Testing the resiliance of SP4020 and SP4021 diodes (and my function generator)

[bloguetronica]

Hi,

This is more a curiosity than anything else. I've read about the USB Killer, which is a stupid way to test for ESD, as virtually any host gets destroyed. However, inspired by the way it works, I've decided to see if the SP4020 and the SP4021 diodes on the outputs of my function generator were resiliant.

So, I've charged a 100uF, 35V electrolytic capacitor with 28V and discharged them both ways on each output. I was surprised to see that the TVS diodes did their job well, apparently not degrading the performance of the circuitry behind. Of course, this is not how an ESD test is performed, and a large enough capacitor with a low enough ESR would kill the outputs.

This test is kind of a preamble to a project that I'm making. It is basically a USB2.0 connect/disconnect switch for testing purposes, with an integrated ammeter. I'm planning to use these diodes to protect the data lines of both upstream and downstream ports (there will be also a Polyzen protecting the power on the upstream port, and a large TVS diode on the power rail of the downstream port). They are not the typical TVS arrays that try to discharge ESD over the 5V rail, which is probably why many computers get killed by the shitty killer, instead of just the USB port. Yes, many 5V DC-DC converters that are used to supply the USB ports cant cope with spikes that are injected via the TVS arrays, especially if its internal Zener goes open, and the surrounding steering diodes survive.

Anyway, you can see the schematic attached, and leave your opinion on how this device managed to survive, multiple times, and some times with visible sparks while "plugging" the charged capacitor into the connectors.

Kind regards, Samuel Lourenço