Non-attached flow was seen when the front end was too small, and then it picked up further back. It was just too steep of an angle (we knew it would be at the time, but we liked testing). We didn't have the front end of the trailer tufted at that time.
We also saw poor attachment along the top when the front end was too high, presumably shoving the air out of the way, and having a hard time re-attaching. This is the case where we had monstrous down-draft. My buddy joked that we could have run a turbine in there. We implemented a stromform (ramped) front end that helped a lot, but did not correct the problem.
All cases had a large (9") diameter leading into the trailer body, so it wasn't some squared-off front end problem.
The most optimal case was a front end that was very close (about 2') to the lead vehicle (a ford 12 passenger van), and well below the roofline. The attachment was so strong that I thought we had no attachment (not a ripple the entire length). It wasn't until we slowed for a turn that the tufts all moved at the same time, and straightened out again when we picked up speed. In this case, we still had a down-draft, but not near what we had with the "high top" we had tried previously, and even a lot less than it had with the stromform front end.
__________________
|