Clamping negative voltage on comparator input (LM393)

Question

I need to compare a AC voltage against a reference voltage. I'm only interested in the positive part of the cycle so I need to clamp the negative part of the cycle to protect the comparator input. I need some degree of precision so I'd prefer not putting a diode in series with the AC signal. The first thing that comes to mind is a simple diode clamp:

This will clamp the input to approximately -0.65V. However, the LM393 datasheet says -0.3V is lowest voltage allowed on a input, stating in the notes that going any lower may cause the output to behave erratically due to parasitic action:

1) Is there a simple, common way to solve this problem?

2) I've seen the above solution with a simple clamp diode used in commercial devices. Does this mean the problem is not as bad in reality and the datasheet is just overly protective or the commercial product was just poor design ?

Answer 1

You could use a Schottky diode, at the currents you're running it should stay in the 0.3V range. Alternatively (or additionally) you could put a resistor from the positive input to ground, creating a divider that will keep the input in the recommended range. This will also reduce the voltage seen on the positive half of the wave, so you may want to modify your REF accordingly.

Answer 2

1. You can use a Schottky diode. There is a lot of leakage, especially at high temperatures, with a Schottky so calculate that. As mentioned in 2. note that the forward voltage drop of the Schottky will go down as the susceptibility of the input goes down at high temperatures, so the -300mV is conservative.

2. You could use an ordinary diode as you have shown but add a resistor from the non-inverting input to ground. If you divide the input by, say, 2.5:1 then you'll have no less than -0.26V at the input. Note that the voltage drops will track with temperature and their -0.3V limit applies over the temperature range, so it's a bit conservative. Of course you would alter the reference voltage by a suitable factor. This will have some effect on the response time of the comparator and noise immunity.

3. Less general, but applicable to your specific case, perhaps, you could use a BJT as follows:

^{simulate this circuit – Schematic created using CircuitLab}

The transistor part number shown has a relatively high Vbe breakdown and gain. An ordinary NPN BJT could be used but the input must never exceed Vref + 5V for most types (including when Vref = 0).

Answer 3

You could use a precision rectifier circuit (OpAmp + Diode) in half wave rectifier configuration.

^{simulate this circuit – Schematic created using CircuitLab}

Answer 4

I overlooked the obvious and realized one could clamp from the other direction with a zener diode. For example, if VCC is 12V and one clamps to VCC with a 10V zener then the comparator's input voltage would never drop below 2V. The reference voltage to compare against is above 2V so it does not interfere.