Any Audio interface (not soundcard) will provide accurate phantom power...
I'd recommand Presonus or Focusrite to start with
What is your mic what do you want to record ?

You don't give the make and model of the microphone you are trying to use, but from your report that it needs 4.5V in order to work I can tell that it's likely to be a domestic type that needs plug-in power rather than the professional type that needs 48V phantom power. You would damage this mic if you managed to plug it into a professional interface and turned on the 48V phantom power.

I don't know what the hobby shops are like in Chennai, but you may be able to find one that sells an in-line plug-in power unit that runs off batteries. The current needed is only a few milliamps, so even button cells would last a reasonable time as long as you remembered to switch it off when you are not using it.

Besides the various types you can buy from audio hobby shops or from Ebay, you can make your own if you are reasonably proficient with soldering and some construction work. Try Googling "microphone battery box" for ideas. Even if you were to get something like the common Vivanco type that takes a 1.5V AA cell (shown [[url=http://[/URL]="http://ecx.images-a…"]here [/]="http://ecx.images-a…"]here [/]with a tie-clip microphone), you can tape a couple more batteries to the box and re-route the internal wiring to get the 4.5V you are looking for.

Or buy a mic that works on 2.87V.

Are you saying you've got a sound card with a typical 3.5mm TRS audio input that provides 2.87 of Phantom Power?

I'm skeptical, but if that's the case, it's a poor design. TRS connectors are a terrible way to provide phantom power because of their potential to short-circuit as you plug and unplug things.

If you're at all serious about getting good results, save yourself a lot of headache and get a proper sound interface that you can plug into a USB port as mentioned earlier.

Best of luck!

Plug-in power (PiP) is for powering small condensers like the ones used with computers. It's typically low current 3-5VDC that can ride on the audio conductor or on its own conductor.

Any day I can learn something new is a good day, thanks.

Boulder has it right - plug-in power is not phantom in the sense that it's not invisible to the signal. In fact it is part of the signal, forming the 0Hz component of it, but flowing the other way. The coupling capacitors that you have to use are what filters out the audio (what can be heard) from the d.c. component (which powers the impedance converter).

For those of us who work more in instrumentation than in audio design (the instrumentation business has not shrunk in the way that audio has in the last few years), signals start at d.c. and go up as far as you need to go. Trying to use those nice audio recorders for instrumentation work brings its own set of challenges in how to record the d.c. - 20Hz frequency band. Some of you may remember the couple of threads we had here last year about recording paranormal happenings using audio gear, and I remarked how similar these probably were to work I had done on recording glacier movement, which is anything but linear. Managing to produce equipment that captured d.c. - 500Hz recordings on a miniature cassette recorder using minimal additional power was one of those moments that makes design work a rewarding delight.