op-amp comparator

Hi,

I was doing the sine to square converter with op-amp comparator LM339.
During the course i was stuck with a problem. The configuration of the
circuit is the sine wave (+/- 10V) was given to the inv. i/p and the
reference voltage of around 0 V was given to non-inv, i/p. To bring
the hysterisis option i have added feed back resistor to the non.inv
input. The op-amp was powered with +10V and Ground.

When i was checking the input signals at inv and non.inv i/p through
CRO, the observation was my reference signal was also follows the
input sine wave during the negative signal period of the input sine
wave. Also my -10V sine wave was clamped to around -0.5V
automatically. Also i have got the high and low transition two times
per cycle of the input sine wave meaning that my square wave output
frequency was also double that of the input sine wave frequency.

Here my doubts are 1. Why was the clamping at the input sine wave had
happened during -ve period? 2. Why was my ref. input followed the
input sine wave during the -ve period alone. 3. Also why was the
double the time frequency have come?

I hope this is general basics of op-amp itself. Could any one please
clarify my basics.

Thanks in advance.

Regards,
M.Nagarajan.

 

Have you looked at the data sheet for the comparator?
http://www.onsemi.com/pub/Collateral/LM339-D.PDF

The absolute maximum input common mode range is -.3 volts
with respect to the negative supply rail (zero volts, in
this case) to the positive supply. The functional common
mode range is 0 to Vcc-1.5 volts.

There is an internal diode junction between the inputs and
the negative supply pin that turns on and conducts if the
inputs are driven negative. Look at the internal schematic
and see the base to collector junctions of the input
transistors.

When this diode turns on, lots of internal functions go to
hell, including the high impedance between the two inputs.

To make your system work, you will have to add resistance
between your +- 10 volt signal and the input and clamp that
input with a diode that conducts before the voltage reaches
-.3 volts. This would take a germanium or Schottky diode.

Of you change the negative supply rail to -10 volts or more
negative.

 

Or use a resistor network to change the input signal to be in the
right range. Perhaps a 10k between signal and input, and another 10k
between input and +10v, giving you a 0-10v sine centered at 5v? You'd
need to adjust the IN- reference too.

 

If you drive either input of the 339 below its ground voltage,
internal substrate diodes conduct, spray charge everywhere, and all
sorts of very strange things happen.

John