The split injection strategy holds a potential for high pressure combustion engines. One advantage of such strategy is the capability to control the heat release rate, which also implies the use of multiple split-injections with relatively short dwell intervals. Most injection rate measurement techniques require installment of the injector on a dedicated test rig. However, these techniques fail to accurately reproduce real-engine operating conditions. Using the spray impingement method, this paper investigates the injection rate of a high flow-rate solenoid injector while being operated on the engine. The aim is to have an experimental configuration as similar as possible to the real engine in terms of the acoustics and the fuel temperature within the injection system. The assumption of spray force proportional to the spray momentum is used to measure the injection rate. The spray momentum is measured while the injector is mounted on the Volvo D13 engine and connected to the in-series fuel rail and pump. A high-natural-frequency piezoelectric pressure transducer is mounted perpendicularly at 4 mm from one of the nozzle holes. The injector and sensor are contained within a specially designed collector for the injected fuel, which is maintained at atmospheric pressure and temperature. Experiments with single injection are conducted varying the Duration of Injection (DOI) from 400 up to 2000 µs. The tests with split double-injections are conducted with fixed DOI of 500 µs while the dwell time are varied from 100 up to 1000 µs. All tests are performed at the rail pressures of 500, 1000, 1500 and 2000 bar while the engine is operated at 1200 rpm. Results show that the injection rate shape of single injections is highly dependent on the rail pressure profile. With double split-injections, the rate of the second injection as well as the total fuel mass injected increases when the dwell time is shortened. Short dwell intervals boost the fuel quantities as a result of the altered needle response. Long dwell time between two equally-long injections generate similar injection rates. The injector hydraulic delay was more pronounced when dwell time was kept long enough. Overall, higher injection pressure advances the effective start of injection while retarding the effective end of injection.