I was looking at the data Sandor provided again because I was curious about the patterns in the data I mentioned earlier in the thread. I decided to plot the change in measured beam height from point to point for both the results he obtained and the flat earth nominal change. I found the results to be interesting to say the least.
- dRes is the change in Sandor's measured values from point to point.
- dNom is the change in "measured" values using his nominal flat earth calculation.
- dNom - dRes is the difference between the two.
- Coalesced - to be explained below.
If you take a look (at
dNom-dRes), it appears that much of the time Sandor's results (for the change in height between points) deviate by 0.00 or 0.01 from nominal. This means they slowly diverge by 0.01 per point in general. At other points, they appear to deviate
by more (marked in blue) but then subsequent data points
appear to compensate for the discrepancy from the nominal values. Or to explain that differently, the
red points in dRes add up to the same value as the
green points in dNom. So, if you then treat these discrepancies as a single point (by adding them together) as shown in the Coalesced column (in
yellow), what you get is is a deviation of only 0.01 maximum at any point. This would explain the patterns in the data (in the chart I showed a number of posts ago), as if someone were trying to consistently under compensate by 0.00 or 0.01 but wanted to spread it over multiple points to make it less obvious. That sort of thing would necessarily result in patterns in the data.
Let me know what you guys think (or if I didn't explain this clearly).
View attachment 21359