The circuit under consideration is to be powered by four 18650 li-ion cells in series. These cells are in a holder and user replaceable so there is risk of the contacts being touched and adding esd protection might be wise.
The biggest problem is a CD40109, which is powered straight from the battery. Parameters of concern for this chip are:
- Absolute maximum supply voltage: 20V
- Recommended maximum supply voltage: 18V
Four, fully charged, 18650 cells in series will provide ~16.8V (4x 4.2V)
The problem is that TVS diodes with a reverse stand-off/working voltage of >16.8V but with a clamping voltage <= 20V don't seem to exist. What would be good ways to handle this situation?
I'm thinking along the lines of the following:
simulate this circuit – Schematic created using CircuitLab
Selecting a TVS diode with a higher clamping voltage such as 30~35V and then feeding the CD40109 trough a RC. With the input clamped at 35V during a ESD event it takes ~40µs for the capacitor to charge from 16.8 to 20V with the values given, which seems enough to survive a standard 8/20µs event. The CD40109 is not driving any significant loads, so 22 ohm in the power feed is no problem. Any problems with this idea? Better ideas?
Best Answer
Don't use a zener or other TVS; not for ESD purposes alone, anyway. The bypass capacitor is sufficient.
You may find this answer I just wrote to be of interest,
Is a 100R/1nF RC enough to protect microcontroller IO pin from ESD?
by changing C2 to a larger value, and minding the same considerations mentioned there (particularly ESL which is your responsibility in PCB layout), you will find the peak voltage is adequately limited even without diodes.
The series resistor is a good idea for a different reason: hot-plugging inrush. Cells have considerable capacitance, so on closing the loop with your bypass capacitor, quite some inrush current can flow, and resonate with stray loop inductance (probably low 100s nH here, more if cable length is present). A small value like 22Ω is enough to dampen \$L \lt R^2 C\$ or several mH.
Mind that type 2 ceramic capacitor value decreases with DC bias, and the value for calculation is the total series equivalent, so will be less than the rated value, perhaps 5µF, perhaps just a couple even.