there whith one more click!

Incredible, innit?

I saw that as well.

Fixed now with a redirect. It'll probably be removed because of some WP:ARCHAIC_WIKIPEDIA_RULE or some such nonsense.

There was a lot of innovative design in the 90s when computers transitioned from being viewed as industrial-like equipment to friendly, humanist tools. I might just be getting old, but there were some really great designs in the 90s. Even things old PC towers like the Inwin Q500 still look good, despite not being the latest fashion.

They are absolutely beautiful, and the Open Windows environment is still gorgeous. Sadly not very useful today but for remote terminal sessions --and you would need a local web browser.

Wondering if the X server would support running one of the newer Firefox builds as a remote display.

I had an Indy at home for some time back then. Now that is a cool looking computerinabox.

This brings back memories. Fond ones, actually. Sure, the machines were sluggish and limited (e.g. 32MB RAM) by today's standards, but they were still very useful. Somehow, the sun systems didn't bog down terribly with load, so they were pleasant to use, since you knew the limitations. And the hardware was lovely. I particularly remember the keyboards, which were easy to use and very robust.

The video on this website is worth a view, especially for those interested in computing history.

"I particularly remember the keyboards"

I particularly remember "Stop A"

I'm Canadian, so I always took that as "Stop, eh?"

I always found it amusing that Sun systems were lumped in with other Unix vendors and called "proprietary": it used IEEE-1496 SBus, IEEE 1275 OpenBoot, and the SPARC architecture could be licensed by anyone (Fujitsu, Tadpole, etc), and later CPU designs were released under the GPL.

Sun used open standards when they suited their purpose, but they also tried for proprietary lock-in. Their alternatives to X Windows, for instance, had licenses that weren't open in any real sense.

And sometimes they tried to have it both ways -- licensing Java, for example, in ways that seemed to allow for third party adaptations and implementations, and then playing games with access to the Java TCK (the test suite) to try to prevent the Apache group from actually shipping the alternate implementation they'd developed (Apache Harmony) -- see, e.g. https://en.wikipedia.org/wiki/Apache_Harmony#Difficulties_to...

Sun was better than some of the other vendors at generating "open" buzz for themselves, but it wasn't really clear to me at the time, nor is it now, that they were better at actually being open than, say, DEC.

IIRC, OpenBoot was not called OpenBoot, nor was it open, until long after Sun started using it. The same was the case for the SPARC processors; the GPL relase was the classic last-ditch efforts of a failing business. Sun, had they succeeded, would almost certainly have stayed as closed as possible.

IEEE 1275 was published in 1994 AFAICT, which was when Sun was still robust. Sbus/IEEE 1496 was published in 1993 (though the first Sun's started using it in 1989). Before that Sun used VMEbus for their Motorola 68000-based systems.

SPARC was licensed starting in the early 1990s as well:

* https://en.wikipedia.org/wiki/SPARC#SPARC_architecture_licen...

You could buy Tadpole's SPARC portable in 1992:

* http://www.computinghistory.org.uk/det/32324/Tadpole-SPARCbo...

While GPLing was late in the game, non-Sun SPARC OEMs were available for decades. You can still buy SPARC servers from Fujitsu it seems.

So much nostalgia; I bought and used a SPARCstation IPC in the early 90s, running SunOS 4.1.3, for a number of years. It felt like living in the future, so much productivity.

Unfortunately, it was lost in a house fire a few years ago.

They don't have quite the same iconic looking design, but a bargain if you have the Sparc retro bug is the Sun Blade 100/150. They sell on eBay for around $100-$200, and typically have a UltraSparc IIe 500Mhz processor and 256-512MB of RAM. They also support IDE, so no need for $100 of SCSI2SD.

That was my main development machine (Sun had a generous developer discount) until customers stopped asking for Solaris builds.

I have a soft spot for the IPX.

It was a wonderful little form factor with ”good enough” computing power and its TurboGx graphics. It was no SPARCStation 20 or anything but it still had a lot of charm.

For a while I had a similar “tower of power” with as many SCSI peripherals in Sun cases stacked on it as I could manage.

The *NIX workstations of the 1990s from Sun, SGI, and others had such interesting industrial design compared to their boring beige PC contemporaries.

Quasi-off-topic: I might have two SunBlade and one UltraSparc workstations looking for a new home in exchange for a postage contribution (they are based in the UK) and, if you feel so inclined, a contribution either to the british heart foundation or our research group (which is primarily funded by them).

These were used as high-end workstations to drive MRI scanner research consoles back in "the day" – most vendors used sun unix at some point, before slowly transitioning to linux. They all have dual NICs and are well engineered, powerful machines. Drop me an email if interested [see my profile for details].

I'm in Ireland and would be happy to adopt one of them. You can reach me at my nickname here at gmail.

I used one of these for programming for a couple years.

I had no idea they made an LCD screen.

One memory I have of this is hauling one of these plus a massive sun CRT monitor on a hand truck on the NYC subway.

Those SPARCStations are fantastic little machines, I still have an LX, has even a 486 SBUS card, was my main PC mid 90s for years (I run win95 on the card)

Most issues with the machine were : bad NVRAM battery, can be fixed by "hack" the NVRAM chip and failed (caps leaking) PSU, its dead easy to rewire it with a SFF PSU (alas the power button on the keyboard does not work to power-on the machine)

I've got an IPX like the linked article. 64MB RAM, 3 external SCSI drive enclosures. I have some SBUS cards around here for it.

It also has the dead NVRAM battery, but I just take the opportunity to set my MAC to de:ad:c0:ff:ee at every cold boot (which isn't often these days).

I have an older version of OpenBSD on it.

I first came across those bad boys at university. Later on I had several generations of SPARC stations 4, 5 and 20 with 24 bit graphics. They were pretty nice to use and Linux had been ported to this architecture. Bus errors, oh the joy. Last Sun machine I had was an Ultrasparc Blade 2000 but finally left that behind for the amd64 architecture a decade ago. ARM will be next if it's more powerful.

There is a chapter in "Expert C Programming" by Peter van der Linden called "Bus Error, Take the Train", which made me laugh out-loud.

Compared to the PC BIOS, OpenBoot (Open Firmware, IEEE 1275) was quite advanced. Network booting is something that took PCs quite a while to get (PXE), and platform-independent drivers still don't exist.

Apple used the system for quite a while as well.

Mitch Bradley (author of Open Firmware) also developed a version of it for the OLPC:


This brings back memories. I had a SparcStation 5 at home for a few years, in the late 90's. It was one of my favorite machines! Also owned an IPC and an Ultra 10 for a while.

Wow. Memories. That was exactly the first "real" computer i worked on... Writing C code for engravers. 1989 or so??

i have one of these sitting in my basement running netbsd. i recently went to power it up and it appears the PSU has gone bad. anyone know where i could source a new PSU?

You may just need to replace the capacitors: http://www.glitchwrks.com/2017/07/24/ipc-recap

The Sun part number is SUN 300-1055, and the model number is APS-02 if you end up trying to find a replacement. It's a 65W power supply that provides +5V/9A, +12V/1.5A, and -12V/0.1A if you want to frankenstein a cheaper PC power supply in as another comment suggested.

IIRC the voltages can all be produced by a PC's ATX power supply, but the connectors are different. If you can't find an IPC or IPX on ebay, that's probably your best bet.

I would say if you can, replace it with a SFF PSU. The work is a bit "messy" (rewiring the ATX with the Sun connector) but no more cap leaking...

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact