0
\$\begingroup\$

I am working on a SMPS and wanted to size a varistor at the input to protect the rectifier against transient overvoltage (let's say a 1 ms surge). However, by looking at datasheets the maximum clamping voltage is always related to a transient current value (because of the maximum energy that can be handled by the varistor). How can I estimate the transient current value for a given transient voltage (stated by standards for surge tests) ? Is it only related to the line resistance ?

Thanks!

\$\endgroup\$

1 Answer 1

1
\$\begingroup\$

How can I estimate the transient current value for a given transient voltage (stated by standards for surge tests) ?

You don't need to estimate it - the value will be in the relevant standard that is applied to your equipment. It's a "numbers game" because eventually (in the fullness of time) there will always be a surge that can blow the back off any surge suppressor. So, you state in your product claims that it is designed to handle such and such a surge as defined by this or that specification (this specification is usually EN 61000-4-5).

For indirect lightning (EN 61000-4-5) the output resistance can be as low as 2 ohms and as high as 40 odd ohms. The value depends on the type of port you are applying the surge to i.e. if applying to a power supply port then it's likely to be 2 ohms hence, a 2 kV specified surge could result in a 1000 amp surge current under near-short-circuit conditions: -

enter image description here


enter image description here

So, if you are trying to be compatible with 61000-4-5, you should be able to get an idea of the numbers from the linked document. If your power supply has power cabling then the effective resistance value will increase above the value stated in the standard and make life easier for your surge suppression devices.

\$\endgroup\$
8
  • \$\begingroup\$ Thank you for your detailed answer. There is still something I don't understand. Let's for example take the S14K320 datasheet (tdk-electronics.tdk.com/inf/70/db/var/SIOV_Leaded_StandarD.pdf) at page 32. The diagram shows the maximum applicable transient current as a function of pulse duration. A customer asked me if the device could handle a 1kV transient for 1 ms. The nominal voltage at the input is 400VAC. What could be the transient current value ? (without any test setup where the output resistance is fixed, let's say a voltage surge coming from the grid). \$\endgroup\$
    – Ultra67
    Commented Sep 10, 2020 at 9:00
  • \$\begingroup\$ TVS devices handle current; you can't talk about handling a transient voltage because that doesn't imply a surge impedance. There has to be a surge impedance so that the voltage is clamped by the suppressor taking a pulse of current for the duration of the event. The device you mention can take a single pulse of 100 amps for 1 ms (then it is at its lifetime limit) - that's the problem with varistors. It would handle 5 amp peaks indefinitely. \$\endgroup\$
    – Andy aka
    Commented Sep 10, 2020 at 9:32
  • \$\begingroup\$ It becomes clearer. From what I understand, how can I determine then the surge impedance? If the voltage transient comes from the grid, do I need to know the grid impedance to estimate the current peak ? \$\endgroup\$
    – Ultra67
    Commented Sep 10, 2020 at 9:48
  • \$\begingroup\$ It's a numbers game. You don't know the surge impedance; however, your product claim is that it is protected against (for example) EN61000-4-5 to a certain impulse voltage level (usually 1 kV or 2 kV) that has a source impedance of (usually) 2 ohms for power lines. Grid impedance only adds to the surge impedance and hence makes life easier on the suppressor. Worst case scenario for 61000-4-5 is 2 ohms; cabling makes it more than 2 ohms and life gets easier. \$\endgroup\$
    – Andy aka
    Commented Sep 10, 2020 at 9:53
  • \$\begingroup\$ Thank you, now I understand better your first answer. \$\endgroup\$
    – Ultra67
    Commented Sep 10, 2020 at 10:57

Your Answer

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

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