[Irinikus]

I've managed to source a replacement chip locally for $6.46!

It will be a good educational exercise to try to have this thing fixed, as ultimately we'll all have to carry out this type of maintenance on our machines more and more, to keep them going into the future!

[Irinikus]

I was moving along reasonably with my IRIX setup, when the KONA EEPROM failed without warning.



I was hoping to release quite a few videos to demonstrate this machines performance while I was down here. This will unfortunately not be possible this time! :(

I have made contact with 3ddoc and will be purchasing a new GE16/4 from him on the 17th of this month. It will unfortunately only arrive on my end at the end of the month, and I will unfortunately have left Cape Town by then!

I will still purchase some replacement EEPROM chips, as I have no idea as to why mine failed (or the conditions which lead to its failure), and @Jan-Jaap has also had the EEPROM in his IR system fail in a similar way. So this is not an isolated incident!!!

So this saga will continue when I'm next down here!

[Irinikus]

Looking at the EEPROM chip on the GE16, the 25TC on the chip denotes a 250ns response time.



I've come across these with a 200ns response time.



Is this a make or break feature?

[weblacky]

I think you’d be fine. Faster access times aren’t a hard timing issue with the communication rate. It’s just the rate you can request info, slower is okay for faster capable chips.

Someone else can chime in and correct me if I’m wrong but I’m pretty sure this would just be like a minor upgrade in the speed and the difference is negligible anyway.

If everything else is the same in terms of memory type and pin location you’re probably good to try them.

[fleedwood]

I was moving along reasonably with my IRIX setup, when the KONA EEPROM failed without warning.

are you sure this meant with KONA EEPROM ? I'd call this chip a flash prom and would expect something smaller in memory sitze as KONA EEPROM.

[Irinikus]

I was moving along reasonably with my IRIX setup, when the KONA EEPROM failed without warning.

are you sure this meant with KONA EEPROM ? I'd call this chip a flash prom and would expect something smaller in memory sitze as KONA EEPROM.

This is what the machine stated during the power on diagnostics:



This is what I got when I attempted to re-flash the EEPROM:



This is the description of the chip in question:



The member with the best knowledge to be able to answer this question is @Jan-Jaap. (As he had the EEPROM transplanted from one GE16 to another.) Let’s see what he says.

[Irinikus]

I was moving along reasonably with my IRIX setup, when the KONA EEPROM failed without warning.

are you sure this meant with KONA EEPROM ? I'd call this chip a flash prom and would expect something smaller in memory sitze as KONA EEPROM.

You may just be right!!!

I just came across this little chip on the GE16, which is also an EEPROM:



Here's a shot from it's data sheet:

[Irinikus]

Cool! Thanks!

Now it comes to the question as to which chip actually serves as the EEPROM on one of these little monsters?

Is it this little EEPROM chip:





Or this one:

[weblacky]

So I can answer half of this after really paying attention to your posts..I should have seen this sooner.

The DS2505 is a special chip that cannot be reprogrammed (one-time data add only). Data is written in something similar to records to this chip. You can always append additional records to a chip but you cannot reclaim the areas that previous records have been written. I don’t know if you can nullify old records by putting in newer records. The point is, you put in a line of data at a time into this chip or more. You don’t flash the entire chip at once. Once that data is on the chip, it’s there forever. Cannot be reprogrammed. It’s still write-once for each storage region on the chip.


SGI often uses it to carry the board’s part # and other data needed early by PROM.  If you have a board part number for this device your troubleshooting show on hinv then likely the DS2505 works. It requires a special programmer to even program that chip the first time that you’ll never get today.  I briefly talked with another SGI forum user about this chip because it’s also used in Octane CPU modules to carry CPU ID data.  SGI loved these chips, and they rarely fail (really). So you’ll find them in several systems peripherals.

I believe it’s also way too small for your data.

Print page 253: https://bukosek.si/hardware/collection/s...ooklet.pdf

“Code+data length:0x1caac”

Which should be like 117Kb?

If you actually verify how big the tport.bin file is, it should be larger than 16Kb, yes?

So I can 100% tell you that there is no way to reprogram a DS2505 on board of system. Not possible. Not supported.  So your EEPROM cannot be the DS2505.

But this doesn’t 100% verify your other chip is the one you’re looking for.

[jwhat]

Hi Irinikus,

does it report same error if you just do an EEPROM clear? :
>> "./ireeprom -c"

This will set IR back to default config.

Cheers from Oz,

jwhat/John