Hot-Wire constant voltage Anemometer

Since the hot wire is always driven with the same polarity, you can add an emitter follower current booster to the opamp output to provide most of the heater current. The output voltage signal is also taken from the follower. You might need a heat sink for that transistor.

The schematic in the application note is not to be taken literally, the op amp symbol is merely an articifice used to illustrate the voltage controlled regulator. And they make a few statements that make me wonder, such as the capability for indefinite coaxial cable lengths from the Vw tap to the hot wire sensor. At a minimum this will introduce a voltage divider action causing a deviation from true CVA operation, and for reasonable vales of Rw and R2 will introduce a destabilizing phase lag into the feedback loop. But then again they are only interested accurate measurement of the hot wire current out to something like

500KHz 3dB bandwidth, and they do seem to have a handle on the concept of deconvolution of the hot wire response, so maybe they can deconvole other effects, but the physics of the heat transfer does not change in that respect. There may be practical advantages to the circuit as well such as the maintainace of Vw at the sensor node whether it is connected or not.