Dear Vishal,
It is a bit late but I see that no-one answered you.
For Qmaxinst it is a good idea to select this once for a series of related scans (as in your case) and to not change it. I usually look at the entire data-range, try and find the highest Q-value (looking at F(Q)) where I don't see any large glitches (sharp rapid changes in the data that are small edge effects on the detector but greatly amplified by the PDF processing) below it. This is usually somewhere around 28-30 on our experiments with 60-70 keV x-rays and a Perkin Elmer detector but will vary with the setup. Once this is set, fix it and don't vary it.
For Qmax, you select this based on your scientific question. If you are looking for small signals, you want less noisy data and set it lower typically, if you want higher resolution and can tolerate more noise, set it higher (If you want high resolution and low noise, you may have to measure longer!). When you are comparing scans with each other it is always advised to compare them when they have been processed with the same Qmax.
So in your case, you could set Qmax to capture all the signal at high-T. You will lose resolution on your lower-T data but have lower noise data. Or you could set it to capture all the signal at low-T. You will have more noisy data at high-T (the noise will be the same but your signal is less so your signal-noise is worse for the high-T data). But in this case you won't be throwing signal away at low-T. You could also set Qmax differently for each temperature to get the "optimal" signal at each temperature, but then be careful about comparing features across different temperature scans. Of course, you can do both or all three and use the differently processed data to answer different questions.
Summary: Always decide what scientific question you are trying to answer and design the analysis guided by that.
S