Fuel prototype (BLUE!)

[jan-jaap]

Having to deal with an old prototype for a business project is a "no-go" for me since there are too probability of being stopped by a dead/defective hardware or simply a device whose behavior might be unexpected. [...] So, what was/is the goal?

The goal is to preserve this system in a functional state.

I'm well aware of what a prototype is. I do software development for a hardware development company so my desk is usually full of proto or modified hardware. This system was under my desk at work for ~ 9 years, and most of that time it was used for software development. It served me well.

The problems with the Dallas chips are well documented. This system had some I2C related problem where the L1 couldn't "see" the CPU anymore, so it refused to power up. It sat there for two years maybe until the company stopped IRIX support entirely and donated the (dysfunctional) system to me. I replaced the CPU, but when I tried to power it up it gave me more DS1780 grief:

Code: Select all

ALERT: Error reading the NODE 1 temperature sensor, no acknowledge
ALERT: Error reading monitor NODE 1 interrupt status 1: no acknowledge
ALERT: Error reading monitor NODE 1 interrupt status 1: no acknowledge
ALERT: Error reading the NODE 1 temperature sensor, no acknowledge
ALERT: Error reading monitor NODE 1 interrupt status 1: no acknowledge
ALERT: Error reading monitor NODE 1 interrupt status 1: no acknowledge
ALERT: Error reading the NODE 1 temperature sensor, no acknowledge
ALERT: Error reading monitor NODE 1 interrupt status 1: no acknowledge
ALERT: Error reading the NODE 1 temperature sensor, no acknowledge

I have some of these chips but the pitch of the pins is really at the limit of what I can do by hand, and there's one between the PCI slots that's even harder to rework. Maybe I'll ask the company that assembles our products for a favor. Instead, I put in the "NextGen" mainboard and now it works again -- with the supposedly non-functional CPU.

[mapesdhs]

So, what was/is the goal?

Presumably the challenge of getting things working again, among other things. Boy, you really are a glass-half-empty kinda guy...

[Y888099]

Boy, you really are a glass-half-empty kinda guy...

Well, in my working experience we do a lot of prototypes, and we usually know that or we work on those made by us, or we for sure have a lot of troubles with incoming prototypes made by others, and since we have deadlines to respect, for us it's not so smart to allow more entropy could income into our laboratories.

We can of course handle it, but we can't respect deadlines as it's usually the source of more problems, so ... it looks silly for me. We are also speaking about SGI, a dead company. What if something goes wrong? Who can be called on the phone? People on HPe?

Last time I emailed MIPS for an Atlas prototype I was said all the engineers who had designed the board are all retired, getting a figurative finger _|_ from the new team. So, at the end, I had to reverse engineering a couple of circuits as they were not present as described in the documentation because the hardware was modified. It was so freaking time consuming, and as you can guess ... I was not able to respect the deadline.

The problem here is ... managers don't understand details, if they are said "documentation was send", they assume you are ready to work, when you speak about external prototypes .... it's viral, since they are usually modified, so the documentation is usually inconsistent.

Now, if you are involved into the development of a prototype with an external company, then there is a natural form of collaboration, you are notified by their changes as you are part of the team.

When the prototype is already formed as definitive product , and you happen to put your hands on it some years later from an external company ... well there is another form of collaboration as where you have to phone to someone who can't be faced in person, which usually ends with a figurative finger if he/she has no interest in collaboration, usually because he/she has other tasks to do, and It means for sure delays.

Now, I am a freelancer. When I don't respect deadlines, customers are not happy, and they might decide to ... delay the payment as well.

Can I delay my electricity bills?

[jan-jaap]

Like I said, this system is retired. There are no deadlines to be met.

Can we keep OT opinions in the Everything Else section? Thanks.

[Y888099]

Like I said.

You said for a project. Which project is it?
A super secret banana project?
Oh, your employer asked to keep silent.
So let's talk about the blue plastic and the turbo air.

Topics like this are ridiculous for me

[mapesdhs]

Topics like this are ridiculous for me

So why do you post?

I enjoy seeing jj's efforts to keep the prototype going, and sometimes it reveals interesting things about where the design came from.

I note for instance that some Fuel mbds have an extra SCSI port on the mbd. Is it functional? Not checked yet. Most mbds have the place for it, but no socket present.

Ian.

[Trippynet]

What Ian said. There's plenty on here that I find ridiculous, and I just ignore this stuff. At the end of the day though, this is an SGI forum, and this is a thread about a prototype SGI computer in a funky coloured case, hence, lots of people are interested in it (myself included).

It's also worth noting that some of the troubleshooting that JJ did earlier in this thread helped me resurrect my own Fuel last year after it suffered from battery failure, so the discussion here does have technical merit/info that has uses beyond this prototype system.

In short though, if you find the topic ridiculous, please don't post and just leave it for others.

[mapesdhs]

I also referenced JJ's repair info while sorting some misbehaving Fuels these past couple of weeks, helped me get three working again, and now I'm more confident I can repair some mbds that are acting weird.

Ian.

[dexter1]

Still looks stunning, Jan-Jaap! Keep that blue beast afloat.

BTW, what is the PROM version? i'm wondering if the NEXTGEN logo remains across PROM updates... I guess a new PROM update will nuke it, so maybe it would be a good idea to do a dump. I have a tool that can search for Images encoded in the PROM, though i might have to dig that up somewhere.

[Dodoid]

Needs badges !

Positioning is imperfect, I blame GIMP because Pixelmator doesn't like to do perspective.

I knew those exactly 1920x1080 slides of every SGI I used in my History series weren't a waste of disk space!
Except the Altix XE slide. That one's really only good for making vaporwave art out of .

[jan-jaap]

BTW, what is the PROM version? i'm wondering if the NEXTGEN logo remains across PROM updates... I guess a new PROM update will nuke it, so maybe it would be a good idea to do a dump. I have a tool that can search for Images encoded in the PROM, though i might have to dig that up somewhere.

It has PROM "SGI Version 6.103 TEST built 10:26:01 AM Nov 29, 2001"

The PROM is normally updated at the end of an IRIX installation, so i installed IRIX 6.5.16 (oldest that supports Fuel, I think, and 6.5.15 is the *only* version I don't have ...) on my 'real' Fuel, and moved the disk to the NextGen system.

I'm happy to report that I was able to boot it and dump the PROM to disk (flash -D). Oh, IRIX 6.5.16 comes with PROM version 6.113. The system isn't on the net right now so I cannot upload the image file.

The mainboard says the L1 (used to be) version 1.9.6, but it's currently at 1.9.9:

Code: Select all

001a01-L1>version
L1 1.9.9 (Image B), Built 11/06/2001 09:18:53    [P1 support]

P1 might refer to the proto level -- browse back to see the sticker on the back of the system (proto level: P3). The PROM chip on the main board says the PROM is version "P3 SN1".

The L1 still knows about "NextGen" and refers to the blue LEDs on the front:

Code: Select all

001a01-L1>help nextgen
nextgen led red on
        turn on red NextGen LED
nextgen led red OFF
        turn off red NextGen LED
nextgen led blue on
        turn on blue NextGen LED
nextgen led blue off
        turn off blue NextGen LED
nextgen led red blink start
        start blinking red NextGen LED
nextgen led red blink stop
        stop blinking red NextGen LED
nextgen led blue blink start
        start blinking blue NextGen LED
nextgen led blue blink stop
        stop blinking blue NextGen LED


There seems to be something wrong with the serial of the system. The Dallas and the yellow 'snaphat' of the NextGen board were both dead, so I used the ones from the 'real' prototype. the barcode of that system (MED907) somehow showed up as the Reference Brick Serial Number of the system (main board barcode MDG756). When I tried to 'serial clear' the L1, it inserted a bogus Reference System Serial Number. That 'MAC' seems odd as well

Code: Select all

001a01-L1>serial all

Data                            Location      Value
------------------------------  ------------  --------
Local System Serial Number      NVRAM         08:00:69:0B:C3:C8
Reference System Serial Number  NVRAM         L0000000
Local Brick Serial Number       EEPROM        MDG756
Reference Brick Serial Number   NVRAM         MED907

EEPROM      Product Name    Serial      Part Number           Rev  T/W
----------  --------------  ----------  --------------------  ---  ------
NODE        IP34            MDG756      030_1707_002          E    00
MAC         hardware detected, no data available
PIMM        IP34PIMM        MDG739      030_1708_002          G    00
XIO         ASTODYB         MDG840      030_1725_001          D    00

EEPROM      JEDEC Info                Part Number         Rev
----------  ------------------------  ------------------  ---
DIMM 0      CE0000000000000026BAAE00  M3 46L3313BT1-CA0   0B
DIMM 2      CE0000000000000026C3AE00  M3 46L3313BT1-CA0   0B
DIMM 1      CE0000000000000028C12601  M3 46L3313BT1-CA0   0B
DIMM 3      CE00000000000000269CF500  M3 46L3313BT1-CA0   0B




[jan-jaap]

Needs badges !