There is a change in stresses also occurred in steel before cracking, but this change can be consider tiny as compared with fy value

Thank you Dr.Abbas,

I think you mean the change in stresses in steel is negligible.

There are two conditions to be considered: 1 Bonded Prestress and 2 Unbonded prestress.

Considering bonded prestress.  In an uncracked member, the  steel and he concrete will undergo the same strains.  Considering a member where the prestress is concentric with the centroid of the member and is subject to a centroidally applied tension P. The result will be a strain of e.  Considering the distribution of forces within the member, P/e=Ac*Ec +As*Es where Ac and As are the areas of concrete and steel, and Ec and Es are the elastic moduli of concrete and steel. Because the area of the concrete is much larger than that of the steel, the majority of the force will be taken by the change in stress of the concrete, even though the change in the stress of the steel is about 6 times greater (by the ratio of Es/Ec).  The argument can be expanded to cover members in bending.

Considering unbonded prestressed members, the result is identical for members in direct compression or tension..  For members subject to bending, the behaviour is different.

The change in the steel stress for an unbonded prestressed member in bending is the proportional to the total change in length of the prestressing steel. Except for large deflections of the member, the change in the length of the prestressing steel from anchorage to anchorage is quite small, and for normal members, can be ignored up to cracking of the concrete.

Thanks Mr.Gordon Chirgwin for your answer,

It seems to be almost the same for both bonded and unbounded presstressed members up to cracking limit, I.e., neglecting stresses in steel especially by the designers to simplify it.