PCI-MXI-2 EEPROM corruption

[RoyO]

While trying to change the configuration of PCI-MXI-2 cards, whenever new config info is attempted to be written in, MAX freezes.

When MAX is started up again and the settings are checked, the following error appears:

  PCI-MXI-2 configuration EEPRON is invalid

  The problem may be caused by:

    - memory conflict with another device

    - incomplete configuration update

    - EEPROM corruption

  To correct the problem, boot with the Factory Override switch set.

This is followed by the error "Unable to access conifguration information".

Then when the card is changed to have the above switch set (per instructions given in the NI document "Restore the Factory Configuratin Defaults for the PCI-MXI-2 and the VXI-MXI-2"), all user configuration is blank.  Again, anytime data is attempted to be written to the card, either by loading an existing config file or by trying to set to the "Default Configuration", MAX freezes.

Put the same PCI-MXI-2 card into another PC to see if I could change it there, but couldn't change it there either.

So thought the card had defective EEPROM, but then tried another PCI-MXI-2 card, and had the same issue again!

Tried both VISA 3.5.1 Beta as well as VISA 4.3, same results.

Note that Resman works fine, just can't change the settings.

Any ideas?

[kcoggins]

You might have a memory conflict with another device.  Check to see what other non-standard hardware you have in your computer, and if you do, see what memory space it uses as compared to the PCI-MXI controller.

How many PCI-MXI-2 cards do you have in this system?

[RoyO]

Thanks for the suggestion, I'll look into possible memory conflicts.  There is only a single PCI-MXI-2 card.

Also, both the cards I have that are showing corrupted EEPROM are 1995 vintage.  I also tried a 2004 card that I configured on a different system and it worked fine on my target system, though I am not brave enough to try configuring the new card on my target system.  Someone at work mentioned there is a difference in the versions of ASICS that could be an issue.

And once the EEPROM is corrupted, it seems there is no way to clear it on any system.

[MXI Master]

RoyO,

To clarify, you are able to run ResMan successfully with the FOV switch in both positions, but when you try to write to the EEPROM it locks up and throws the error like you specified above. Correct me if I am wrong, please.

Also, what version on NI-VXI are you running on these systems?

Do you have switches set like the pictures in Appendix B of the MXI-2 Refrence Manual?

[RoyO]

Yes, you have it correct.  And yes, the switches are set as described in the manual you referenced.  I am running with NI-VXI 3.5.0 but have also used 3.3.0 on another system.

Memory usage appears clean.  There is a shared interrupt, but I don't think that should be a problem, here's my system:

0xD0000000-0xEFFFFFFF PCI bus
0xD0000000-0xEFFFFFFF Intel(R) 975X PCI Express Root Port - 277D
0xD0000000-0xEFFFFFFF NVIDIA Quadro FX 3500
0xF4000000-0xFEBFFFFF PCI bus
0xE5000000-0xE71FFFFF Intel(R) 975X PCI Express Root Port - 277D
0xE5000000-0xE71FFFFF NVIDIA Quadro FX 3500
0xE6000000-0xE6FFFFFF NVIDIA Quadro FX 3500
0xE0500000-0xE0503FFF Microsoft UAA Bus Driver for High Definition Audio
0xE0200000-0xE04FFFFF Intel(R) 82801GR/GH/GHM (ICH7 Family) PCI Express Root Port - 27E2
0xE0400000-0xE040FFFF Broadcom NetXtreme Gigabit Ethernet
0xE0504000-0xE05043FF Intel(R) 82801G (ICH7 Family) USB2 Enhanced Host Controller - 27CC
0xE4200000-0xE4207FFF PCI-MXI-2
0xE2000000-0xE3FFFFFF PCI-MXI-2
0xFEC00000-0xFEC00FFF Advanced programmable interrupt controller
0xFED00000-0xFED003FF High Precision Event Timer
0xE0504400-0xE05047FF Intel(R) 82801GR/GH SATA RAID Controller
0x0000-0x9FFFF System board
0xFEC01000-0xFECFFFFF System board
0xFED00400-0xFED3FFFF System board
0xFED40000-0xFFFFFFFF System board
0xF0000000-0xF3FFFFFF System board
0xA0000-0xBFFFF PCI bus
0xA0000-0xBFFFF Intel(R) 975X PCI Express Root Port - 277D
0xA0000-0xBFFFF NVIDIA Quadro FX 3500
0xD0600-0xE3FFF System board
0xD4000-0xFFFFF System board
0x100000-0xCFFFFFFF System board

IRQ 0 System timer
IRQ 1 Standard 101/102-Key or Microsoft Natural PS/2 Keyboard
IRQ 4 Communications Port (COM1)
IRQ 6 Standard floppy disk controller
IRQ 8 System CMOS/real time clock
IRQ 9 Microsoft ACPI-Compliant System
IRQ 12 PS/2 Compatible Mouse
IRQ 13 Numeric data processor
IRQ 14 Primary IDE Channel
IRQ 15 Secondary IDE Channel
IRQ 16 Intel(R) 975X PCI Express Root Port - 277D
IRQ 16 NVIDIA Quadro FX 3500
IRQ 16 Intel(R) 82801GR/GH/GHM (ICH7 Family) PCI Express Root Port - 27E0
IRQ 17 Intel(R) 82801GR/GH/GHM (ICH7 Family) PCI Express Root Port - 27E2
IRQ 17 Broadcom NetXtreme Gigabit Ethernet
IRQ 18 Intel(R) 82801G (ICH7 Family) USB Universal Host Controller - 27C9
IRQ 18 PCI-MXI-2
IRQ 19 Intel(R) 82801GR/GH SATA RAID Controller
IRQ 20 Intel(R) 82801G (ICH7 Family) USB Universal Host Controller - 27C8
IRQ 20 Intel(R) 82801G (ICH7 Family) USB2 Enhanced Host Controller - 27CC
IRQ 21 Microsoft UAA Bus Driver for High Definition Audio
IRQ 21 Intel(R) 82801G (ICH7 Family) USB Universal Host Controller - 27CA
IRQ 22 Intel(R) 82801G (ICH7 Family) USB Universal Host Controller - 27CB

[RoyO]

Yikes, after I posted that I noticed a conflict with the graphics card, looks like I better try a different one.

[RoyO]

Alright, tried it with a different graphics card with none of the memory shared with the PCI-MXI - same results.

Another question: is it possible to corrupt the cards when setting the configuration so that it is impossible to clear them (other than sending them in for repair)?

[TobyHex]

Hey Roy,
I looked into your issue a bit more and wanted to make sure that you have run across a couple of different articles that relate to your issue. From what I can gather you are still experiencing the original problem and with that I was wondering if you might respond with the type of PCI-MXI-2 card you are using. As well if you could get me the serial number and part number for your VXI-MXI-2 card so that I can look into that a bit further. Attached are the links that I would like for you to look into and let me know if these are of any help to your problem or what might be the problem we are facing at the moment.

Links-
Developer Zone-Restoring EEPROM Configuration Settings on VXI/VME Embedded and Remote Controllers
KnowledgeBase- Corrupted EEPROM Message When Using PCI-MXI-2

Thanks again and let me know what I can do to help.

Regards,

Nicholas K
Applications Engineer
National Instruments

[RoyO]

Thanks for the response.  The first document I am familiar with, the second is new to me - I'll go through it and see if it will help.

I am currently working with 3 PCI-MXI-2 cards - two of them show up as having corrupted EEPROM, the other works great.

Corrupted card #1 has the following numbering:

A1BA1B

ASSY: 182685C-01

PCB:  182687B-01     847578309   UL94V-0  (C)1995

Chip U8: 70236    1B.01  1796 H

Chip U12 (ASIC): MITE-AQ

Chip U24: 78235  3A-01  1796 H

Corrupted Card #2 has:

11CCFA9

ASSY 182685K-01

Chip U8: 70236  1F-01  4905  HO

Chip U12 (ASIC): MITE-BQ

Chip U24: 702353B  -01 M V  2505

(c) 2004

Good Card:

11C29C1

ASSY 182685K-01

Chip U8: 70236  1F-01  4605  HO

Chip U12 (ASIC): MITE-BQ

Chip U24: 702353B  -01 M V  1005

(c) 2004

Again, I have two PC's I have been working with, both HP xw4400 Workstations.  Fairly similarly configured but not exactly.  One has seemingly corrupted two PCI-MXI-2 cards so far, the other PC has let me change configuration fine on the third card but can not change configuration on the other two cards.  I have also tried a third PC, an older IBM, and can not change configuration on the two corrupted cards but can on the good card.  I still do not dare to try changing configuration using the first PC.

[WillPage]

On some newer/faster computers, EEPROM writes happen too quickly, corrupting the EEPROM.

NI-VXI 3.5.1 adds a rate-limiter to the EEPROM writing code so that the corruption you're seeing should no longer happen, regardless of processor speed.