This L5P program is built around one idea: don’t kill an engine blindly. We instrument the engine, intake, exhaust, intercooler, and dyno so we can see what is happening as power goes up. Then we log it all through Banks iDash Pro so the failure point is documented instead of guessed at. That matters because the first thing to fail is not always the engine itself. A turbo, intercooling, oiling, or cooling shortfall can end the test early and tell you nothing about the real strength of the long block. So if a support system becomes the limit, we fix that and keep going. The point is to find the physical limit of the L5P engine, with data behind it, not just make noise and scatter parts.
Gale Banks introduces an L5P Duramax mounted in the dyno cell at Max Power. This is the same basic engine used in current pickup trucks, and the purpose of the program is to determine how far the platform can be pushed. Rather than guessing at limits or relying on anecdotal failures, the plan is to document the engine's behavior methodically as power is increased step by step.
The engine has been instrumented over the course of roughly three weeks so the team can monitor it in detail. Banks emphasizes that every sensor was chosen for a reason and that each measurement is considered mission-critical to understanding how the engine survives, where it begins to struggle, and what ultimately causes failure. The goal is not simply to make power, but to know exactly what the engine is doing throughout the process.
A major part of the project is data acquisition through the iDash DataMonster. Banks says the system will log conventional readings as well as entirely new channels that have not previously been shown publicly. Those additional measurements are expected to reveal operating conditions and trends that are normally hidden, giving a clearer picture of the engine's true limits under load.
Before beginning any serious testing, Banks wants to fire the engine and confirm that the setup is healthy. On startup, he notes that the L5P sounds surprisingly refined for a diesel, especially considering it is running outside the truck with no vehicle sound insulation around it. The engine exhibits very little traditional diesel clatter, which he points out as an impressive characteristic of the platform.
Banks makes it clear that the objective is to take the engine far enough that it eventually fails, but to do so intelligently rather than carelessly. In his view, destroying an engine without understanding why proves very little. The value of this program is in documenting the exact chain of events that leads to failure so the team can identify the true mechanical limit instead of confusing it with a preventable supporting-system problem.
To reach the engine's actual structural boundary, Banks does not want a peripheral component to become the reason the test ends early. If the combination needs a larger turbocharger, it will get one. If charge-air cooling becomes inadequate, it will receive more intercooling. The same approach applies to the lubrication and cooling systems: if the engine needs a higher-capacity oil pump, a larger oil cooler, a higher-flow water pump, or a bigger radiator, those upgrades will be made. The strategy is to remove external bottlenecks so the test exposes the physical limit of the engine itself.
The end goal is to identify the point at which the L5P's core hardware can no longer survive increased output. Banks describes that limit in blunt mechanical terms: he wants to push the engine until a major internal failure occurs, whether that means catastrophic bottom-end damage or the top end being forced apart. By proceeding in measured steps and recording everything, the project is intended to show not just that the engine can be broken, but exactly how and why it reaches that point.
Banks frames the dyno program as an ongoing, fully documented development effort rather than a single power pull. Viewers who follow the series will see each stage of the process, from baseline operation through successive upgrades and escalating stress. The emphasis is on learning from the data at every step so that the final failure, when it comes, is meaningful and technically explainable.