Because of the nature of flowing billets in a metering device, in current billet planters, excessive billet consumption, low precision and control in placing the billets on the furrow is caused and use of billets with different length becomes limited. To modify current billet metering device used on that planters, Khani (2007) based on a patent by Pupolin and Pupolin (1976) developed a prototype of a billet metering device capable of one-by-one billet picking and evaluated it on a laboratory test rig. His results showed an undesirable effect of unarranged billets fed to the secondary holding tank which attached to the one-by-one billet metering system. This research was conducted to devise a system for continuous transfer and longitude arrangement of billets coming from a main holding hopper and delivery to the secondary holding tank. A prototype of the system was designed and developed and the effect of some factors such as billet feeding rate (1.5, 2.2 and 3.0 billet/second), angle of the chassis of the system (60°, 65° and 70°) and position of outlet of the system relative to the secondary holding tank walls (near the center wall, center and near the side wall) on some parameters such as standing, vertical and oblique billets were evaluated in laboratory and "total error" parameter was calculated. In order to save time and money for modifications of the prototype, a discrete element model of the system was developed using "Visual Nastran" software. Two important physical properties of billets, i.e. coefficient of dynamic friction ( µ k ) and coefficient of restitution ( e ), which required by the software were measured by developing tow special devices. Calculation of motion acceleration of a body on an inclined surface was the base of µ k measuring device. Therefore three pairs of infrared emitter-receiver for body motion detection and an electronic processor for processing signals sent by three receivers were developed and used in the device. The base of e measuring device was to measure the impact and rebound velocities of a wooden pendulum end impacting on a concrete anvil. Then, µ s (coefficient of static friction), µ k and e of billets were measured. In addition, to obtain a criterion for validating the developed model, developed system was evaluated in another experimental plan. During this evaluation, density, center of mass and number of nods of 180 billets with length of about 0.5m were measured. Finally, a model of the system was developed and its results compared with results of the pervious evaluation of the system. Then, this model was used in modification of the prototype. The results of system evaluation showed that total error significantly had maximum value in
