Early reliability evaluations of a microprocessor-based system can be achieved by means of fault injection tools based on simulation. This is a widely used method to validate the effectiveness of fault mitigation strategies, and also to assess the reliability of mission-critical systems. However, more research is needed for a better understanding on how accurate are the results obtained by means of fault injection campaigns using simulated models of the system in comparison to emulated faults on the real processor system. In this work, we present an experimental comparison of the results obtained using two simulation-based fault injection tools against the results obtained by emulating faults on the real processor. To carry out the comparison, a series of fault injection experiments are performed on the Texas Instruments MSP430 microcontroller.