Peter did that for educational purpose and it is fine at its place ,I dont think a discrete solution could fit my project as I have to deal with ESD also which comes ready built in 232 chips .If we use transistors I am not sure about issues like temperature stability and high baud rates 57Kbps.

Sanjeev, can you imagine that there was a time, when RS232 transcievers where all built with discrete transistors, because MAX232 wasn't borne? You would be surprised, how clever people designed their circuits...

I saw the 232 datasheet and it states that the charge pump can deliever 2Vcc both phases,so with 3.7V it is +/-7.4 Not Bad ! if we compare with +/-3V .It is given that the charge pump could accept voltages upto 5.5V but the lower side is left undocumented .But one thing is sure that according to standards it requires atleast -/+9V at the output .

Which standard do you mean? In 1964 (!) the EIA (Electronic Industries Association) formulated RS232 standard for the first time. Beside many other features, input levels and output levels were defined. At input logical low state ('SPACE') must be represented by +3V...+15V signal and logical high state ('MARK') by -3V...-15V signal. At ouput logical low state ('SPACE') must be represented by +5V...+15V signal and logical high state ('MARK') by -5V...-15V signal. Signals showing a voltage between these levels, means at input between -3V and +3V and at output between -5V and +5V, were defined to be NOT allowed, means should not change logical state.

If you now have a look at datasheet of MAX232, you will see that not even this chip is fullfilling this standard word by word, because transciever inputs are nearly TTL-compatible. So, all this discussing about 'standards' looks a bit bizarre to me...

I know it will work with 3.7v battery but I have no gaurantee so as others have stated I must not use such a quick solution even if it works !

Sorry, you don't really know whether it 'works' at 3.7V! You chip is doing something, which looks like as if it WOULD work. But this is far, far away from REAL working!
If your developement is meant seriously, stop fumbling with 3.7V, when datasheet clearly states, that minimum operating voltage is 4.5V!!! It's not professional like and the only 'benefit' is a damage of your reputation...

RS232 transcievers are used for mainly two reasons:

1. Long distance communication (>20cm!) needs a bigger 'distance' between logical low and logical high representing signal voltages. Then, a higher noise level can be accepted without eroding signal transmission quality.

2. RS232 transcievers provide a drastical decrease of slew rate at outputs (about 100 times slowlier than 74LS or 74HCMOS outputs), making it possible to work with cables, which need NOT to be terminated with characteristic impedance at both ends, what would drastically increase signal currents and supply current requirements.


Regards,
Kai