When I realized that the AEM V2 Intake for the 2002 Civic Si was a "warm air" intake (it draws air from the engine bay), my expectations for massive power gains similar to those found in cold air applications (i.e. S2000 or GS-R) were deflated. In fact, looking at AEM's published values for their Si cold air application, it only produced modest gains of about 5HP (AEM CAI for 2002 Civic Si - AEM published). Looking at these results, it was difficult to see how the V2 warm air intake would produce any more power than the CAI.

Before discussing our results, we need to cover a few preliminaries. We performed our testing using a Dynapack Chassis Dynometer at Church Automotive Testing in Torrance, CA. The Dynapack Dyno is unlike the Dynojet in that the wheel hubs are directly connected to the dyno. The result is a more repeatable and consistent measurement of engine performance since it bypasses variable such as tire pressures, how the car was strapped in, and where the wheels sit on the rollers. Using a OBDII hand scanner, we monitored the intake air temperature as well as the engine coolant temperatures to ensure that for each run, we started at a consistent point. Both intake air temperatures and engine temperatures can affect the measured results. Final notes: the car was tested with 91 Octane gas and in fourth gear.

After the second pull, Shawn noticed that the intake air temperature rose significantly. He tried to bring it back down to where it was in the previous runs, but was unable to do it. After verifying that the IAT sensor was securely mounted, Shawn noticed that the radiator hose ran right under the intake tube (as mentioned in the previous section). For a temporary fix, we placed a towel between the radiator hose and the intake tube to provide some insulation. The intake temperature stabilized and we continued the dyno runs.

The dyno graph above shows the baseline run (blue curve) and the AEM V2 run (green curve). Note that the Dynapack software does not calculate a SAE corrected torque, only SAE corrected horsepower. Under 4000 RPMs, the V2 shows a mild loss of torque (2-5 ft-lbs loss). Beyond 4000RPMs the results are positive, and there's a 1-3 ft-lbs gain. The horsepower graph shows that past 4000RPMs, we measured a 3-5 HP gain with a 1-2HP loss below 4000RPMs. The intake does work, at least on the dyno. AEM also has a dyno plot of the V2 with an exhaust showing a 17HP gain. This plot caused some confusion and AEM's Jon Gleason provided this response for us.

During the development stage of the Civic Si kit, two different cars and three different dynamometers were used. One car is the AEM Civic Si, which is fitted with a cat-back exhaust, AEM fuel rail, and AEM fuel pressure regulator. On our car, the regulator was set to the factory-recommended psi setting and the injectors were factory units, rendering the rail and regulator essentially as stock replacement components. The other car was equipped with a cat-back system too, so generally speaking, the vehicles were at the same level of modification. Unlike all of our other products, we develop the V2 on vehicles with modified exhausts and try to test them with different combinations as we expect V2 customers will already have performance modifications on their vehicle. This is reflected on our published dyno, and we regret the confusion it has created. In the future all published dynos of products that were tested on modified vehicles will list the modifications of the tested vehicle. This dyno will also be updated on our website to reflect the presence of a cat-back exhaust during testing.
... After reading the forums and concluding that our website numbers have people in disbelief, I decided to post the numbers we got testing the second Civic Si. However, in an effort to prove the accuracy of this data to the skeptics, the numbers posted will consist of both the Excel graph and the raw data scanned from the printed XS Engineering dyno page. As most people know, AEM takes three consistent runs and averages them together in order to accurately portray realistic results. Since the dyno numbers I achieved with the other Civic Si were never meant to be published, nor used for actual intake development, the three runs were made with a small amount of time between each run and thus were not as consistent as they would otherwise be when we test for publication. Typically, we take more time between the runs to normalize all of the vehicle's temperatures so that they are consistent for each dyno pass. I hope that the data from this second test, coupled with the published data on our website - and the fact that we neglected to indicate that development was done on a vehicle with a cat-back exhaust - will clear up the confusion. As you can see, an exhaust plays a vital role in freeing up backpressure and when combined with a V2 will make considerably more power than a V2 system on a bone-stock Si. In fact, the testing we performed on a stock Si for an Internet performance review on Temple of VTEC was the first bone stock Si we tested a V2 on.

We already know that the Si exhaust system is restrictive. Jon's explanation certainly makes sense. The Si's midpipe has an outer diameter of 1.75"! It appears that the combination of the V2 intake and an exhaust system frees up the engine. We will look at the exhaust system in Part 4.

The loss of torque and power is interesting because one of the claimed benefits of the V2 system is that it works better across the entire rpm range. When we test the Hondata user programmable ECU, we will see if we can turn things around.