We have a a project where there a system powered by 13s Li battery - 54.6V max voltage of pack. The buck regulator we're choosing for this has an absolute max of 76V. For this we're considering two TVS diodes

824521481 Vbr=56.1V Vcl=77.4V

SMBJ48CD-M3/H Vbr=54.1V Vcl=76.3V

In the first option quite small current will flow through the TVS, but max clamping voltage is greater than 76V rating of buck regulator. In second case there is a possibility of more than 1 mA current flowing through diode when battery is fully charged but clamping voltage is much closer to max rating of buck regulator.

Which of the two options is better? Or will both of them not work and we've to look at a higher voltage rating for the buck?

Edit 1 for more context: This is a small add-on circuit with 2G GSM modem (2A peak) being powered from the battery of an EV that has a motor size of 0.5 to 2kW and battery capacity of 4 to 8 kWh. So at 48V from battery during the GSM peak current also there is sub 300 mA current draw. We are looking at protecting the circuit from any transients or surges from the motor operation, charger connection or BMS battery cut-off during any fault. We are not yet sure of what kind of transients to expect in such a system.

  • 1
    \$\begingroup\$ Can you explain the scenario where the diodes have to clamp the voltage? Is this a motor driver that you are talking about? Is it possible for the battery to get disconnected during charge? I am having a hard time seeing how a system with a 13S battery can get to 76 V. Also, how much input current does the regulator need? If the input current is low you may be able to protect it with an RC. Edit your question to include all this information. Don't reply with a comment. \$\endgroup\$
    – user57037
    Jun 7, 2021 at 6:32
  • 2
    \$\begingroup\$ I don't think there is any way you can use the first one, the 824521481. Vbr is 56.1 +/- 5%. So the min is 56.1 - 0.05*56.1 = 53.295. \$\endgroup\$
    – user57037
    Jun 7, 2021 at 6:36
  • 1
    \$\begingroup\$ Yes if the BMS disconnects you have to disable regen very quickly. The voltage rise rate can be calculated using I = Cdv/dt. dv/dt = I/C. I is the regen current going into the battery before the BMS disconnects, and C is the total capacitance on VBAT. You don't have a lot of time. But a TVS doesn't buy you that much, either, because they can only absorb transient over-voltages. If you try to continue regen after battery disconnect your TVS will blow for sure. The motor controller really needs to be "in the loop" and disable regen BEFORE the BMS disconnects. \$\endgroup\$
    – user57037
    Jun 7, 2021 at 22:32
  • 1
    \$\begingroup\$ The motor controller generally has large capacitors. The regen current is not infinite. When the battery disconnects you may have some ms of time to halt regen before a too high voltage is reached. You can figure it out using dv/dt = I/C. One product I worked on had a braking resistor to dissipate extra energy. But another option is to simply disable braking. It could be based on voltage. Once the controller sees 60V, it should disable regen braking, for example. However, this would only work if the controller samples the voltage pretty frequently. \$\endgroup\$
    – user57037
    Jun 8, 2021 at 15:40
  • 1
    \$\begingroup\$ @mkeith A comparator feeding a break input could be used instead of high frequency sampling, and react faster. \$\endgroup\$
    – DKNguyen
    Jun 9, 2021 at 13:31

1 Answer 1


We've come up with two approaches since we don't know all the EVs that we are going to work with. So both these need to be tested to see which is better for us in various parameters. Also test with various EVs.

1st approach would be as @DKNguyen suggested to use a higher wattage TVS diode and hope the surge current would not cause either the TVS to blow or cross the buck regulator's max.

2nd would be to change the input power stage. Offline switcher ICs such as LNK3207 or UCC28881 can operate as buck regulators and can handle inputs up to 700V. A low-power high-voltage TVS diode is need with this as the area under the curve of the spike would be smaller at higher voltage, thus would need to dissipate lower energy. Also if higher current is needed for the application, it can be two stage buck regulator - offline switcher can come to 35V at 0.5A so that it can give enough power to another buck regulator to operate at higher current at 3-5V.

  • 1
    \$\begingroup\$ I think your second approach, or a variant of that is prudent. As you have discovered, TVS diodes have a big disparity between working voltage and clamping voltage. You run into trouble if the working voltage is only a little lower than the absolute max for your device at risk. Another option is to use a surge stopper device (i.e., LTC7860). A device like this will raise the voltage tolerance of the circuit overall - it will protect the buck. Now, there is large gap between the working voltage and absolute max for the surge stopper, and you can easily find a TVS for that. \$\endgroup\$
    – Troutdog
    Jun 9, 2021 at 13:38
  • \$\begingroup\$ One key point about these type of devices is that they provide precision thresholds, unlike a TVS. \$\endgroup\$
    – Troutdog
    Jun 9, 2021 at 13:39
  • \$\begingroup\$ Thanks for letting me know about LTC7860. Since we are cost sensitive, LTC7860 won't work for us, I can look for a lower cost solution like this. \$\endgroup\$
    – EarthLord
    Jun 9, 2021 at 13:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.