LMV393 comparator damaged? input with voltage before V+

Hi Dan ,

I have moved the post to the appropriate forum for support.

Hi Daniel,

How is the 393 damaged? Shorted supply pins? Shorted input pins? Does it still "kinda" work with reduced output swing?Shorted output pin? If so, to which supply line? Does it get hot?

What is the load? Is it inductive?

The input should be able to handle 230uA.

If you expect the input will always have voltage before the supply comes up, then add an external Schottky diode (BAT54 signal type) between the input and V+ to take the brunt of the current (the Schottky has a lower forward voltage than the internal ESD diodes, so it will conduct first).

If you suspect supply transients, add a 5.1V zener across the supply.

Can you provide a schematic?

Regards,

Hi Paul,
Thanks for answering so quickly. Haven't a clue how "exactly" it was damaged; it was working for months and then... Can't seem to attach a schematic through the forum so a simple description:
The input is monitoring the presence of a 28V power supply through a simple resistor divider where the comparator (with an external hysteresis) is to trip when 25.4V (on rise) and 23.5V (on fall). The IN- reference is set for 3.3V (LMV393 powered by 5V). At 28V IN+ should read about 3.8V but now it read about 3.1 (lower than the 3.3V IN-), indicating the input impedance somehow has dropped.
Regards,
Dan

Hi Dan,

You cannot paste directly into the editor. Click on "Use rich formatting" at the lower right under the reply box. A new editor will appear.

Click on the "Edit/Insert Media" icon (with the "+" to the left of the paper clip on the lower tool bar). You can then upload the graphics file.

Regards,

Hi Paul,

Schematic inserted.  The Vext is the 28Vdc power mentioned earlier.  This can regularly be active about 30ms before the +5V is present.

Regards and thanks,

Dan

Hi Dan,

I do not see any immediate issues with the circuit. Although you are close to the common mode limit (3.5V) at room temp , and over it over temperature (3V). But that would not cause damage.

The input has enough series resistance to limit the current to a safe level.

What about the 3.3V ref line? Have you monitored it for any spikes? Is there any series resistance there?

The inputs and output can handle transients up to 32V even without a power supply (the LM393 does NOT have ESD diodes on the I/O pins). But it really does not like voltages below ground since there is no controlled path (ESD diodes) for negative voltages and the current goes through the bottom of the die - and that CAN cause damage.

The LM393 is a fairly tough device. I would look towards transients on the power supply and reference line, particularly transients below ground.

Does anything else share the output line that could pull it below ground? Same for the reference line?

Use a scope and watch the entire turn-on and turn-off sequence - and even wait a few seconds afterwards for "thumps", "chirping" switchers and other events that can happen as supply caps slowly discharge.

Is the layout tight? Is tehre any "noisy' circuits near by (switchers, motor drivers, relays/coils, etc) that may radiate EMI or magnetic fields?

Regards,

Dan,

The problem I see is the hard short for the inverting input to 3.3V supply.

The LMV358 would be greatly offended (high input current) if the 3.3V supply was higher then 5V supply at any time.

Power up, power down, or supply fault.