Getting Started, Documentation, Tips & Tricks

it appears as though the NVRAM battery on my Onyx is kaput... date and enviroment variables reset after hard power off, and a nice stream of NVRAM battery messages in the console output....

ive looked all over the IO4 and VCAM and cannot see anything that resembles a battery, removed the system monitor and there is even less on there! had a look on both sides of the backplane.....

where the dickins is it hidden? and is it an easilly replaced battery or a soldered on nightmare????

MattPayne wrote: the NVRAM battery on my Onyx is kaput... date and enviroment variables reset after hard power off, and a nice stream of NVRAM battery messages in the console output....

ive looked all over and cannot see anything that resembles a battery.

where the dickins is it hidden? and is it an easilly replaced battery or a soldered on nightmare????

I hate to pull up an old thread, but I'm in the same boat. Anyone dealt with replacing (or know the location of) the NVRAM battery in an ONYX Deskside?
Thanks.
-milatchi

The Onyx manual states that the battery is in the System Controller. I have no idea what the battery looks like in there, but that's the place to start.

System Controller

Located just above the SCSI drive rack is an on/off key switch and the System Controller display panel. The System Controller is a microprocessor-controlled, battery-backed subsystem that is mounted directly to the system backplane. It monitors various system operations, including chassis temperature, system fan speed, backplane voltages, and the system clock. The battery backup supports only the System Controller's NVRAM and internal system clock.

When any operating parameter exceeds or drops past a specified limit, the System Controller can execute a controlled shutdown of the Onyx deskside system. During such a shutdown procedure, the System Controller maintains a log with the last error message(s) received before the shutdown.

Chapter 2 shows the location of the System Controller's front panel on the chassis. Figure 3-13 in Chapter 3 identifies its control buttons.

To understand and use the System Controller, see “Using the System Controller,” in Chapter 5

nekonoko wrote: The Onyx manual states that the battery is in the System Controller. I have no idea what the battery looks like in there, but that's the place to start.

Thanks nekonoko. I'll check it out.

yearly thread bump! lol!!

any joys in finding the battery?? time isnt somuch of an issue, but the default boot is to disable the GFX so requres a terminal and a reboot to play with...

checked the back pf the system controller, but nothing... its purely a display and buttons... cable through to the backplane, which is also devoid of anything even remotely battery looking

Ok, I don't have an Onyx...So hold your pistols please!

I was looking through the Gallery and came across an image of Pentium's Onyx IO4 board. On it there was a Dallas chip. Now I know (maybe that should be 'I think') the Octane gets it's battery backup from the Dallas chip, would the Onyx not do the same? Would the Dallas chip also hold the System MAC address?

See Image here - > IO4 Board

-J

JacquesT wrote: Would the Dallas chip also hold the System MAC address?

Nope. Not for an Onyx

nekonoko wrote: The System Controller is a microprocessor-controlled, battery-backed subsystem that is mounted directly to the system backplane.

edited for clarification:
The system controller is the most-left board plugged in to the back of the backplane, closest to the AC inlet. But to fix the symptoms described above (bad MAC address, etc) you need to check the dallas on the IO4. The IO4 and system controller both have a Dallas chip.

I'm going to take a good guess that the battery in one of the dallas chips (one on the IO4 and I THINK one on the system controller) have gone south.
It should not be too hard to grind out the terminals for the battery and solder a CR2032 in its place.

If you need more info on what to do, read this.


I also see my persistance to photograph every part of my system has paid off for someone.

your a star! I noticed the Dallas chips and was looking for a battery closeby, I didn't think of looking inside!

its all bit scary fixing! but the machine is a semi brick (or should that be pallet of bricks?!) without the battery! there are 2 timer chips... ine on the IO4 and one on 'one of those power regulator thingies' whats the betting i get the wrong one!

pentium wrote: I'm going to take a good guess that the battery in one of the dallas chips (one on the IO4 and I THINK one on the system controller) have gone south.
It should not be too hard to grind out the terminals for the battery and solder a CR2032 in its place.

If you need more info on what to do, read this.

**Photo Link to IO4 chip**

I also see my persistance to photograph every part of my system has paid off for someone.

If you need more info on what to do, read this.

those dallas chips are a real pain in the az ... i recently saw the clock reseting on my Indigo2 and this is first sign of a dying one ... they should last for at least 10 years ?

from the linked hack - i`m wondering what is the repercussion of leaving the old - empty - battery inside and soldering contacts together with the new one. depending of what electrocemichal process happened inside the old battery and what is the actual internal resistance of it, you can find yourself shortcutting the battery u re attaching outside or u can short the life of the new one significally ? dunno, should be checked out ...

there are brand new Dallas replacement parts/chips with newer serial numbers that are compatible with those used in sgi`s and aren`t so expensive ~$10 - $25

from wiki: "The original battery was made by Dallas Semiconductor, now a division of Maxim Integrated Products. The original unit was marked the "DS-1386-8K-150", however its replacement unit, the "DS-1386-8K-120" can be directly substituted with no ill effects."

The hacking process physically disconnects the battery from the chip. Unless your soldering skills are not too good you should not have a new issue with the old battery killing the new battery.

question-
does this hacking technique work on the "watchdog timekeeper" chip (the one with the "bulldog" on the face instead of the "alarm clock") as well? is it in the same location? I'm not seeing the blank pins on the datasheet where it would be....

compuman86 wrote: question-
does this hacking technique work on the "watchdog timekeeper" chip (the one with the "bulldog" on the face instead of the "alarm clock") as well? is it in the same location? I'm not seeing the blank pins on the datasheet where it would be....

Same idea - can't recall exactly the details. Try using a magnet to find the side with the battery in it.

SAQ wrote:

compuman86 wrote: question-
does this hacking technique work on the "watchdog timekeeper" chip (the one with the "bulldog" on the face instead of the "alarm clock") as well? is it in the same location? I'm not seeing the blank pins on the datasheet where it would be....

Same idea - can't recall exactly the details. Try using a magnet to find the side with the battery in it.

yeah, that should work as long as it has iron or steel in it. i'm planning to slowly and carefully disassemble one, hoping to get to the pins. too bad we don't have access to an x-ray machine, that would definitely make it easy to find the exact location and depth of the pins...