INA137: Differential line receiver input protection

Hi Paul,

Thanks for your question. I’d like to point you to an application circuit from our INA1620 datasheet (section 8.3.1). The ESD protection scheme used here can be reused in your application with the INA317:

If an ESD event hits either input, it should be controlled somewhat by the 100pF capacitors, but any residual that couples across the 47µF blocking capacitors will get shunted to the supplies through D1-D4. Adding TVS diodes for the supplies is important here to ensure the ESD event is coupled back to ground. If the application isn’t single-supply, you may add a TVS diode to each supply and then tie D3 and D4 to the negative supply, rather than to ground.

You may also look into the INA1650 as a modern differential line receiver with precision-matched on-chip resistors for high CMRR performance.

-Tamara

Hi Paul,

I would do it this way:

V1 and V2 stands for the SMD varistors or TVS to clamp ESD to signal ground. Then the ESD currents are shunted from signal ground to chassis via hybrid bonding. The ESD protection parts can have a parasitic capacitance of up to 100pF or so. Choose parts with low leakage currents, though, which can easily withstand the maximum input signal voltage.

V1 and V2, the ferrite beads FB1 and FB2 and the NP0 ceramic caps C1 and C2 form PI filters. I wouldn't increase the capacitances C1 and C2 to over 1nF, because these capacitances also present a capacitive load to the output stage which drives the INA137 input. That said, in some mixing consoles capacitances of up to 2.2nF can be seen.  Also, FB1 and FB2 need not necessarily to be ferrite beads but can also be standard chokes. Chokes of up to 330µH can be seen in some mixing consoles. But I think to filter cellphone frequencies the ferrite bead is the much better choice. It also improves the suppression of ESD.

R1 and R2 are a bit critical. For protection purpose they should be high. But high resistor values can degrade the symmetry of input circuitry of INA137 and by this ruin the common mode rejection. See the section "applications information" of datasheet of INA137. So, I would use something between 4R7 and 47R. Take 0.1% resistors, if you can. In any case you should check by measurement whether they have any unwanted impact on the common mode rejection of INA137.

Keeping R1 and R2 small is not so much of a disadvantage as it seems at the first sight, because the BAV99 can withstand huge currents, if the current flow lasts very shortly, which is the case for ESD currents. Also, the residual ESD currents through the BAV99 are heavily reduced by the action of the ESD absorbers V1 and V2 and the PI filters. So, the BAV99 only have to withstand rather little residual ESD currents.

Another issue is, when the INA137 is powered down while being driven by an input signal. In this case the BAV99 will profit from the output current limitation of the driving output stage. But even if the driving output stage isn't current limited, 5V / 47R is only 106mA, which the BAV99 can easily withstand.

In the supply voltage lines of INA137 you can see RC low pass filters. When it comes to audio I often use such RC low pass filters. The 47µF cap is a standard aluminium electrolytic. The advantage of this "huge" decoupling cap is, that the residual ESD currents flowing through the BAV99 can be fully absorbed by this capacitance. So, a zener diode clamp across the supply voltage pins of INA137 might no longer be needed.

Kai