The challenges and requirements of partial discharge (PD) measurements on long extruded high-voltage (HV) cables are often suppressed, ignored or simply unknown to HV laboratories, test facilities and grid operators. Long extruded HV cables do not represent a lumped capacitance. Therefore, wave propagation, attenuation and reflections of travelling waves along a cable caused by PD pulses must be considered. The requirements of PD measurements on long HV cables are defined in the standard IEC 60885-3. Due to the reflection of PD pulses at the far cable end, superposition errors may occur and affect the recorded PD magnitudes. While higher PD magnitudes caused by positive superposition are tolerated by the standard, negative superposition exceeding 10% of the original PD pulse value is not permitted. Since negative superposition errors are related to the PD instrument and the cable length of the test specimen, these only represent a problem for specific cable lengths. These critical cable lengths have to be determined by using double pulse diagrams. Depending on the results, additional procedures may have to be applied. Checking the double pulse behaviour of the entire test circuit is mandatory before PD testing of cables during routine tests, factory acceptance tests and on-site tests. The double pulse behaviour of the test circuit depends on the PD measuring instrument, the PD band-pass filter settings of the PD instrument, the test circuit and the characteristic impedance of the HV cable. To investigate the influences of different parameters and the feasibility of implementing the standard IEC 60885-3, several modern digital PD measuring instruments, different PD band-pass filter settings, characteristic impedances and test circuits were used. The PD measuring instrument and the PD band-pass filter settings have a major impact on the double pulse behaviour and the diagram. Several procedures are recommended by IEC 60885-3 to overcome and avoid negative superposition errors on long HV cables. These include the use of a reflection suppressor and terminal impedance. The feasibility and the drawbacks of these methods are described in detail.