After having the car for a long time in the garage and after fixing an issue I had with my overpipe (broken stud), I did a few more road dyno runs. Following are the findings ...
Here is a factory dyno comparing the newer and the older header design. The red torque curve is of the +'17 MT header and the blue one of the older. Notice that there is a noticeable bump in performance in the area of 5500-6700 rpm:
Here is a 3rd gear virtual dyno graph with a similar increase at same rpm's (marked in red boxes):
The torque curve is pretty much similar and accurate. Even the small decrease in two points @6000 rpm is because I had to change lanes during the run. Otherwise, the curve would remain same up to 6700 rpm. So, what are the conclusions ...
First of all, you can get the complete gains of this change without a re-tune of the engine for the particular header. Secondly, factory focused on the higher rpm's area and not trying to fix the torque dip. Such a decision is certainly good for performance on the track, but not for a day to day driving where a flatter torque curve in mid rpm's would make more sense. Finally, the JDM +'17 MT cars are spec'ed at 207ps with a peaked difference of 7ps if you compare with the blue torque curve. If I make the same assumption and compare my own curve with a flat curve in the 5500-6700 rpm area, then I have the same gains and maybe a bit more because of other changes.
Overall, I would say that Subaru did a great job if we consider that older and newer headers are catted designs. Well done Subaru!
