This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LM5163: Buck IC fails sometimes during hot plug

Part Number: LM5163
Other Parts Discussed in Thread: BQ76952, TIDA-010247

Hello everyone!
We have developed a BMS with 2 BQ76952 ICs, using the TIDA-010247 reference design. We are using LM5163 IC as a buck converter for the supply. We have observed that the buck IC fails sometimes (roughly 1 in 20) during hot plug of the cell voltage connectors. I'm attaching the schematic for reference.

We have made the following observations:

  1. We are using a 100V part for this. We even captured the voltages on the DSO when the buck converter failed. VIN at the time was around 80V and it overshot a bit like 85V, but never exceeded the 100V limit. The reference design suggests a 120V part of the same family. We figured that this is because the design is intended to be used with higher voltage applications. Our max voltage will never exceed 85V, should we also use the 120V variant?
  2. Sometimes when the buck fails, the ferrite bead on the VIN pin also blows up.
  3. TVS diodes on the VIN input don't fail, suggesting that the overvoltage conditions didn't occur. Is that right?
  4. The issue is not consistent and only happens rarely. But is recurring enough to be a problem for us when we deploy our BMS in the market.
    Thanks in advance!
  • Hi

    The absolute maximum rating for Vin of LM5163 is 100V. 85V Vin should be ok. But can you please monitor the Vin when hot plug, will >100V transient vin happens?

    As seen from the SMBJ85A Datasheet, the maximum clamping voltage is 137V, so the TVS don't fail doesn't mean the overvoltage doesn't happens.

    Can you please try placing TVS with lower Vclamp? like the SMBJ78A

    Thanks and Best regards.

  • Hello Gui,
    Thanks for your prompt reply. We have captured the waveform when the hotplug happened and IC blew up, Even in that case the battery voltage was 80V and the transient peak didn't go above 85 V. 
    On side note, Even we are considering using SMBJ78A. I'm a little confused about the fact that our max pack voltage can go up to 85V. If we use SMBJ75A, won't the TVS will always be conducting a small current (<1mA)? How will it affect the performance of TVS diode during actual transisent? Also, Clamping voltage of SMBJ85A is also 126V. Does that mean we can't ensure that it will always clamp the transient to a value below 100V?
    My apologies for the basic question, but I can't seem to find any practical guidelines around this.

  • Hi

    when you see the transient peak didn't go above 85V , you mean VIN PIN of LM5163 or input port?  can you help monitor the VIN PIN voltage? and do you have a chance to add a bulk cap 22uF 100V with C31 see any improvement? 

    I see you have a EN control signal, whether the device will failure when EN is low(Device is off)


  • Hello Daniel,
    Yes, I meant the transient voltage on the VIN pin. The DSO probe was connected to the VIN pin at TP53 of the schematic. 
    Yes, we can explore adding a bulk cap of 22uF. 
    I did not quite understand your question. I'll try to provide an explanation for the EN control part.
    The EN pin is connected to a 3.3V signal, which is coming from the Battery management IC (BQ76952). The EN signal gets active after around 50 ms of connecting the battery to the BMS. The VIN signal gets active instantly after connecting the battery. So we can say that there is a 50 ms delay between the VIN signal and EN signal. When we send a shutdown command to the BMS, the EN signal gets low while the VIN signal is still at 80V. This shuts down the BMS and the buck port and allows us to save power. This functionality is tested thoroughly and hasn't caused any issues yet. The only time LM5163 might get damaged is when we connect the battery to the BMS (Hot-plug). This is also rare.

  • Hi

    TP53 is the right position for testing the PIN VIN voltage. 

    1. For EN question, if En signal get active after 50ms, which means the device is not working when LM5163 hot plug into the battery. Then if LM5163 still damaged, highly possible it is overvoltage, the spike may only couple us, DSO probe need using 10us-100us/div with full bandwidth.