Matt Bearman Asks a Question

Matt Bearman asks a quesion.. what is going on with the area below the cockpit?

Time for another arm-waving session. There has been much debate around this area:

  • Is it flat, or just appearing so where several subtle curves meet?
  • Is it vertical, or does it break convention by being at an acute angle to the wing, as indicated by the wind tunnel model data tables?
  • Why the odd, slightly awkward slab-sided effect anyway?

I confess I was on the wrong side of this debate – I now stand corrected, and somewhat astounded by what Petter was actually doing here.



To understand his surprising innovation it is necessary for a minute to go back to the 1930’s and aerodynamic thinking at the time..  it was recognised that if one went beyond the vertical and the angle between the fuselage side and the wing became acute, one would get a serious drag penalty, called ‘interference’ drag. This was demonstrated by tests both side of the Atlantic, eg http://naca.central.cranfield.ac.uk/reports/1936/naca-report-540.pdf



These tests were on circular or ovoid section fuselages, and as one moved the wings up and down, the interference drag was recorded. Low wing arrangements created the most drag, and it was all put down to there being an acute angle formed between the local fuselage side (at the bottom of the ‘circle’) and the wing. And there stood the orthodoxy.



Designers at the time strove to create fairings between wing trailing (lifting) portion and fuselage to overcome this. All mainstream fighter designs of the period acquired a fairing, being low-wing and round-fuselage – from the big one on the Spitfire to little ones such as on the 109.



However, the Whirlwind did not. Furthermore, it had an acute angle between fuselage and wing. What on earth did Petter think he was doing?

The answer lies deep in a NACA Technical Note from 1938. TN.642 also tested what they called ‘Straight Sided’ wing intersections. In these, round fuselages had been locally altered around the wing junction to produce a straight ‘seam’ – the fuselage section was flat and vertical where it met the wing. The drag was reduced enormously compared to circular sections, and the report concluded “A modification of a round fuselage to provide straight side junctures appeared very effective aerodynamically”



Crucially, the wing/fuselage intersection was straight and parallel to long axis when viewed in plan form. This was not enlarged upon by the Tech Note, though it did describe the junctions as ‘straight’ rather than just the sides being ‘flat’. This must have been noticed by Petter!

Here’s the thing. The slab sides on the WW are not a flat surface. In fact they are ever so slightly twisted, as shown by the positioning of point ‘A’ in the diagram and table above such as to ensure a constant distance from centreline as the lifting, trailing part of the aerofoil intersects the inclined plane. The wing/ fuselage ‘seam’ as viewed from above is STRAIGHT AND PARALLEL.


What Petter had picked up on was this. It was not simply the angle of wing to fuselage that caused interference. In fact it was irrelevant, despite what aerodynamic texts at the time were saying (and what a lot of on-line resources still say now).


What is important is the straightness of the wing/fuselage intersection. When you intersect the curve of an aerofoil with the curve of a fuselage, you get an even more curved ‘join’. That was what caused the interference. This is backed up by the curious effect whereby if you increased incidence, you increased interference drag. What you are doing is decreasing the radius of this curve.

So, the answer is to correct the fuselage profile locally to make this join straight and parallel when viewed in plan, as in the Airbus design study below. Upper bad, Lower better:

This is exactly what designers of airliners, conventionally low-wing and tubular, started doing in the 1960’s thus dispensing of fairings. And it is what Petter did in 1938.

I always thought ‘Genius’ was a bit strong for Petter. Now I am re-assessing that.