0
\$\begingroup\$

I have a dual-channel TVS diode from LittelFuse SM24-02HTG.

In its datasheet, a maximum value of 7A is indicated as Peak Pulse Current.

Does this current refer to the single TVS diode in the component or to both?

If it were on both then it would be 3.5A on each of the diodes?

Thank you

\$\endgroup\$
6
  • 1
    \$\begingroup\$ I would say each diode. As to whether the device would thermally survive is another question. \$\endgroup\$
    – Kartman
    Commented Nov 18, 2022 at 13:17
  • \$\begingroup\$ @Kartman Thank you for your reply. I too would say for the single diode but in the datasheet unfortunately I don't see any reference to this. The thermal issue is a separate topic. \$\endgroup\$
    – Alinik
    Commented Nov 18, 2022 at 13:21
  • \$\begingroup\$ 7A each diode. If in doubt, contact the manufacturer. They have teams of people dedicated to support. If you want to use their part, then what they tell you holds more weight to what any of us says. \$\endgroup\$
    – Kartman
    Commented Nov 18, 2022 at 13:31
  • 1
    \$\begingroup\$ Please do take note of the target application and thermal data in the datasheet. TVS sometimes get confused or misused as a sort of bi-directional Zener diode for clamping circuit operating currents, or unexpected stuff from power supplies or mains, which they're not. As per it's datasheet, they're designed for dissipating ESD, electrical fast transients (EFT) and lightning induced surges into cables/connectors i.e. infrequent and very brief discharges. Their capabilities and characteristics are centred around that. \$\endgroup\$
    – TonyM
    Commented Nov 18, 2022 at 13:40
  • 1
    \$\begingroup\$ Good question, because the image on the first page shows them in series. I'd imagine 7 Amps should be physically survivable in series for a very short time. The datasheet specs max 400 peak Watts, and Vc for this model as 42V*5A = 210W so they seem to be conservative. In reality, ESD / EFT events are going to be faster/less current than this. And if a pin is ever connected to 120VAC or gets a lightning surge, nothing in this category is going to save it. \$\endgroup\$
    – rdtsc
    Commented Nov 18, 2022 at 13:43

2 Answers 2

1
\$\begingroup\$

This is for each diode.

A simple way to get convinced is to check at the VCL values for 8/20 µs waveform just above: datasheet crop

If it were for both, it means each diode could withstand only 3.5 A (destruction beyond this value). So they could not give the VCL at IPP=5A for Pin 1 OR Pin 2 to Pin 3.

I hope it clarifies.

\$\endgroup\$
0
\$\begingroup\$

Each.

The 8/20us surge isn't long enough for heat to diffuse over the die area, or across the lead frame anyway. It's fairly safe to assume this is a thermal limit and will not be exceeded if both devices hit the same waveform simultaneously.

I can imagine some possible exceptions, however:

  • Monolithic dual. If they share temperature much quicker than separate dies, they may heat each other up, reducing total capacity in the double-reverse-surge direction. That is, using the two channels in parallel, will still handle a higher peak current than a single channel, but probably less than double. If this is the case, mind. Most likely the two channels are individual dies. But who knows.
  • Shared ground return path. The peak current limit could be driven by bondwire dimensions, and if both currents flow through the same bondwire(s), the wire itself might blow in the process.

These aren't really obvious, short of decapping or x-raying the thing. You can always ask the manufacturer for clarification.

Note that, in bidirectional mode, one device is forward-biased, so has a much lower voltage drop and power dissipation. They don't specify forward surge characteristics, but it's safe to assume it's at least as robust. It's not actually clear offhand what bidirectional voltage drop would be; presumably it's more than the single case, but by how many volts, isn't apparent. (Another good thing to verify with the manufacturer!)

\$\endgroup\$

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.