solved now. sx6012 hardware problem

Freebsd1976

Active Member
Feb 23, 2018
255
39
28
Last week play with my sx6012 , when use programmer clip on U5 (st24c32wp), I damaged the Resistor on R7 (which should be 0402 4.7k ohm) .
then resolid a 0603 4.7K ohm Resistor on R7 for replace , now switch report "fatal error" when start , linux and bash still work, I can login in os and bash etc.

error message are:

Code:
 chad[5883]: [chad.ERR]: check_base_header(), fru_info.c:299, build 1: Unsupported fru info version of the base block fru_ver: 0, exp_fru_ver: 0x1
chad[5883]: [chad.ERR]: st_open(), subsystem.c:8165, build 1: Auto detect of the subsystem "" failed due to MLXI2C_UNSUPPORTED_FRU_INFO_VER
chad[5883]: [chad.ERR]: check_base_header(), fru_info.c:299, build 1: Unsupported fru info version of the base block fru_ver: 0, exp_fru_ver: 0x1
chad[5883]: [chad.ERR]: st_open(), subsystem.c:8165, build 1: Auto detect of the subsystem "" failed due to MLXI2C_UNSUPPORTED_FRU_INFO_VER
chad[5883]: [chad.ERR]: check_base_header(), fru_info.c:299, build 1: Unsupported fru info version of the base block fru_ver: 0, exp_fru_ver: 0x1
chad[5883]: [chad.ERR]: st_open(), subsystem.c:8165, build 1: Auto detect of the subsystem "" failed due to MLXI2C_UNSUPPORTED_FRU_INFO_VER
chad[5883]: [chad.ERR]: mi_opend(), mlx_i2c.c:727, build 1: mi_opend retry number 3 failed due to MLXI2C_AUTO_DETECT_FAILED
chad[5883]: [chad.ERR]: wrapper_mi_open: sysapi failure:[21] on mi_open 'MLXI2C_AUTO_DETECT_FAILED' flags mask:[105]
chad[5883]: [chad.ERR]: do_i2c_init_ha(), chad.c:2122, build 1: Error code 14004 (item not found) returned
could anyone give some advise what to do next for repair the sx6012?
according error messages it seems related to i2c or fru error , maybe replace the st24c32wp or something else?


1.jpg
 
Last edited:

klui

Active Member
Feb 3, 2019
291
123
43
Are you sure you resoldered the resistor correctly? Do you have another 6012 where you can swap the chip? Hopefully the contents are still good but it's obvious the system can't read it. Hopefully it's not because your clip was mislabeled and you attached it backward or off-by-one row.
 

Freebsd1976

Active Member
Feb 23, 2018
255
39
28
Are you sure you resoldered the resistor correctly? Do you have another 6012 where you can swap the chip? Hopefully the contents are still good but it's obvious the system can't read it. Hopefully it's not because your clip was mislabeled and you attached it backward or off-by-one row.
1 resoldered the resistor should be correctly , because a hardware engineer did it.
2 this is the only one switch, if needed can find another one(need sometimes)
3 yes the clip is mislabeled , will it casue problem? and what kind of problem ? fried the chip or damage the content ?
 

klui

Active Member
Feb 3, 2019
291
123
43
3 yes the clip is mislabeled , will it casue problem? and what kind of problem ? fried the chip or damage the content ?
Reversing the pins, chip enable (e2, e1, e0) would be swapped with serial data, serial clock, and write control; along with power and ground. My guess--and I'm not a HW engineer--is the part would have shorted and the chip most likely damaged if the FRU isn't readable (VCC connected to VSS, and vice versa). I've read chips should have internal protection but since MLNX-OS can't read it and the resistor was replaced properly, it would point to chip damage. If you have access to rework engineers, have them remove the chip and place it directly into a programmer to see if it could be read.
 

Freebsd1976

Active Member
Feb 23, 2018
255
39
28
ok
Reversing the pins, chip enable (e2, e1, e0) would be swapped with serial data, serial clock, and write control; along with power and ground. My guess--and I'm not a HW engineer--is the part would have shorted and the chip most likely damaged if the FRU isn't readable (VCC connected to VSS, and vice versa). I've read chips should have internal protection but since MLNX-OS can't read it and the resistor was replaced properly, it would point to chip damage. If you have access to rework engineers, have them remove the chip and place it directly into a programmer to see if it could be read.
thanks , contact an engineer now , will report back

edit:

repaired through resolid the chip by engineer , luckly there are no chip damage or content damage .
 
Last edited: