I have finally resolved some minor issues I was having with the Ta152 main fuselage section, which were mainly small dimensional discrepancies that bothered me. I have now completed this aspect for the Ta152 and have temporarily diverted back to the Spitfire project.
The results of my studies are represented in a series of 2d drawings that are fully dimensioned with point details and accompanying ordinate tables.
This illustration shows progress to date with the Spitfire (in progress) at the top and the Ta152 (completed) below. The latter study is the result of 4 months intensive work, involving analysis of all documentation, cross referencing of assemblies, ordinates, parts and curvature analysis.
I was fortunate with the Ta152 partly because the original dimensions were already in millimetres so there was no deviations as a consequence of unit conversions. However there were some discrepancies attributed to mistakes on the manufacturers drawings and some misalignments, although these were marginal they still required correction…this accounted for the bulk of my work..because to change something it has to be verifiable.
In contrast, the Spitfire is dimensioned in inches on the original manufacturers drawings, albeit to 2 decimal places, this still results in a deviation when converted to mm.
In real terms this may not account for very much, but accumulatively it can add up and cause some issues with geometry alignments. The Spitfire drawing as it stands, represents over 40 man-hours on this project. The main problem I was having with the Spitfire was the canopy/windscreen – this required full development of the contour drawings initially to check ordinate accuracy – there were no less than 14 discrepancies found. Once that was developed, I needed to check this with the part drawings that make up the supporting structure in this area to verify.
The main issues tend to relate to aspects of the structure that are inherently in a straight line – the manufactures for both aircraft provide data sheets with ordinates that are supposed to define straight line segments. In consideration of the above noted conditions we invariably have some deviation in the plotted points – marginally above or below the perceived straight line. So that needs to be sorted – its not just a simple case of visually determining where this line should be, it has to be a calculated statistic of all the points to determine a “best fit” line using Linear Regression.
A calculated “best fit” line in itself is still not completely correct without first defining a datum point – in each case there is usually a point that is critical to the association with all the other components – so some study work is involved to figure this out.
The above table shows a straight line segment analysis – the first column is the ‘x’ value and the second is the ‘y’ value as per original data – the third column shows the corrected coordinate with the difference shown in the last column. Generally the difference is very small, but as you can see the last row shows comparatively a large deviation. It may be worth noting that the supporting coaming channel did not align properly with the canopy ordinates, even before this change, as there was a mistake on the original drawing, which incidentally was only noted as a consequence of doing this exercise.
This is typical of the process undertaken at each stage of development, for example the entire systemholm line was studied for the Ta152 to find the correct location of the many ordinate points that intersect along this line.