2
\$\begingroup\$

I'm working on a battery powered device and would like an LED to come on when the battery is getting low. In the given schematic, the LED is a red LED having a forward voltage of 2V, Lite-On product number LTL-307EE. I've come up with the following quick and dirty design to check it out:

Comparator Schematic

The battery passes through a switch (so it's not always on), and then a 3V3 buck-boost supply to run the system. I only want the LED to come on when the system is powered and the battery is below a certain threshold (3.4V in the design below for prototyping).

I'm using an LM393 as I have some available and have made sure to attenuate the inputs to not exceed the input limits (Vcc - 2V).

My concern is that when the system is unpowered, the battery will still be feeding up to 4.2V into one of the comparator inputs. The other inputs will be floating about. The current will be very low (uA) but is this likely to cause damage? If so, is there a way to add protection to this design?

Also, is there a better option for the comparator? The input voltage limit on the LM393 is a bit restrictive at such a low supply voltage.

\$\endgroup\$
2
  • 1
    \$\begingroup\$ You can do better with discrete BJTs. The LED itself, of course, won't turn on when the supply voltage is below its own minimum voltage level. So that's free, after a fashion. Beyond that, it's possible to use 2 BJTs to set up a minimum required for one of them to pull down hard enough when the voltage is above any desired given voltage. Given BJT variations, it's won't be precision work. But it is very cheap. For precision, a TL431/TLV431 might be considered, in addition. \$\endgroup\$
    – periblepsis
    Commented Mar 25 at 16:53
  • 2
    \$\begingroup\$ The 393 is a solid part, if you treat it with respect. As you seem to do this, I expect no issues in this circuit. \$\endgroup\$
    – tobalt
    Commented Mar 25 at 17:47

2 Answers 2

Reset to default
4
\$\begingroup\$

There is no need to add any circuitry. The inputs can withstand -0.3V to +36V regardless of whether supply voltage is applied or not, and even without current limiting.

Many other comparators (such as CMOS ones) do not have this feature.

Adding a diode can cause problems such as lifting the supply voltage or inaccuracy because of diode leakage (the latter being more a more likely malign effect in your case- if a Schottky diode is used).

\$\endgroup\$
1
\$\begingroup\$

I don’t think there can be problem with feeding the comparator input with higher voltage than Vcc since there is a 250k in path. So it is going about very low current, no damage should happen.

Lm393 has inputs made from bjts (not fets), so there is no risk with damaging fets gates.

But anyway, if you worry just place a clamping diode on comparator input.

enter image description here

\$\endgroup\$
3
  • 1
    \$\begingroup\$ Such diode is likely present for ESD purposes internally. Usually TI specs 10 mA clamping current for comparable analog parts, though not for the LM393 itself, so one should be much more conservative. With 250 kO, the current would be only 15 uA, which - even as a continuous condition - will be totally fine IMO. \$\endgroup\$
    – tobalt
    Commented Mar 25 at 17:45
  • 1
    \$\begingroup\$ I see the input protecting diodes in LM393w version only. \$\endgroup\$
    – Michal Podmanický
    Commented Mar 25 at 18:01
  • 1
    \$\begingroup\$ But they all have some ESD rating for HBM, so there should be something in there ?! \$\endgroup\$
    – tobalt
    Commented Mar 25 at 18:03

Not the answer you're looking for? Browse other questions tagged or ask your own question.