thanks again,as per mr eric scheme i did one thing...i gave a fixed 2 millisecond pulse i,e 20 % pw,and at the end of this pulse i took a sample butthings remained same...
ok one more information >>> the same winding+bridge rectifier+.1 cap arrangment also feeds a lm 317 voltage regulator which is used to charge small battery,the 317 is enabled only when mains is present by a npn transistor + controller pin.so as mr kai sugeested that due to a fix load there might be a problem like this..so if while taking a sample this 317 is enabled then it will act as a load,i,e in pseudocode >>>
1. enable the 317.
2.wait for 100 microseconds.
3.get a sample.
4.again disable the 317.
so any comment?
SAIF ULLAH KHALID.