I apologize for starting yet another thread on this subject, but thought that the topic was of sufficient interest to deserve it…
Anyway, here’s the deal-io: back in 2007, my wife went to the Texas A&M wind tunnel to refine her position aboard her P3C. Among the many things that we compared was her CdA when using a pair of Zipp clinchers (808 front and ‘bulged’ disk rear) shod with Veloflex Record tires versus when using a pair of Blackwell tubulars (100 front and flat disk rear, the latter being a rebadged, non-dimpled Zipp disk) fitted with Continential Podium tires. Despite the numerous differences between conditions, I thought it might be interesting to revisit these data, to see if it might be possible to ascertain any difference due to the shape of the rear disk alone. For starters, though, here are the raw data (mean +/- SD for 2-3 separate runs):
As I mentioned elsewhere, the protocol entailed a yaw sweep that went from 0 ->2.5 → 5 ->10 ->5 ->2.5 → 0 deg, with the left side of the bike facing the wind. As can be seen in the plot, when using the Zipp wheels there was little hysteresis in the data, with drag being quite similar at 2.5 and 5 deg regardless of the direction from which that yaw angle was approached. This indicates that minimal flow separation occurred; i.e., that the equipment used did a good job of ‘managing’ the air flow to prevent excessive turbulence.
In contrast, when using the Blackwell wheels, there was clearly some hysteresis. In particular, note how although CdA was higher even at 0 deg of yaw, the decrease that occurred at 2.5 and 5 deg of yaw on the way ‘out’ paralleled that found when using the Zipp wheels. CdA then leveled off, however, and was then clearly higher at 5 and 2.5 deg of yaw on the way ‘back’ before then dropping back down to equal that originally found at 0 deg. What this indicates is that the pedaling rider+bike ‘stalled’ somewhere between 5 and 10 deg on the way ‘out’, leading to flow separation. Flow then remained detached, such that CdA was higher, until everything was pointing straight into the wind again.
The above in itself would be good reason to choose the Zipps over the Blackwells (and in fact we already had). In the present context, however, I was curious about the extent to which this behavior could be explained by the differences in the front wheels alone. The logic here was that if you could adequately account for that difference, then any remaining difference would have to be due to the rear wheel, and thus might provide some idea as to whether a ‘bulged’ disk really helps on a P3C. Roady was kind enough to dredge up the necessary drag-vs.-yaw-angle data for the Zipp 808 and Blackwell 100 to permit such a correction, which I then used to produce the plot shown below (note that since the Blackwell disk is a rebadged Zipp, I’ve taken the liberty of labeling it as such):
To come up with this graph of hypothetical data, I converted the drag data from grams to CdA (assuming a tunnel speed of 30 mph and an English unit standard air density of 0.002378 slugs/ft^3) and substracted the difference between the 100 and 808 from the data obtained when using the pair of Blackwell wheels. I then further offset the data by the difference in CdA at 0 deg of yaw, under the assumption that any difference there is likely to be due to other factors (e.g., differences in tires, presence/absence of dimples) and not due to the bulge per se. (Note that the difference in tires is such that, if anything, I have probably undercorrected the data.)
In any case, as shown above the results now parallel each other much more closely, in particular in the way that CdA continues to decrease all the way out to 10 deg (note that since the wheel data provided by roady were obtained during ‘reverse’ sweeps, i.e., with decreasing yaw angle, I have only applied the correction to the comparable rider+bike data). The conclusion here, then, is as Tom A. just pointed out: most, albeit certainly not all, of the difference in what we found when comparing pairs of wheels seems to have been due to the difference in front wheels. In turn, this suggests that there is, in fact, limited (albeit not insignificant) benefit to using a ‘bulged’ disk on a P3, just as Gerard has stated. Due to the fact that this was not a planned comparison, however, one should not place too much faith in this conclusion.
FWIW…
