ESD on SUPERVISOR POWER IC

Hello everyone!

While doing electrostatic discharge (ESD) testing, it seems that noise is coupling into the MR pin
of the TPS3823 causing my microcontroller to be reset. I took a 0.22 uF capacitor and soldered this between pins 3 (MR) and 2 (GND) and now I am not seeing the unexpected resets during an electrostatic discharge.

Here is a screen shot of the circuit:



I was wondering if there is a better value to use for the capacitor? The 2k resistor (R200) and the added
capacitor should give a low pass cutoff frequency of:

f = 1/ (2 * pi* 2k * 0.22 uF ) = 361 Hz

1.) When solving ESD problems how should I be thinking about the low pass filter here? Should I try to move as close as possible to DC? (0 Hz for the cutoff?)

2.) What about the impedance of the capacitor? We know that the impedance for a capacitor is

Zc = 1 / (j*w*C)

As frequency gets larger the impedance of the capacitor decreases routing noise to ground.


If I had my choice between difference values of capacitors such as 0.68 uF, 22 uF or 47 uF what would be the best
selection for this problem I am having with the TPS3823.

Thank you.

Hi,

adding a capacitor in parallel to your reset-switch is favourable and is commonly done to debounce the switch [1].

As the current rise of an ESD test is in the range of nano seconds you should consider the impedance of your capacitor, as it is small at high frequencies. The caracteristic test current curve for the human body model (HBM) lasts about 100 ns [2], which leads to a (idealized) frequency of around 10 MHz. A capacitance of 0.22 µF represents an impedance of about 75 mΩ at ~10 MHz. Further, your 2 kΩ acts a current limiter, limiting the current and thus the energy flowing towards your IC.

I'm not sure if there is a "how-to" recommendation how to choose this capacitor, but in gereneral a higher value with a low ESR will be beneficial. Further, it highly depends on your performed test (placing of your ESD gun) and layout by means of a surrounding (non-isolated) ground plane.

[1] https://hackaday.com/2015/12/09/embed-with-elliot-debounce-your-noisy-buttons-part-i/
[2] https://www.onsemi.com/pub/Collateral/TND410-D.PDF

greets

I would expect that 1nF up to 10 nF are sufficient to achieve the intended effect. Just try.