On June 17, 2005, about 2310 Pacific daylight time, a Beech C35, N5813C, experienced a loss of engine power and collided with an automobile during a forced landing on a freeway near Pasadena, California. The student pilot/owner was operating the airplane under the provisions of 14 CFR Part 91. The CFI sustained serious injuries, and the student pilot/owner sustained minor injuries. The airplane sustained substantial damage. The personal flight departed Red Bluff, California, about 1930, with a planned destination of Fullerton, California. Visual meteorological conditions prevailed, and a visual flight rules (VFR) flight plan had been filed, but it had not been activated. The approximate global positioning system (GPS) coordinates of the primary wreckage were 34 degrees 8 minutes north latitude and 118 degrees 9 minutes west longitude. The student pilot stated that the flight originated from a private airfield in Independence, Oregon. She had just purchased the airplane and was flying it home. She asked the CFI to help her get the airplane home since she was a student pilot. They took off from Independence about 1530 with full fuel. They encountered winds en route that were stronger than forecast, so they changed their plans, and landed in Red Bluff for fuel. They pumped 36.1 gallons of 100LL from an unmanned pump. She obtained an updated weather brief in Red Bluff. They departed Red Bluff on the right main tank, and then switched to the auxiliary fuel tank. They followed the 99 freeway to the 5 freeway to the 210 freeway. The ceiling at Fullerton was 3,000 feet; they started to descend around Valencia. They were about 2,800 feet mean sea level (msl) over the 210 freeway near La Canada. They had been on the auxiliary fuel tank for about 50 minutes, so they switched to the left main fuel tank, which was indicating full fuel. The right tank was in the yellow range on the gauge. The airplane was in the vicinity of the Rose Bowl, when the engine started to lose power. The student pilot turned on the electric fuel pump, but it had no effect. She transferred control of the airplane to the CFI, and began to operate the hand pump. The engine speed increased as she pumped, but fell off when she stopped. The CFI had her switch back to the auxiliary tank. When this had the same results, she switched to the right tank. The right tank was still at the top of the yellow band. However, the engine only produced power while she operated the hand pump, and at a low rpm. The airplane had lost altitude, and was down to about 1,000 to 2,000 feet over the 210 freeway. She couldn't keep pumping the hand pump due to fatigue. She and the CFI looked for a place to land, and found the 134 freeway. They turned on to final about 300 to 400 feet. About 10 to 15 feet above ground level (agl), they flashed the landing light to warn drivers on the freeway and it appeared to her that "all lanes opened up." She was going to operate the brakes on landing; the CFI had the control wheel. Within 2 to 3 feet of landing, she noticed a car to the left of the airplane. The left wing hit the bottom part of the car, and the "left wing stayed where it was and the plane flipped over." After exiting the airplane, she noted that fuel was leaking from both wings; the anti-collision lights and strobe lights were still operating. The student pilot stated that she used the electric fuel pump as a primer to start the airplane at Redding. Fuel pressure went to 9 psi, and she didn't use it again until the accident. A National Transportation Safety Board investigator interviewed the CFI. He said that the student pilot started pumping the wobble (hand) pump, and turned the auxiliary fuel pump on. The engine started to catch. The engine would only run for 3, 4, or 5 seconds after they stopped pumping. They tried all tanks to no avail. The student pilot was unable to continue pumping due to fatigue. They were down around 500 feet agl, and they decided, "we are going down." The student pilot gave the control wheel to the CFI. The student pilot would work the brakes once they were on the ground. They looked around, and decided to land on the 134 freeway as it was well lit. About 1/2 mile from their landing point, they dropped the gear, and turned final. As they got closer to the ground, there was a gap between the cars. The cars in front of their flight path were well ahead, and the cars behind slowed down. In the flare, just feet above the ground, the student pilot noticed a car off to the left. The left wing struck the car, and the airplane apparently cartwheeled, ending upside down. He closed his eyes as they hit the car, and opened them again as his face was sliding down the pavement. His door was no longer on the airplane, or it opened during the accident sequence. He struggled to get out of his seat belt and exit the airplane. A passerby broke the window on the left side of the airplane to help the student pilot get out. The CFI indicated that they tried the wobble pump for about 2 minutes, they did not declare an emergency, the fuel gauges indicated 3/8 in the right tank, 5/8 in the left tank, and gas was gushing out of the airplane following the accident. When asked what he thought the problem might be, he initially thought that it might be the fuel selector valve, but then he thought it might be the fuel pump. He said they had plenty of fuel. According to first responders, an 8-foot puddle of fuel was under the airplane. The FAA and Teledyne Continental Motors (TCM) were parties to the investigation. Investigators from the Safety Board and the parties examined the wreckage at Aircraft Recovery Service, Littlerock, California, on July 27, 2005. Investigators removed the engine. They slung it from a hoist, and removed the top spark plugs. All spark plug electrodes were clean with no mechanical deformation. The spark plug electrodes were gray, which corresponded to normal operation according to the Champion Aviation Check-A-Plug AV-27 Chart. Plugs no. 2 and 4 were light gray. A borescope inspection revealed no mechanical deformation on the valves, cylinder walls, or internal cylinder head. Investigators manually rotated the crankshaft with the propeller. The crankshaft rotated freely, and the valves moved approximately the same amount of lift in firing order. The gears in the accessory case turned freely, and they obtained thumb compression on all cylinders in firing order. Investigators manually rotated the magnetos, and both magnetos produced spark at all posts. The oil sump screen was clean and open. The oil filter was clean. The carburetor finger screen was clean. Investigators used a handheld drill to operate the fuel pump, and observed output flow volume that varied in proportion to the speed of the drill. The electric fuel pump did not rotate. The IIC observed a teardown of the pump at its manufacturer's plant. The roll pin inside the pump had sheared into two pieces; one piece lodged in the vanes, which prevented them from rotating. The technician that disassembled the pump remarked that this usually occurred when the pump had been run while dry.