The use of small indirect injection (IDI) diesel engines in passenger cars has gained popularity, as fuel economy becomes a salient feature of vehicle running costs. The broadening in the application of this engine from light commercial vehicles has placed a greater emphasis on customer acceptability. Particular concern is being given to noise levels at low loads and speeds and at idle. Coupled with this is the need to meet the more stringent exhaust emission regulations being imposed both in the United States and Europe. The IDI engine is fitted with variable geometry pintle type nozzles, which allow small quantities of fuel to be sprayed into the combustion chamber before the main injection period. Combustion is initiated by the fuel/air mixture produced during the nozzle prelift, and the remainder of the fuel burns steadily as it is injected.Deposit formation in injector nozzles (Nozzle Coking) causes partial or total blockage in the prelift range. Thus there is little or no fuel/air mixture prepared to ignite before large quantities of fuel are injected. In these conditions, spontaneous ignition of the charge causes high rates of pressure rise with a consequent increase in noise. The resultant decrease in combustion efficiency gives less complete combustion with increased hydrocarbon emissions.Methods used to assess nozzle coking vary from subjective evaluations of noise levels, to removing and flow checking the nozzles. There is a need, therefore, for a standard method for characterising nozzle coking.This paper discusses results obtained from an air flow test rig, which was used at BP Research Centre. Sunbury to measure flow rates through clean and partially blocked injector nozzles from IDI diesel engines. A method of analysing the results to give levels of nozzle blockage on a scale of 0 to 100 is described and the ability of the method to define levels of deposit formation is demonstrated.