Voltage Spike Protection

Discussion in 'General Electronics Chat' started by Joetheeskimo, Dec 29, 2015.

  1. Joetheeskimo

    Thread Starter New Member

    May 15, 2014
    6
    0
    Hi Guys

    I need some help for voltage spike protection. My product is a sensor that operates from a 28VDC supply. The voltage can vary from 18VDC to 32VDC. I have got an adjustable linear regulator (LT3010) on the front end. The regulator can operate from 3VDC to 80VDC. I have got SMLJ36V TVS diode on the front end to provide protection against transients.

    I have put the product through about 3 weeks of EMC testing and so far it has passed all RF susceptibility, Emissions, Voltage variation, 600V 10uS Spike ETC.
    However the product failed on 80Vdc 100mS spike test, i believe the TVS diode failed as its gone short circuit.

    The TVS diode specification is as follows;
    Working Peak Reverse Voltage - 36V
    Maximum Reverse Voltage - 58.1V
    http://www.bourns.com/docs/Product-Datasheets/SMLJ.pdf

    So my question is as follows;
    I need to be able to clamp the voltage at 80Vdc or less because of the voltage regulator and withstand the 600V 10uS spike.
    Can you get TVS diodes with a clamping voltage lower than the working peak reverse voltage?

    Any help would be gratefully appreciated.
     
  2. ErnieM

    AAC Fanatic!

    Apr 24, 2011
    7,394
    1,607
    Is there any limit on how much current this 80 V signal can source? Did you add any impedance in series with your device to limit the current?

    You chose a device rated for 3000 Watts for 1 ms and expose it to some huge current for 100 times that limit.

    Gee, no wonder it fails.

    Let's see if someone knows a good part before I can get to a PC and look something up.
     
  3. Joetheeskimo

    Thread Starter New Member

    May 15, 2014
    6
    0
    Hi ErnieM, thanks for your reply. There is no series impedance, i thought using a 3kW TVS diode a power resistor would not be needed.

    I dont want to change the circuit layout if i can get away with it other wise i will need to repeat parts of EMC testing. Hopefully i can just uprate the TVS diode.

    There is no limit on how much current the 80V signal can source. Not that i know of anyway.
     
  4. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    Then you have to limit the current that you pull off it. A series resistor, even a quite small value could limit the current enough so that you don't cross the max current. How much power does your device draw?
    Maybe you could use a PTC to limit the peak, but that depends on how fast it is to react and how long it takes to recover to the normal state versus the normal load current and if it can be interrupted.
    Another way to survive such overvoltage would be to use some form of disconnect. That would probably mean enough capacitance behind the diconnector to ride out the 100ms drop out if you cannot afford the device to turn off during that period, e.g. what standard you are trying to pass.
     
  5. Joetheeskimo

    Thread Starter New Member

    May 15, 2014
    6
    0
    Trying to pass DO160 test.
    Ok I will need to look at a series resistor. But I may need to look at redoing the 600V spike test.
    It's a Shame I can't just fit a larger TVS diode with a higher clamp voltage.
     
  6. Joetheeskimo

    Thread Starter New Member

    May 15, 2014
    6
    0
    My device only draws around 30mA max
     
  7. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    Lets say you put in there a 47 ohm resistor, it will limit the current to a bit less than 1A at 80V, which will be more than enough. If you use a surge-rated resistor that is also rated for the peak voltage then you should be ok even with other transient overvoltage tests I would think.
     
  8. Bordodynov

    Active Member

    May 20, 2015
    643
    188
    Effectively suppresses RFI ferrite beads.
     
  9. Joetheeskimo

    Thread Starter New Member

    May 15, 2014
    6
    0
    Thanks for your feedback on the current limiting resistor.
    If you don't mind me asking how did you calculate it will be less than 1A. When I calculated it, it works out as 1.6A.
     
  10. kubeek

    AAC Fanatic!

    Sep 20, 2005
    4,670
    804
    In combination with the 33V transil the current will be about 1A.
     
Loading...