• Resolved

ISO7321-Q1: ISO7321 issue

Part Number: ISO7321-Q1


  i have used the  ISO7321  in the FAN control and monitor,and the ISO7321-q1 use to isolate from high voltage domain and low voltage domain,the low voltage domain is 3V3,the GND is digital ground,and the high voltage is 5V,and the GND is -48V(the FAN array power supply is 0V),for the detail,please see the figure 1.


                                               figure 1

  but up to now there are at least 5 chips was damaged,the high voltage domian(the pin5,6,7,8 get together short,and the pin7 was black),but sometimes,the pin7 get together short whit pin8,

by the way,if this case was happen,the fuse on the power supply always damage.

please help me to solve the problem,thanks

  • Hi,

    Thank you for checking with us and for sharing the schematic. We have looked at it and have the following comments and questions.

    1) The ISO7321 provides high voltage isolation from the Vcc1/Gnd1 side to the Vcc2/Gnd2 side. But, locally on each side the voltage must be < 6V. That is, between pin 1 and 4, the voltage < 6 . same thing for pins 5-8, the voltage must <6V. Otherwise, catastrophic damage can occur and this may be what you are seeing. This is documented in the Absolute Maximum rating section of the datasheet (section 6.1)

    2) Is it possible to please share any additional schematics, part number of the fan module U3? We'd like to see if there are any other components that connect to pins 6 and 7 of the ISO7321.

    3) Do you happen to know if the fuse trips first or if the parts get damaged first?
    You maybe seeing a high voltage stress on pins 6,7 --> damages the part --> pins get internally shorted to supply (or gnd) --> creating high short current --> tripping the fuse;
    there is a separate high current event when the fan is plugged in/turned-on --> fuse trips and opens - -> high voltage (48V) appears on pin 6 --> stresses and damages the part due to chip abs max violation.

    4) A suggestion for you to consider: a safer location for fuse f1 would be between "Vcc_48V" and the top plate of local decoupling cap for the fan, C2. This way, C2 will be directly in parallel with the fan and connect directly to ground on one side (instead of through the fuse) and give better performance. More importantly, in the new location, if the fuse f1 opens due to high current, it will prevent the high voltage stress from appearing across the chip and protect it better.

    Thanks and regards,

  • In reply to Abhi Aarey:

    hello Abhi
    thanks for your help
    1)i'm sure the both side voltage are <6V,between pin1 and 4,the voltage is 3.3V,and the other sider ,the voltage is 5.25V
    2)for the FAN monitor and control ,this is completed schematic,there are not any other components ,just like the figure 1,and the U3 is FAN unit
    3)because the damage was done in a short time,so,i'm not sure the sequence
    4)mybe this is a good suggest ,but i can't verification,because the PCBA is completed ,and this problem is not occurs on every board and every time

    otherwise ,i found ,if the chip was damaged,and FAN must be bad too,and ,if the chip is ok,but i insert a bad FAN on the fan connector,the chip is immediately broken.

    can you give me the other suggestion ? thank you.
  • In reply to user1359349:


    Another experiment/suggestion would be to use a protection/clamp diode (6V Zener) on pins 6 and 7 w.r.t GND2. This will protect the pins and if the diodes get activated, then you can be sure that high-voltage stress is still appearing on them from somewhere.