Tree autotrophic respiratory processes, especially stem respiration or stem CO2 efflux (Estem), are important components of the forest carbon budget. Despite efforts to investigate the controlling processes of Estem in recent years, a considerable lack in our knowledge remains on the abiotic and biotic drivers affecting Estem dynamics. It has been strongly advocated that long-term measurements would shed light onto those processes. The expensive scientific instruments needed to measure gas exchange have prevented Estem measurements from being applied on a larger temporal and spatial scale. Here, we present an automated closed dynamic chamber system based on inexpensive and industrially broadly applied CO2 sensors, reducing the costs for the sensing system to a minimum. The CO2 sensor was cross-calibrated with a commonly used gas exchange system in the laboratory and in the field, and we found very good accordance of these sensors. We tested the system under harsh tropical climatic conditions, characterized by heavy tropical rainfall events, extreme humidity and temperatures, in a moist lowland forest in Malaysia. We recorded Estem of three Dyera costulata (Miq.) trees with our prototype over various days. The variation of Estem was large among the three tree individuals and varied by 7.5-fold. However, clear diurnal changes in Estem were present in all three tree individuals. One tree showed high diurnal variation in Estem, and the relationship between Estem and temperature was characterized by a strong hysteresis. The large variations found within one single tree species highlight the importance of continuous measurement to quantify ecosystem carbon fluxes.