A farmer bought it that week for a tractor that needed a starter, and Malik felt, for a few days, like he’d figured the whole thing out. Then the farmer brought it back. The starter bolted onto the mounting pattern just fine. But once it was in place, the drive gear didn’t engage the flywheel the way it needed to.
The nose housing on that particular model positioned the pinion just slightly differently than the tractor required. Close enough to look right sitting on a bench. Wrong enough that it wouldn’t actually crank the engine once it was installed. Malik refunded the farmer’s money that same afternoon. Mr. Bell didn’t scold him for it. But he made the lesson clear enough at the shop counter that evening.
A starter that spins freely on a test bench is not automatically a starter that belongs on any given tractor. Spinning proved the motor worked. It proved nothing about whether it was the right motor for that machine. Malik took that seriously. He spent the next several weeks building an actual classification system instead of just testing whatever came off the shelf first.
Recording voltage, rotation direction, mounting bolt pattern, nose-housing type, pinion tooth count, pinion diameter, shaft length, solenoid position, and the manufacturer part number wherever it was still legible, cross-checking all of it against tractor application charts pulled from old Delco and Prestolite service books Mr. Bell kept in a back cabinet. He made handwritten index cards for each family of starters. One card per application.
Calling dealerships two counties over when a number he found didn’t match anything in the books they already had. Some calls leading nowhere. A part number too old for anyone to place. Others turning up exactly what he needed from a parts-counter clerk two hours away. He learned to check brush wear under the loop, to look at a commutator for scoring, to test bushings for play, to inspect field coils and armature windings for the telltale discoloration of a unit that had actually burned rather than just worn out.
Mr. Bell helped him rig a proper bench-test setup, checking current draw, free-running speed, and voltage drop under load, so a starter wasn’t just judged on whether it spun, but on whether it spun the way it was supposed to. One afternoon that October, Malik pulled two starters off the shelf that looked, at a glance, almost identical. Same manufacturer, same rough size, mounting bolts in what looked like the same pattern.
He assumed for a few minutes that one could stand in for the other if he ever ran short. And he was ready to write both onto the same index card as interchangeable. Mr. Bell stopped him before he got that far, had him lay the drive ends side by side and count the teeth on each pinion gear under the loop. They weren’t the same. Close, but different enough that swapping one for the other would have left a farmer with the same problem Malik had already refunded once.
Mr. Bell reminded him, not unkindly, that this was exactly the kind of mistake that could put a customer’s tractor back on Bell’s counter a second time. And a shop only got so many chances before farmers stopped trusting the work. “Bolting on and working correctly,” Malik wrote in his notebook that night, “are not the same thing.” And he underlined it twice.