Thanks for the answer.
Yes I am using MRF, where volute and ported shroud are statinoary, and backspace, impeller and inlet region are rotating, which I connected with mixing plane interface. I tried different physical time step upto low as 2e-6 sec. but it is not working. on the other hand if I dont include backspace then it works, so now I am initialising 'the complete model' with converge 'without backspace model' and this time I am using local time scale value of 3, as it allows different physical time in different region depending on the speed of the flow. I will update if it works.
By the way I use this speed ramping method and it works for this 'without backspace model' but not for 'with backspace model', but in this study i used fixed physical time step.