The real gain here came from fixing restriction outside the long block. We straightened the turbo inlet, built larger up-pipes with gentle bends into a forged collector, and replaced a choke-point turbine outlet with a long conical diffuser and larger exhaust. That reduced exhaust pumping loss, improved cylinder scavenging, and let the engine make more power with less fuel per horsepower. The data shows why that matters. Intake improvements helped the compressor inlet conditions, but the big win was backpressure reduction. Total exhaust manifold backpressure dropped even while airflow and power went up, which means less pressure fighting the piston on the exhaust stroke. That also improved the engine scavenge ratio, leaving less residual exhaust in the cylinder during overlap and making room for a better fresh charge. At 912 horsepower, the stock-internal L5P was still alive, but the turbo was at its match limit. Compressor output only climbed slightly while turbine pressure requirement climbed harder, pushing the boost-to-backpressure ratio negative. That tells you exactly where the next restriction is: more turbo is needed to keep airflow moving without driving backpressure up.
Gale Banks opens in dyno cell number two with the current status of the Duramax L5P program. In the previous round, this stock-bottom-end L5P produced 852 horsepower. The objective for this session is to push past 900 horsepower while also reducing fuel used per horsepower.
Banks emphasizes how little has been changed internally. The engine remains stock from the oil pan to the valve covers, with no internal teardown. The external changes are limited to upgraded components such as the oil cooler, vibration damper, turbocharger, injectors, and the revised intake layout. The significance of the result, he notes, is that this is still fundamentally a stock L5P taken from a pickup truck.
To reach the new target, the team reworked the engine's external airflow path. On the intake side, they laid down the two large filters that had previously stood upright. That eliminated a 90-degree bend and allowed the turbocharger to be repositioned up and out of the valley, creating a straighter path into the compressor inlet.
Raising the turbo required replacing the aftermarket up-pipes with custom-fabricated 2.25-inch pipes using very gentle bends that merge into a 3-inch forged collector. In the process, they also eliminated the stock turbo mounting foot. Banks describes that mounting foot as acceptable for the factory turbo, but a major choke point once power is doubled.
On the turbine outlet side, the short cone adapter feeding the 4-inch exhaust was replaced with a long conical diffuser connected to a 5-inch exhaust pipe. From there, the system remains 5 inch to the exterior of the building, then expands to 10 inch up to the roof. The team also replaced a worn rubber vibration damper with a viscous fluid damper to protect the crankshaft and to support future crank-twist measurement.
Before the pull, Banks outlines the warning thresholds programmed into the iDash DataMonster system. Compressor discharge temperature is limited to 465 degrees Fahrenheit, turbocharger shaft speed to 130,000 rpm, and turbine inlet temperature, or EGT, to 1,650 degrees Fahrenheit. The dyno room is loud enough that the crew is wearing hearing protection during the run.
The pull reaches 912 horsepower. That is roughly a 60-horsepower gain over the previous 852-horsepower result, and it immediately prompts a deeper review of the logged data. Banks notes that the iDash system records up to 100 channels, allowing them to compare the revised hardware package in detail rather than relying on a single peak number.
The first area Banks analyzes is the cold-air intake arrangement. There is no Ram-Air effect in this dyno setup because the engine is drawing air from the room, so inlet temperature is essentially the same as shop air and remains fairly consistent from day to day. Even without vehicle-speed ram effect, the revised intake path improved conditions at the compressor inlet.
By laying the filters down and creating a straighter shot into the compressor, the team gained a little over a quarter pound of pressure at the compressor inlet. Removing the 90-degree bend also reduced the work being done to the air as it moved through the tubing, and compressor inlet air temperature dropped by about 3.2 degrees Fahrenheit.
That did not create a true density gain above ambient, but it reduced the density loss. The setup went from losing 4.1 pounds per thousand cubic feet to losing 2.3 pounds per thousand cubic feet, an improvement of 1.8 pounds per thousand cubic feet. According to Banks, that change alone was worth 5 horsepower. The horsepower penalty from density loss dropped from 12.3 horsepower to 7.3 horsepower simply by revising the filter orientation and inlet path, while the engine was ingesting nearly 1,400 cfm from the dyno room.
Banks then works backward through the exhaust system to show how back pressure affects engine output. At the earlier 852-horsepower level, the tailpipe contributed 2.3 psi of back pressure, the turbine contributed 42 psi, and the up-pipes plus the turbo mounting foot contributed 7 psi. That produced a total exhaust manifold outlet pressure of 51.3 psi.
He translates that pressure into pumping load on the piston during the exhaust stroke. Using the piston crown area of 12.9 square inches, 51.3 psi creates 662 pounds of static exhaust pumping pressure on the piston before it even begins moving upward. At the horsepower peak of about 3,400 rpm, the dynamic effect is even greater.
After the revisions, the tailpipe side effectively showed zero back pressure. The team verified this with both a gauge-pressure sensor and an absolute-pressure sensor. When the gauge sensor read zero, the absolute sensor read ambient pressure, meaning the tailpipe was no longer imposing measurable restriction. Banks attributes this to the long conical diffuser, the 5-inch pipe, and the unchanged 10-inch vertical outlet acting like a chimney and helping scavenge the system.
The turbine's contribution to back pressure did rise, from 42 psi to 47.4 psi, because the engine was moving more air and making more power. However, the revised Y-pipe, forged collector, and large-radius 2.25-inch bends reduced the contribution of the up-pipes and former mounting-foot area from 7 psi to just 0.2 psi. As a result, total exhaust manifold outlet pressure dropped from 51.3 psi to 47.6 psi even while power increased to 912 horsepower. That reduced static exhaust pumping pressure from 662 pounds to 614 pounds, a 7.3 percent reduction, freeing parasitic horsepower that would otherwise be spent pushing exhaust out of the cylinders.
Banks next introduces a turbocharger metric he calls the boost-to-back-pressure ratio, based on the pressure differential across the compressor compared with the pressure differential across the turbine. Rather than looking at the whole intake and exhaust system, this isolates the turbocharger itself to determine whether it is the right machine for the engine.
At 852 horsepower, the compressor was flowing 91 pounds per minute and producing a pressure gain of 46.1 psi. The turbine required 42 psi. That meant turbine pressure requirement was 4.1 psi lower than compressor pressure gain, yielding a positive boost-to-back-pressure ratio of 9.8 percent. Compressor efficiency at that point was 74.5 percent, which Banks considers healthy.
At 912 horsepower, airflow rose to nearly 95 pounds per minute. Compressor pressure gain increased only slightly, to 47.1 psi, about a 2.2 percent increase. But turbine pressure requirement climbed to 47.4 psi, a 12.9 percent increase. Compressor efficiency also slipped from 74.5 percent to 73.1 percent. That is not yet disastrous, but it shows the compressor is being asked to move more air less efficiently, adding heat in the process.
The critical issue is that the turbine now requires slightly more pressure than the compressor is producing. Instead of being 4.1 psi lower than compressor gain, turbine requirement is now 0.3 psi higher. That flips the boost-to-back-pressure ratio from positive 9.8 percent to negative 0.6 percent, a 10.4 percent deterioration. Banks' conclusion is straightforward: the current Precision turbocharger has been pushed beyond its match point, and the engine now needs more turbo.
To show how the exhaust changes affected the engine itself, Banks introduces another parameter: engine scavenge ratio. This compares intake plenum pressure feeding the intake ports with exhaust manifold pressure at the outlet, effectively measuring the pressure relationship across the engine from port to port.
At 852 horsepower, intake plenum pressure was 43.2 psig while exhaust manifold pressure was 51.3 psig. That is a negative 8.1 psi difference, or a negative scavenge ratio of 15.8 percent. In other words, exhaust pressure was substantially higher than intake pressure, which works against cylinder scavenging during valve overlap.
At 912 horsepower, intake plenum pressure rose to 44.4 psig while exhaust manifold pressure fell to 47.6 psig. Intake pressure improved by 1.2 psi and exhaust pressure dropped by 3.7 psi. The scavenge ratio improved from negative 15.8 percent to negative 6.7 percent, a 9.1 percent improvement.
Banks argues that this is direct proof that reducing back pressure matters. A less negative scavenge ratio means less residual exhaust remains trapped in the cylinder during valve overlap. That improves cylinder fill, lowers initial in-cylinder temperature, and reduces exhaust pumping losses. Using 1,600 degrees Fahrenheit EGT as the comparison point, the team gained 51 horsepower. Since 5 horsepower came from the intake revision alone, the remaining 46 horsepower is attributed to the exhaust-side and scavenging-related improvements.
The team ultimately allowed EGT to rise to 1,650 degrees Fahrenheit, which is where the engine reached 912 horsepower. Banks notes that this occurred even though injection duration was approaching 40 degrees of crank angle and the turbine was entering choke. The turbocharger is still a strong piece, but every turbo has a limit, and this one has reached it.
Even so, the result was not just more power. Fuel used per horsepower also improved. Banks measures this through brake specific fuel consumption using a fuel-flow meter and dyno horsepower data. BSFC dropped from 0.428 pounds per horsepower-hour to 0.416 pounds per horsepower-hour. That is a 2.8 percent improvement in fuel economy while making about 60 more horsepower.
For Banks, that reinforces the central lesson of the session: reducing back pressure anywhere in the system can improve both power and efficiency because it reduces the work the engine must do to expel exhaust gas.
The final part of the recap looks ahead. The factory Denso high-pressure pump carried the engine to 912 horsepower, which Banks says many people would have doubted. However, on the 912-horsepower run the team observed some rail-pressure loss, about 40 bar, even with 90 psi of lift-pump pressure feeding the Denso pump. That shortfall is part of why injection duration had to stretch so wide.
To address that, S&S supplied a set of 100 percent over injectors to replace the existing 50 percent over injectors, along with a Cummins-style CP3 conversion. Specifically, the team is replacing the Denso pump with an S&S Diesel Motorsport 12 mm stroker Bosch pump. They are also installing a Comp camshaft and changing the oil filter in an effort to reduce pressure drop in the lubrication system.
Banks expects the new pump and injectors to provide the fuel volume needed, the camshaft to process more air per cylinder cycle, and the larger Precision turbocharger now being developed with Dan at Precision to deliver greater mass flow with less back pressure. The plan includes trying multiple turbine housings to tune shaft speed, mass flow, and A/R ratio. With those changes, the next target is clear: push this L5P beyond 1,000 horsepower.