hi Hero,
Are you asking me the questions in yout last post? If yes, here is my response, if not, it may help the OP.
As he is working the system over ranges/distances of 1 mtr, I don't think the velocity of sound versus temp is going to a problem.
If he calibrated at 20degC, vel sound is 343.5m/sec, so if he operated the system +/-10degC on the cal temp, he would get a change of +/-1.75mm over that temperature range, based on a maximum distance of 1mtr.
As he is using a freq of 40kHz, whose wavelength is 8.5mm and assuming at best, a resolving power of a quarter wavelength, ie; 2.1mm
I cannot see a problem.
Semiconductors usually get faster with temperature, within their specified operating range, due to the reduction in threshold switching levels. However I think it reasonable to assume that as the 'TX' delay increases with temp so will the 'RX' delay, so the temp differences should tend to 'cancel' out.
If he 'calibrates' the 'system' TXR as advised, the emitter switching times will be part of the overall response delay time.
I have already pointed out to the OP about variations from one TX,RX pair of the same type.
Eric