LF398-N: Input protection/ranges

Hi Phil,

The LF-398-N is a legacy product originally introduced by National Semiconductor. I have little information about the device other than a hand-drawn schematic that was likely drawn sometime in the 1970s or 1980s. 

Examining that schematic it doesn't show any ESD diodes. However, it does state in datasheet section 12.5, Electrostatic Discharge Caution, "These devices have limited built-in ESD protection." That may not be ESD current steering diodes, but instead another semiconductor structure that handles an ESD event by other means. That could be the PN junctions of transistors that connect to the various pins. When common ESD steering diodes aren't employed, a transistor's avalanche breakdown can provide ESD clamping action. Do note that the LF-398-N datasheet doesn't mention any HBM, or CDM ESD voltage levels so what it can actually tolerate may be quite limited.

What the schematic does show connected internally to the LF-398-N input pin 3 is a high resistance 37 kilohm resistor, in series with the base of a Darlington connected NPN bipolar transistor. A very large resistance such as 37 kilohm will very much limit the input current that could flow into that base circuit, or any of the other internal diode/transistor paths connected to the input pin. It does seem reasonable for that design that the input voltage can be +/-18 V when the supplies are at a different power supply voltage.

TI does not provide a Spice model for the LF-398-N. Unfortunately, many legacy devices did not have one developed. Although TI cannot vouch for the correctness of any LF-398-N Spice that may be on line, I did find that when I did a Google search on "LM398 Spice Model" one source provided a model file. You might want to give that a try.

Regards, Thomas

Precision Amplifiers Applications Engineering

Thanks Thomas,

I was thinking to add some external ESD protection and just a current limit resistor.  In my case, +/- 18V is high enough for overvoltage and if there's already a 37k internal resistor, an external current limiting resistor should be redundant, which will help my error to now have to do externally.  I'm just a bit wary of connecting things to the outside world pretty much 'directly' (with exception of ESD prot).  But if the input section is as you describe there isn't too much I should have to do!

Thanks again and LMK if I've said anything above that sounds wrong :)

Phil

Hi Thomas, hi Phil,

I have here an TI databook from 1989 which also shows an internal schematic of the LF398 (not LF398N). I also see a 36k resistor close to the input. The maximum specification of supply voltages is +/-18V, but the specified maximum input voltage range is VCC+/-. For me this means that the input voltage shall not exceed the supply voltages of LF398.

So, I think it would do no harm to add a BAV99 (or BAV199) from the input to the supply rails and to limit the diode's current by an adequate current limiting resistor.

Kai

Hi Phil,

We will certainly keep them in mind if the LF398-N datasheet comes up for rev.

Thanks, Thomas

Precision Amplifiers Applications Engineering