Skip to main content
V
Vipuls
Associate II
March 4, 2025
Solved

AEK-POW-BMS63EN Short CIrcuit on Daisy Chain

  • March 4, 2025
  • 3 replies
  • 1758 views

Hello Community,

I have an observation from my attempt to daisy chain multiple AEK-POW-BMS63EN devices. 

I have made a battery pack which has two 14-cell Battery modules in series. One module is supposed to be connected to one AEK-POW-BMS63EN and send data over ISOSPI lines back to the AEK-ISOSPI-SPI Translator and then to the master, which is AEK-MCU-C4MLIT1. 

I validated the individual slaves first, and they were working. However, when I connected the ISOH line of my first slave with the ISOL line of the second slave, it created a short circuit, and the slave IC burnt. I want to know what the correct configuration is for connecting the ISOSPI lines for a daisy chain. 

I have attached my configuration in the attached photo. 

Thank you

Regards,

Vipul

Best answer by Vipuls

Hello,

First of all, thank you so much for a quick reply.

I understood yesterday that the IC was getting fried because of the creation of "ground loops" or high voltage differentials as the GND of both the modules would differ by 48V in my series connection. 

I would like to explain a few things regarding the design of my modules and ask a few doubts,

 

1. I have connected the GND_Ext to the C0, i.e., the negative terminal of the first cell. I have been able to get data from the individual module with this configuration. But now I am confused about the difference between GND and GND_Ext on the board. Are they isolated? Is there a provision for two different configurations?

I was expecting that the GND used for the communication interface (ISOSPI) and GPIOs would be GND and that the GND_Ext would only be used for the power circuit. Maybe that is not the case. 

I tried reading about it in the documentation, but it isn't explicitly mentioned, so it is still a bit difficult to understand.

 

2. We still haven't made up our minds about whether we would like to use the RSense functionality or not. Maybe we would want to leave it as is, as we are planning to have a hall effect sensor for pack current measurement. 

 

3. I might try to connect the connect only the ISOH lines of the first module to the ISOL Lines of the second module but still the two modules would be on two different GND levels if I have connected the negative terminal of the first cell to the C0 of the slave. Is there a way where I can have a isolated common GND for the communication/control line between the slaves in daisy chain and individual GND for the voltage sensing for each slave?

 

Please see if you could answer these doubts or if you direct me to correct sources. My background is automotive engineering but I am still trying to learn.  

Thank you. 

Vipuls

 

3 replies

SRomeo
SRomeo
ST Employee
March 5, 2025

Hello Vipuls,

BMSHOLDER connection works fine when x14 batteries are used, and when the current needs to be measured.
However, for custom mockups, in case less than 14 batteries or no Rsense want to be used, some of the electrical connection need to be revised.


Let me share a "simplified" version of this functional block for the two boards:

SRomeo_2-1741193083396.png

  • Individually, Check if C0 of a mockup have almost the same voltage of GND and GND_ext of the same mockup.

This will happen only if Rsense is present, the Jumper JP1 is inserted and BMSHOLDER mockup has been reproduced correctly. If not, something is missing on your custom mockup.

 

When you say "The IC burns" I'm assuming you are referring to the L9963E mounted on the second BMS node. This usually happen if too high voltage is supplied to GPIO or communication pins.
By investigating AEK-POW-BMS63EN schematic, FAULTL line, ISOL line and ISOH line are each electrically isolated from the L9963E. The only line which seems to be directly connected to L9963E is the FAULTH one.

VBAT_UP of ISOH port carries (according to your setup) 48v with respect to its C0 pin voltage so, if connected wrong, something may happen.

 

The USB cable used to connect 2 BMS nodes should make the following connection:

SRomeo_4-1741195939079.png

  • Check that the ISO cable perform the latter connection

If you are not interested in FAULT transmission between nodes, then you can just connect the two boards as follows:

SRomeo_5-1741198473314.png

 

Thanks
AEK team

    V
    VipulsAuthorBest answer
    Associate II
    March 5, 2025

    Hello,

    First of all, thank you so much for a quick reply.

    I understood yesterday that the IC was getting fried because of the creation of "ground loops" or high voltage differentials as the GND of both the modules would differ by 48V in my series connection. 

    I would like to explain a few things regarding the design of my modules and ask a few doubts,

     

    1. I have connected the GND_Ext to the C0, i.e., the negative terminal of the first cell. I have been able to get data from the individual module with this configuration. But now I am confused about the difference between GND and GND_Ext on the board. Are they isolated? Is there a provision for two different configurations?

    I was expecting that the GND used for the communication interface (ISOSPI) and GPIOs would be GND and that the GND_Ext would only be used for the power circuit. Maybe that is not the case. 

    I tried reading about it in the documentation, but it isn't explicitly mentioned, so it is still a bit difficult to understand.

     

    2. We still haven't made up our minds about whether we would like to use the RSense functionality or not. Maybe we would want to leave it as is, as we are planning to have a hall effect sensor for pack current measurement. 

     

    3. I might try to connect the connect only the ISOH lines of the first module to the ISOL Lines of the second module but still the two modules would be on two different GND levels if I have connected the negative terminal of the first cell to the C0 of the slave. Is there a way where I can have a isolated common GND for the communication/control line between the slaves in daisy chain and individual GND for the voltage sensing for each slave?

     

    Please see if you could answer these doubts or if you direct me to correct sources. My background is automotive engineering but I am still trying to learn.  

    Thank you. 

    Vipuls

     

    SRomeo
    SRomeo
    ST Employee
    March 7, 2025

    Hi Vipuls,

    1. If you take a look to the schematic, GND/GND_Ext relationship is managed by the following section:
      SRomeo_1-1741340074324.png
      Regarding the ISOSPI, it is a differential protocol, so no need to refer to GND at all.
    2. It's your choice to use or not the embedded sensing mechanism but, in case you decide to, just make sure that C0-GND cabling for each module is the following: SRomeo_0-1741339851909.png
    3. If your aim is to connect 2x14 battery packs in series (each managed by an AEK-POW-BMS63EN) to reach 96v, then it's correct that the two modules have different ground levels.
      SRomeo_4-1741341617681.png
      As I mentioned, ISOSPI is a differential protocol, so just ISOXp/ISOXm cables are needed to send commands to a BMS device via ISOline.

      For more information about BMS daisy chain please take a look to this thread.

      For any further question related custom mockups I suggest opening a separate thread.
      Best regards,
      AEK team

     

     

     

     

      V
      VipulsAuthor
      Associate II
      March 11, 2025

      Hello,

      Thank you, SRomeo, for your help. Since I will have to build a series of connections for these battery modules, for now, I will have to connect only the ISOH and ISOL lines separately. 

      I am currently working with the temperature measurement for a single module. 

      Thank you again for your help. 

      Best Regards,

      Vipuls

      Terms & ConditionsAccessibility statement