0
\$\begingroup\$

I'm looking at a few 4 channel ESD diodes for my USB circuit, specifically the ESD0524P and the PESD2ETH-X. For my circuit, I only need 2 channels, D+, D-, and a channel for VBUS.

Looking at these two diodes, the PES has a Zener for the VBUS line, while the ESD0524P does not. However, I initially connected VBUS to one of the ESD0524P I/O lines which are just normal TVS. Thinking about it now, I feel like this might be wrong.

Can I connect the VBUS lines to the I/O channels of a normal TVS or is this the completely wrong approach? In using either chip, my VBUS line would be connected in parallel to the ESD diode.

\$\endgroup\$
1
\$\begingroup\$

USB is positive signal voltages only. Vcc maybe On or OFF depending on the controller but it is positive if ON. TVS (Transzorbs) are bi-directional and meant for very short transients of either polarity.

Zeners are often better when signals and power are of a single polarity. Zener diodes are built with more junction mass to tolerate more than just surges, they can clamp continuous over-voltages as long as current is limited. Also the reverse clamp is less than -1 volt and forward clamp should be close to 5 volts, though for some head room a 6 volt clamp is common. TVS diodes have the same clamp voltage in both directions, exposing the IC pins to as much as -6 volts.

To protect power the zener should be more beefy. There is no capacitive loading issue on power feeds so a dedicated SMD zener of 5.1 volts to 6.2 volts can be used. On the data lines you need to use ultra-low capacitance TVS diodes or zeners, or you will not get USB 3.0 full speed to work.

\$\endgroup\$
3
  • \$\begingroup\$ Thanks for the answer! Would I need a bi-directional Zener for the VCC line then? I don't expect current to ever flow the other way, so I'm thinking that a unidirectional should be enough. I also suppose it should conform to the standard 15kV/8kV tests like the ESD chip itself? \$\endgroup\$ – Felix Jen Jul 23 '20 at 4:49
  • \$\begingroup\$ Nothing wrong with TVS diodes which are like back-to-back low current zener. Conformity is important if predictable behavior is required. \$\endgroup\$ – user105652 Jul 23 '20 at 5:08
  • \$\begingroup\$ Awesome, I don't expect to be drawing >500mA anyways, so I guess I'll just stick with the TVS ESD0524P chip I was already using. \$\endgroup\$ – Felix Jen Jul 23 '20 at 5:15
1
\$\begingroup\$

The ESD0524 part is rated up to 5V. So based on that only, it should not be used for VBUS which might be up to 5.25V. Having said that, the datasheet has breakdown rating of 6V and the example uses the part for HDMI 5V supply and Hotplug protection, which are also lines that are allowed to go at least up to 5.3V. So while it is meant primarily to be a low capacitance ESD protection for high speed data lines, it can be used for power line protection, it just is not optimal for it.

\$\endgroup\$
1
  • \$\begingroup\$ interesting, I completely forgot to look into V_standoff. I probably should be spec'ing for V_{standoff} instead of V_{breakdown}, but I wonder how much of a functional difference that would actually make. \$\endgroup\$ – Felix Jen Jul 23 '20 at 5:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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