Several methods are nowadays used by OEM’s in order to determine engine friction through experiments to help them develop friction correlations to be used in 1D simulation models. Some of the friction measurement methods used are; Willans Line, Morse test, Teardown test and Indicated Method. Each of these methods have their own disadvantages, with some reliant on heavy assumptions. In this paper a friction measurement method is discussed which requires a conventional motoring dynamometer cell by which the engine can be motored at different speeds. The exhaust manifold of the motored 2 litre, 4 cylinder diesel engine was shorted to the intake manifold with an unrestrictive ‘shunt’ pipe which reroutes the exhausted air to the intake [1]. The shunt pipe was pressurized by an external source of compressed air to make up for blow-by losses. It is noted that the compressed air supply is thus a small fraction of what would be required if no recirculation is used. In fact a small compressor that supplied shop air to the laboratory was more than adequate. In this manner the engine could be loaded with peak in-cylinder pressures reaching those reached whilst firing, thus engine friction can be determined at any particular combination of engine speed and load. This method computes the rubbing friction mean effective pressure (RFMEP) from a subtraction of the brake mean effective pressure (BMEP) and indicated mean effective pressure (IMEP). Since in motoring both the IMEP and BMEP are comparable in magnitude to the FMEP, error propagation is kept to a minimum. Thermodynamic loss angle, cylinder to cylinder variability, cycle-to-cycle variability, setpoint stability and repeatability of measurements are presented for a 5 rpm by 4 MAP test matrix, 1100, 1400, 2000, 2500 and 3000 rpm and 1.0, 1.5, 2.0 and 2.5 bar MAP.