Differences between revisions 4 and 5
Revision 4 as of 2018-05-15 08:57:08
Size: 6284
Editor: frekk
Comment:
Revision 5 as of 2020-05-26 20:17:16
Size: 8199
Editor: AlexDawson
Comment: Update status and location of PPC95111000
Deletions are marked like this. Additions are marked like this.
Line 6: Line 6:
UCC has two [[https://en.wikipedia.org/wiki/BeBox|BeBoxen]], a revision 6 production model called {{{magenta}}} and a pre-release rev 5 prototype model which has no name and is missing the front panel. {{{Magenta}}} is stored in the cupboards upstairs in the loft and the rev5 BeBox is currently stored on the top shelf behind the door. UCC has two [[https://en.wikipedia.org/wiki/BeBox|BeBoxen]], a revision 6 production model called {{{magenta}}} and a pre-release rev 5 prototype model which has no name (referred to as PPC95111000 through this article) which has never had a front bezel. {{{Magenta}}} is believed to be stored in the clubroom somewhere, while PPC95111000 is currently with AlexDawson for advanced hardware diagnostics and repairs.
Line 14: Line 14:
{{{ {{{[
Line 31: Line 31:
The rev5 prototype BeBox has one of the lowest serial numbers known (PPC95111000) according to a [[http://www.bebox.nu/registry.php|list of known BeBoxen]] (link not working as of 2018). The rev5 prototype BeBox has the lowest serial number known (PPC95111000) according to a [[http://www.bebox.nu/registry.php|list of known BeBoxen]] (link not working as of 2018). If UCC wishes to divest this system in the future, it would be advised to contact the Computer History Museum in Sunnyvale CA for potential donation.
Line 33: Line 33:
Only about 100 prototype BeBoxen were ever produced, handmade by Be Inc. employees and delivered to developers in 1995. They bore a label stating the following: Only about 100 prototype BeBoxen were ever produced, handmade by Be Inc. employees and delivered to developers in 1995. They bear a label stating the following:
Line 42: Line 42:
The revision 5 BeBox was installed a number of times shortly after its donation, however it died completely within hours of being powered on, which was taken to be an indication of a faulty motherboard. The second rev6 BeBox was successfully installed and was operational for a number of years. PPC95111000 was installed a number of times shortly after its donation, however it died completely within hours of being powered on, which was taken to be an indication of a faulty motherboard. The second rev6 BeBox was successfully installed and was operational for a number of years.
Line 44: Line 44:
During 1998 and 1999, inquiries were made by UCC members to repair the motherboard on the club's revision 5 BeBox, since the motherboard seemed cracked or otherwise faulty. In order to effect the repairs, the machine would have had to have been sent to the service center in California. Given its current state, it is unlikely that this ever happened. During 1998 and 1999, inquiries were made by UCC members to repair the motherboard on PPC95111000, as the motherboard is believed to be faulty. In order to effect the repairs, the machine would have had to have been sent to the service center in California. Given its current state, it is unlikely that this ever happened. The board shows traces of rework to the coin cell battery holder, and an undocumented header on the IO Board.
Line 49: Line 49:
{{{Magenta}}} was booted up again in 2012, after much difficulty resulting from one of the RAM modules being loose. A range of different video cards were tried in order to find one that could be used with the prototype machine, but this was unsuccessful. In order to connect the prototype to the network (in the case it was able to boot successfully), a DEC Tulip PCI or RTL8029 (or NE2000 compatible) PCI Ethernet card would have been required. {{{Magenta}}} was booted up again in 2012, after much difficulty resulting from one of the RAM modules being loose. A range of different video cards were tried in order to find one that could be used with the prototype machine, but this was unsuccessful. In order to connect the prototype to the network (in the case it was able to boot successfully), a DEC Tulip PCI or RTL8029 (or NE2000 compatible) PCI Ethernet card would have been required. A replacement AT power switch for {{{magenta}}} was ordered and installed in 2012-04 by ElliotNunn.
Line 54: Line 54:

As of 2020-05-26 (and some time before frankly), PPC95111000 is with AlexDawson, who is attempting a restoration of the system.

Work Performed so far is as follows:

 * The push buttons on the front panel had their "mushrooms" snapped off, and have been replaced with identical ones, with the originals preserved.
 * The Trident TGUI9440 video card that was in the machine, has been removed as it is not on the published Video card HCL for BeOS
 * A S3 Trio64V has been obtained, repaired and tested in a PC and installed.
 * CR2025 in coin cell battery holder has been replaced with a CR2032 battery
 * Cosmetic restoration of FCC warning label
 * General system cosmetic repairs

Parts are pending to repair the original AT PSU, but an AT to ATX converter harness has been used for initial testing. This harness does not provide -5V, but it is unknown if there is a need for that.

At this point, PPC95111000 is still non-functional. With a floppy drive connected, the system should seek for a BIOS recovery at boot, however it does not do so. On inspection under power, CPU0 becomes quite warm, as do some of the FPGAs. On Visual inspection, some of the Programmable Array Logic chips exhibit slight corrosion. A connected keyboard does not react to caps-lock. Once the power supply is repaired, or an alternative AT power supply obtained, the system will be in the most optimal state for booting, and attempts will be made to probe logic with an oscilloscope.
Line 61: Line 76:
 * Systems have a Serial ROM monitor on Serial port 4, 19200,8N1. Press F1 at power on to activate.
Line 66: Line 82:
It is likely that the club will try to sell them in the near future to a computer museum or other interested parties. It is estimated that BeBox units are worth around $1500-$2500 per piece today. A discussion on the mailing list in 2018 came to the conclusion that the club should retain the machines in the near term.
Line 70: Line 86:
Thanks to James Cox, Andrew Williams, Alex Dawson, Andrew Adamson and Nick Bannon for contributing to this documentation of the history of these machines. Thanks to James Cox, Andrew Williams, Alex Dawson, Andrew Adamson, Elliot Nunn and Nick Bannon for contributing to this documentation of the history of these machines.

BeBox (n; plural BeBoxen)
    A Box manufatured by the company Be Inc.

UCC has two BeBoxen, a revision 6 production model called magenta and a pre-release rev 5 prototype model which has no name (referred to as PPC95111000 through this article) which has never had a front bezel. Magenta is believed to be stored in the clubroom somewhere, while PPC95111000 is currently with AlexDawson for advanced hardware diagnostics and repairs.

Be Inc. was started in 1990 by Jean Louis Gasse, an early Apple pioneer. The company became defunct in 2001. For more information about the company, please see Be Inc. (Wikipedia). As of 2018, they are the last company to attempt to enter the home PC market with their own hardware and OS, and are considered an example not to follow.

These machines utilise dual PowerPPC 603 CPUs and were considered far ahead of their time. Be Inc developed their own operating system, intially only for PowerPC but was later ported to other platforms, which was called BeOS.

BeBox-HWBlockDiagW.gif

{{{[ The big problem they had was that they irritated their developers due to the brittle base class problem. They broke ABI a lot – even if your program would compile without changes, it still needed recompilation and redistribution to users. Pre-ubiquitous-fast-internet connections, that was ridiculously impractical. - Grahame Bowland, 1/04/2012 }}}

[...] Those guys were way ahead of their time. Database-based filesystems with indexing and
searching; ubiquitous threading and multi-processor support; metadata mania; the whole works.
It's only in the last few years that we've started to get the sorts of functionality they had
back then.
- Daniel Axtens, 31/03/2012

Pre-UCC history

The two BeBoxen were donated sometime between 1995 and 1999 by past UCC member Peter N. Lewis, who was at the time a Mac developer and interested in trying the Be. As of 2018, he has given permission for UCC to de-acession them.

The rev5 prototype BeBox has the lowest serial number known (PPC95111000) according to a list of known BeBoxen (link not working as of 2018). If UCC wishes to divest this system in the future, it would be advised to contact the Computer History Museum in Sunnyvale CA for potential donation.

Only about 100 prototype BeBoxen were ever produced, handmade by Be Inc. employees and delivered to developers in 1995. They bear a label stating the following:

Attention! This device has not been approved by the Federal Communications Commission. This device is not, and may not be offered for sale or lease, or sold until the approval of the FCC has been obtained.

Only about 1800 BeBox units were ever sold. Magenta is one of the production units that were available to consumers after 1996.

UCC machine history

PPC95111000 was installed a number of times shortly after its donation, however it died completely within hours of being powered on, which was taken to be an indication of a faulty motherboard. The second rev6 BeBox was successfully installed and was operational for a number of years.

During 1998 and 1999, inquiries were made by UCC members to repair the motherboard on PPC95111000, as the motherboard is believed to be faulty. In order to effect the repairs, the machine would have had to have been sent to the service center in California. Given its current state, it is unlikely that this ever happened. The board shows traces of rework to the coin cell battery holder, and an undocumented header on the IO Board.

Magenta was in regular use until about 2003 when it was presumably decommissioned and placed into storage in the clubroom / loft. It was pulled out again in 2007 set up as a display on O-Day. There are some nice photos here, courtesy of Alex Dawson.

Magenta was booted up again in 2012, after much difficulty resulting from one of the RAM modules being loose. A range of different video cards were tried in order to find one that could be used with the prototype machine, but this was unsuccessful. In order to connect the prototype to the network (in the case it was able to boot successfully), a DEC Tulip PCI or RTL8029 (or NE2000 compatible) PCI Ethernet card would have been required. A replacement AT power switch for magenta was ordered and installed in 2012-04 by ElliotNunn.

On 2018-05-14, magenta was successfully booted, but on inspection, the video card was moved to PCI slot 2, due to problems with it booting in slot one. The prototype machine's power supply is mostly dead and probably unsafe, so the power supplies were temporarily swapped, and while the LEDs on the front of the prototype came on, and chips began warming up, it would not boot with magenta's power supply, graphics card and hard drive. Thus as it has been for the last twenty years, it is indeed probably still non-functional. At the conclusion of the hacking evening, the two systems were restored to the previous configuration and both placed in cupboards in the loft.

The machines are able to perform basic networking functions and authentication over LDAP. Due to the fact that BeOS failed to become widely used, there is very little software available that runs on it. At some point, there was a BeOS installer CD in the clubroom.

As of 2020-05-26 (and some time before frankly), PPC95111000 is with AlexDawson, who is attempting a restoration of the system.

Work Performed so far is as follows:

  • The push buttons on the front panel had their "mushrooms" snapped off, and have been replaced with identical ones, with the originals preserved.
  • The Trident TGUI9440 video card that was in the machine, has been removed as it is not on the published Video card HCL for BeOS
  • A S3 Trio64V has been obtained, repaired and tested in a PC and installed.
  • CR2025 in coin cell battery holder has been replaced with a CR2032 battery
  • Cosmetic restoration of FCC warning label
  • General system cosmetic repairs

Parts are pending to repair the original AT PSU, but an AT to ATX converter harness has been used for initial testing. This harness does not provide -5V, but it is unknown if there is a need for that.

At this point, PPC95111000 is still non-functional. With a floppy drive connected, the system should seek for a BIOS recovery at boot, however it does not do so. On inspection under power, CPU0 becomes quite warm, as do some of the FPGAs. On Visual inspection, some of the Programmable Array Logic chips exhibit slight corrosion. A connected keyboard does not react to caps-lock. Once the power supply is repaired, or an alternative AT power supply obtained, the system will be in the most optimal state for booting, and attempts will be made to probe logic with an oscilloscope.

Hardware Notes

Magenta takes a standard AT keyboard and VGA monitor. The twin lines of vertically running LEDs on the front panel give an indication of the usage of the two CPUs. One LED on the right column of magenta is non functional, while two on the left are slightly dim.

  • BeBoxen are notoriously picky about video hardware. They will allegedly only work with the esoteric card they were shipped with.

  • BeBoxen take 72 pin 60ns RAM SIMMs

  • Systems have a Serial ROM monitor on Serial port 4, 19200,8N1. Press F1 at power on to activate.

Future plans for the machine

Since the machines receive attention very infrequently, they are not of great value to the club as long as they continue to gather dust on shelves.

A discussion on the mailing list in 2018 came to the conclusion that the club should retain the machines in the near term.

Thanks to:

Thanks to James Cox, Andrew Williams, Alex Dawson, Andrew Adamson, Elliot Nunn and Nick Bannon for contributing to this documentation of the history of these machines.


CategoryMachines