Research has shown that reversible phenomena such as non-linearity, self-heating and thixotropy occur during laboratory fatigue testing on bituminous mixes. This work aims to develop a test protocol to separate and quantify the recovery of reversible phenomena and the restoration of damage. Besides proposing some adjustments to the protocol, in the present paper, its repeatability is evaluated by comparing the results obtained for several samples of the same bituminous mixture. The test procedure is composed of two parts: short complex modulus tests (200 cycles at 10 Hz) at temperatures ranging from 8°C to 14°C with strain amplitudes ranging from 50 to 110 µm/m are used and five partial fatigue tests (at 10°C) are carried out at 100,000 cycles, 100 µm/m and 10 Hz. Each fatigue lag is followed by a 48-h rest period which consists of short complex modulus tests. The test protocol appears to show reliability and provide repeatable results.