I did more testing. I have good news and bad news.
Good news: A 4ft-kammback extension improves my FE.
Bad news: It only improves by 0.2mpg (1%)
Some pictures: First, the nose aeromods I will reference, filling some gaps and a half upper grill block.
Before:
After, with prototype nose aeromods:
Next, the new version of the prototype kammback. This version sits flush with the sides and the roof of the car and would (I believe) provide for a pretty smooth transition from the car to the kamm.
Result averages:
Stock car: 21.1mpg
Added nose aero: 21.7mpg (3% increase over stock)
Nose aero and Kammback: 21.9mpg (1% increase, 4% increase over stock)
I saw similar on a previous day (without making panels exactly flush with car sides; there was a step, to sit just inside the tail lights):
Stock car: 21.1mpg
Added kammback: 21.3mpg (1% increase over stock)
This kammback has a 9-degree top panel slope and 7-degree side panel angles inward to approximate the template (which is relatively flat for this low of angles, since I can only get to 30% of the template in 4ft).
Tuft Testing Side panels:
7 degrees: attached flow, somewhat turbulent
9 degrees: attached, more turbulent
11 degrees: usually attached, very turbulent, sometimes flow briefly detached in places
Tuft Testing Top panel:
9 degrees: strings barely ever move at 65mph (either super laminar and attached, or fully detached)
11 degrees: strings barely move at 65mph (either super laminar and attached, or fully detached)
15 degrees: strings move some
18 degrees: strings move consistently a normal amount for attached, slightly turbulent flow
Because of how I designed the egg-crating on the back of my panels, I didn't allow for going any steeper than this; I didn't expect to have attached flow at this steep of an angle.
Not gonna lie - I'm a little confused by the tuft tests, since I saw expected results for pretty good flow at an 18-degree downward slope, but not at 9 degrees. At 9 degrees, there are several-second intervals where it looks like the strings don't move at all in the center of the panel. Does this mean flow is fully detached?
On all top panel tests, the strings near the side edges move more than in the center. Maybe I have a really strong downwash in the center of the panel and turbulence near the edges. Currently, there's a sharp corner between my top and side panels, which would get a large radius if I make a permanent version. I saw a more normal attached, slightly turbulent flow on the 9 degree panel when I had a step down from the top of the car (to not cover the top brake light) than when the panel started flush with the top of the car.
More mpg results, if helpful to interpret the puzzle:
9 deg. top panel: 22.0mpg
11 deg top panel: 21.5mpg
15 deg top panel: 22.3mpg
18 deg top panel: 21.8mpg
Any ideas as to what is happening and how this flow could be improved?
Test an angle greater than 18 degrees? (This seems crazy, since it's close to the "terrible" 22-degree angle)
Maybe I can't expect good MPG gains, since I can only get to 30% template in this length...