Hiko said...
We decided at the time that the slot where the wheel went into the fairing was the problem so I think that particular attention should be paid to that area
Cheers Hiko
I remember a model airplane nut who did similar testing, and he came to a similar conclusion... he found that the critical part was the distance between the fairing and the 'axle'. The axles in this case are plate spring steel, which run vertically parallel to the wheel before bending inwards and attaching to the fuselage.
By adding and removing washers to change the gap he found a massive difference in drag. He eventually found the best solution was no fairings, with a certain gap between the 'axle' and the naked wheel.
Not sure if it applies here though... you may find that adding golf ball dimples to the fairing might help create a better boundary layer, similar to how Mythbusters tested dimples on a car.
(fast forward to 2:00 or so)