Prahlad,

I am quite sure that the on board LPC932 is also not getting full 3V3 supply. It is going to suffer from same problem as your device under programming (on ICP header).

Now coming back to your original problem, you can use the following circuit to control VCC of the device under programming.

If you are successful, I leave it upto you to find the reason and post back here. If not, post back the result.

Good Luck

PS: I see some other problems in the circuit. For example IC2B is a HCT NAND gate driven by +12/-12V from COM port (although you have 47K in series) and this gate turns on Q1 which controls VDD to LPC932A1 (having prog90x.hex).
Why not use a spare gate in MAX232 to convert this signal to proper level and then use it as VDD control?