Hi Blogs,
>When I saw the origonal plots I wondered about the standard errors of these estimates.
The problem is that the Finnish test only measured speed over altitude, not power over altitude, and the engine was not performing in accordance with the DB605A power graph.
However, from observation of the full throttle heights and the drop of the power it's possible to arrive at an approximated power curve (using a 3rd order polynom) that can be used for further calculations in order to estimate the standard deviation.
Here is the calculated drag (relative to the average) for the full set of data from the Finnish table:
140 m: 95,3%
1130 m: 96,8%
2130 m: 96,5%
3120 m: 95,5%
4120 m: 96,6%
5120 m: 98,7%
6120 m: 97,2%
6420 m: 96,6%
7110 m: 109,7%
8110 m: 103%
9110 m: 95,4%
10100 m: 118,8%
Standard deviation: 7,2%
This gives the following data:
631 km/h @ 6420 m +/-11 km/h
581 km/h @ 10000 m +/-11 km/h
However, one look at the drag values should be enough to recognize that the data gathered above full throttle height is considerably less reliable than that gathered below full throttle height.
Standard deviation drops from 7.2% to 1.1% by simply disregarding all (four) values above full throttle height.
Here's the data resulting from that decision:
0140 m: 98,6%
1130 m: 100,1%
2130 m: 99,9%
3120 m: 98,8%
4120 m: 99,9%
5120 m: 102,2%
6120 m: 100,6%
6420 m: 99,9%
Not considered:
----
7110 m: 113,5%
8110 m: 106,5%
9110 m: 98,7%
10100 m: 122,9%
----
Standard deviation: 1,1%
635,6 km/h @ 6420 m +1,9/-1,3 km/h
586,0 km/h @ 10000 m +1,6/-1,6 km/h
Due to the incomplete nature of the Finnish test, this should be only taken as an illustration of the inherent inaccuracy of performance measurements.
>I think its under appreciated how noisy these numbers can be when only 1 or 2 tests are run under any set of parameters.
Well, the values below full throttle height are quite consistent in my opinion.
Above full throttle height, they're extremely poor, but considering that up there, the limited excess power up combined with the increasing difficulties of getting accurate measurements at high altitude, that shouldn't come as a surprise :-)
Regards,
Henning (HoHun)