IRIX Network Forums
Tezro Dallas Chip Orientation - Printable Version

+- IRIX Network Forums (//forums.irixnet.org)
+-- Forum: SGI/MIPS (//forums.irixnet.org/forum-3.html)
+--- Forum: Hardware/Triage/Repair (//forums.irixnet.org/forum-11.html)
+--- Thread: Tezro Dallas Chip Orientation (/thread-2413.html)

Pages: 1 2 3 4 5 6

Tezro Dallas Chip Orientation - jwhat - 08-22-2020

Hi SGI'ers,

I sacrificed the Dallas DS1742W-120 out of my Tezro some time ago so I could get new serial number onto Numalink Router by way of chip swap.

I am now wanting to put a Dallas back into Tezro to see if I can get L1 contact with machine back up and running as I want to try to do another chip swap.

Unfortunately before pulling Dallas from Tezro, I did not take any pictures ..

Could someone possibly open up a working Tezro and take a picture of Dallas orientation. The chip is in socket on the front part of the system board, near the black plastic PCI board positioning and fan holder on the same side as the IO9 & other cards.

For those interested I have posted Dallas orientation picture for Numalink and Origin 350 (Chimera) here: 
https://tips.graphica.com.au/sgi-irix-notes/

Thanks in advance.

Cheers from Oz.

John.

RE: Tezro Dallas Chip Orientation - weblacky - 08-22-2020

eBay auction for Tezro board, I think I see the RTC...https://www.ebay.com/itm/NEW-SGI-Tezro-Tower-IP53-Workstation-System-Board-030-1881-007-WKSTA-INTFC-2-/253680787894

Stupid RTC Dallas logo is facing towards main PSU power connector in this picture.

Also if this a new chip, we've had a debate on how to default it correctly, I left what I thought were the right instructions but the previous user never came back to try them...simply placing a new RTC in the tezro won't be enough, it doesn't auto-format it, Tezro will keep complaining about RTC after insertion.

You need to hook to the L1 serial port, perform a "help eeprom" and see if an option appears in help that says something about "write defaults". This info comes from Fuel RTC instructions (same chip) but no one has confirmed if similar instructions for tezro work/exist. You may be the first!

This is the thread where I left the last possible hints. http://forums.irixnet.org/thread-452-post-15681.html#pid15681

We've never gone farther. Read this and see if that gets you any new clues. Please post the full output of the help "help eeprom" command on your tezro once you have the RTC in place. Hopefully that's the answer.

Thanks!

RE: Tezro Dallas Chip Orientation - jwhat - 08-22-2020

(08-22-2020, 10:00 AM)weblacky Wrote:  eBay auction for Tezro board, I think I see the RTC...https://www.ebay.com/itm/NEW-SGI-Tezro-Tower-IP53-Workstation-System-Board-030-1881-007-WKSTA-INTFC-2-/253680787894

Stupid RTC Dallas logo is facing towards main PSU power connector in this picture.

Also if this a new chip, we've had a debate on how to default it correctly, I left what I thought were the right instructions but the previous user never came back to try them...simply placing a new RTC in the tezro won't be enough, it doesn't auto-format it, Tezro will keep complaining about RTC after insertion.

You need to hook to the L1 serial port, perform a "help eeprom"  and see if an option appears in help that says something about "write defaults".  This info comes from Fuel RTC instructions (same chip) but no one has confirmed if similar instructions for tezro work/exist.  You may be the first!

This is the thread where I left the last possible hints.  http://forums.irixnet.org/thread-452-post-15681.html#pid15681

We've never gone farther.  Read this and see if that gets you any new clues.  Please post the full output of the help "help eeprom" command on your tezro once you have the RTC in place.  Hopefully that's the answer.

Thanks!


Well spotted on eBay auction... that has the chip the right way, as I tested with a couple of my chips.

Chip #1 - DS1742W-120+ from lot I purchased quite some time ago on ebay
     Nothing... on Console port completely dead. I recall having no luck with this a few years back which was why I took the chip out of Tezro


Chip #2 - DS1742W-120 from any of O350 or Numalink chassis .. get scrolling INFO & ALERT:

ALERT: NODE EEPROM read error, no acknowledge
INFO: Cannot disable VRM: 9
INFO: Cannot disable VRM: 10
INFO: Cannot disable VRM: 11
INFO: Cannot disable VRM: 9
INFO: Cannot disable VRM: 10
INFO: Cannot disable VRM: 11
ALERT: NODE EEPROM read error, no acknowledge
INFO: Cannot disable VRM: 9
INFO: Cannot disable VRM: 10
INFO: Cannot disable VRM: 11
INFO: Cannot disable VRM: 9
INFO: Cannot disable VRM: 10


The VRM 9,10,11 errors are to be expected as I have no IP53 or IP59 board in the machine, so there are no VRMs which I think are on the CPU boards.

Putting in an IP53/IP59 is quite fiddly as I would have to take board out of another machine.
I might try to plug it into an l2 controller to see if that allows me to get to L1 prompt.
I also have a few more of the eBay 120+ chips but I suspect they are all going to be duds...

Thanks for pointer.

Cheers from Oz,



John.

RE: Tezro Dallas Chip Orientation - Geoman - 08-22-2020

(You can also change the snaphat on the multi-purpose io-board board, since mine was exhausted alsoSmile

After changing the Dallas, you have to boot up the L1-controller via serial terminal and enter

Code:
serial set A1234567

this serial number is to be found on the back of the machine, and consists of an upper case letter and seven digits. After that Tezro will boot-up again.

RE: Tezro Dallas Chip Orientation - jwhat - 08-22-2020

thanks Geoman, and appreciate tip on Snaphat. I have another machine getting nvram error and likely need to refresh that one.

cheers from oz,

John

RE: Tezro Dallas Chip Orientation - jwhat - 08-23-2020

Hi Weblacky & Geoman,

hooked up L2 (easier than swapping in IP53/IP59) and I cannot get to l1 even via this as L1 is cycling up and down (as seen in USB open/close). Here is L2 via telnet:


telnet serial.XX.XX
Trying 192.168.XXX.XXX...
Connected to serial.XX.XX.
Escape character is '^]'.

Linux 2.4.7-sgil2 (192.168.XXX.XXX) (ttyp0)


SGI L2 Controller

INFO: connection established to localhost, to quit enter <ctrl-]> <q>
serial-001-L2>ver
L2 version: 1.48.0
serial-001-L2>config
L2 192.168.XXX.XXX: - 0001 (LOCAL)
serial-001-L2>INFO: opened USB device at b1;p1;d2 (/dev/sgil1_0)
INFO: closed L1 USB device /dev/sgil1_0 [b1;p1;d2]
INFO: opened USB device at b1;p1;d3 (/dev/sgil1_0)
INFO: closed L1 USB device /dev/sgil1_0 [b1;p1;d3]
INFO: opened USB device at b1;p1;d5 (/dev/sgil1_0)
INFO: closed L1 USB device /dev/sgil1_0 [b1;p1;d5]
INFO: opened USB device at b1;p1;d7 (/dev/sgil1_0)
INFO: closed L1 USB device /dev/sgil1_0 [b1;p1;d7]
INFO: opened USB device at b1;p1;d8 (/dev/sgil1_0)
INFO: closed L1 USB device /dev/sgil1_0 [b1;p1;d8]
INFO: opened USB device at b1;p1;d10 (/dev/sgil1_0)
INFO: closed L1 USB device /dev/sgil1_0 [b1;p1;d10]
INFO: opened USB device at b1;p1;d11 (/dev/sgil1_0)


I also have an Onyx4 G2 chassis (the infamous ATI graphics), which as Dallas of unknown origin (as I did quite a bit of chip swapping when working on getting Numalink working ;-) ).
This one has the follow errors and like the Tezro I cannot get L1 prompt to reset it..


...
...
ALERT: PS 2 EEPROM board info checksum error                                    
ALERT: NODE EEPROM read error, no acknowledge                                   
ALERT: PS 2 EEPROM board info checksum error                                    
ALERT: NODE EEPROM read error, no acknowledge                                   
ALERT: PS 2 EEPROM board info checksum error                                    
ALERT: NODE EEPROM read error, no acknowledge                                   
ALERT: PS 2 EEPROM board info checksum error
...
...

So in both cases I am not able to get to L1 to do a serial reset or EEPROM init ...

As you asked here is l1 help from O350 (which I think will be the same as that of the Tezro):


SGI SN1 L1 Controller                                                           
Firmware Image B: Rev. 1.44.0, Built 07/17/2006 18:20:50                        
                                                                                
                                                                                
001c01-L1>help eeprom                                                           
eeprom                                                                          
        show brick eeprom data.                                                 
eeprom <exp> <exp> <exp>                                                        
        show brick eeprom data at <eeprom> <offset> <length>.                   
eeprom Fuel write default                                                       
        write standard Fuel EEPROM data to MAC EEPROM                           
001c01-L1>

So it mentions Fuel write default explicitly...
And also seems to be making some reference to MAC EEPROM, which would indicate that Fuel programs its Ethernet MAC in EEPROM somewhere .... ???
This was like the older machines which have MAC address in EEPROM and on losing battery the ethernet MAC got lost.

Last thing I could do would be to put IP53/IP59 board into Tezro and see if that fixes errors and lets L1 come up ok.

Cheer from Oz,

John

RE: Tezro Dallas Chip Orientation - weblacky - 08-23-2020

Sorry I wasn't more clear....L1 Serial port, USB may not work when serious crap happens. Use a DB9 null modem cable directly to the Tezro's L1 Serial on rear of machine (that's how I talk to the L1) and try again.. Also, don't try to start the system, plug in serial, then apply power, we're trying to talk to L1 while OS/PROM are powered down.

Also, again the fuel instructions are the only thing we have, it's unknown. What previous Fuel-specific posts found was the write defaults eeprom command "formats" the NVRAM and adds back stuff when executed. But we don't know if the Tezro needs that or not.

I do agree with what was said about possibly being able to use set serial on a blank chip *just once. if you know your tezro serial and want to set things back up. Can you try the serial set command posted previously (after you get stable L1 access of course).

Let us know, we don't have good info on this...we just know that unlike the Indigo2/O2/Octane the Fuel and tezro don't autocorrect and fill the NVRAM with defaults automatically on checksum failure...it just stays and complains...what to do next...that info (if it existed) was lost with nekochan.
Oh, I forgot...yeah, please put the Tezro back together...I wouldn't try this NVRAM stuff without the IP Board in...sorry brushed over that.
Not to be a broken record on this too. Here's another previous post where the user got farther (though no real good summary of events): https://forums.irixnet.org/thread-1470-page-3.html

I *THINK* he got into L1,
typed: let the carnage begin
then typed: serial set SSN

Where SSN is replaced with the serial on the back of the systems starting with a letter and the rest are numbers.

RE: Tezro Dallas Chip Orientation - jwhat - 08-23-2020

Hi Weblacky,

you have been clear enough, thanks.

I should have made it clear that the L1 diagnostics I first posted are coming from the serial Console port... (not L2).

I have been monitoring via both serial Console and L2 via USB.

I did read your prior thread, but my problem is that I can't get L1 prompt, likely as you say due to not having an IP53/IP59 installed.

I have also read the Pymble Software method to flip Numalink serial and other Dallas stuff, as I used this guidance when first swapping chips to get Numalink going. The full thread of this is still available here: http://pymblesoftware.com/blog/?p=295

I was going to pull one from O350 but the easy to access machine has IP59 (1GHz board) in it which means to pull it out requires taking out fans to access retaining screws and the machines with IP53 boards are in deep wiring entanglement with Numalink cabling and stuff, so again hard to get access. I should have tried this when I had spare IP53 lying around ;-( .

Maybe this is task for next weekend...

Cheers from Oz,


John

RE: Tezro Dallas Chip Orientation - jwhat - 08-23-2020

Hi Weblacky & Geoman,

bit the bullet and pulled IP53 out of O350.

this got rid of the (L1 Console) error:

ALERT: NODE EEPROM read error, no acknowledge

But the L1 still never came up and you could see the LED on system board cycling GREEN/RED and L2 getting USB opened/closed message as it cycled.

I get the same results when using either of the Dallas DS1742W-120 taken out of other machines.

I also retested with 3 "new" Dallas DS1742W-120+ chips and in all cases got:

1. Constant GREEN on system board LED
2. No USB opening up on L2
3. Nothing being logged on L1 Console port
4. No L1 Prompt on Console Port or access via L2

Again as I am not able to get into L1 it was not possible to initialise the chip via eeprom command.
So possibilities are:
1. All the new Dallas DS1742W-120+ Chips are duds
2. Tezro cannot work with uninitialised Chip
3. Tezro will not accept chips already initialised from non Tezro (Chimera ?) machines.

So maybe the Fuel has found a new purpose as chip initialiser... ;-)

I think what is needed is for someone with Fuel & Tezro to see if you can swap Dallas chips from Fuel to Tezro as way to get new chips going.

As none of the new DS1742W-12+ appeared to work, I will Dremel one, so we can see where the battery is.

I will not do any more testing on this one, as I am putting the IP53 board back into its original chassis.

TRIVIAL aside, the Tezro uses small gauge screws to hold the IP53 board down in chassis than the O350.


Cheers from Oz,

John.

RE: Tezro Dallas Chip Orientation - weblacky - 08-23-2020

OK, Obviously I'm out of element here, but that's never stopped me before.

Let's go back to basics:

Here's my dumb question...what IP board did you actually put in and do you know it actually works with the current firmware your Tezro currently has? Are you sure you haven't accidentally flipped your L1 into its backup firmware version (A vs B) which would possibly be an older firmware?

I mean, your L1 isn't coming up. The don't believe in "bad" RTCs right now...because unless it was physically abused or shorted, it's a NVRAM storage with a clock and battery...what SHOULD go bad is the battery. On clusters, some people WANT a bad battery so the serial is forgotten at rack power-down and the L2 can assign them new serials everytime they boot (so it's ultra-flexible to add/remove bricks). If you had a dead battery...everything would still work...you'd just lose settings and L1 time when you take power off the system. That still doesn't explain the symptoms you have.

The Tezro manual claims a solid green means it's booting, so sounds to me like it's stuck booting, so I'd ask the same question again...what L1 version do you think you have and what IP board did you try in it when you saw all this?