Looking at the Elipse WingTips (Jim Smith's airplane)

Paul designed new tips for Jim Smith's RV-6 and was surprised that the performance gains were even better than expected. Good work!

But, I was still curious. So I reviewed the available information.

First, I have seen this picture of the tips and they appear to be tapered on the trailing edge as well as the leading edge:

But this is what they looked like at AirVenture2011:

I don't know for sure that there are two versions. I also don't know, if there are two, which one produced the 193.4 mph at 8000'.

Paul thinks that the CAFE RV6A's Oswald factor is .81 rather than the .85 they computed. That would be, roughly, within the 4% margin of error they cited. Taking those numbers and carefully recomputing them, I get .8493. It's in the "triangle tool" on my website: LINK if anyone wants to review it. CAFE's numbers balance to their speed and fuel flow, so I think they are probably pretty close. However, I, too, think that number seems high for a low AR, constant chord wing.

If I use +4% induced drag for the CAFE RV6A and leave parasite drag the same, I get an Oswald of .817 which is very close to Paul's result and within the stated CAFE report margin of error. I used that in the steps described below after first doing it with the original CAFE numbers.

As I described in an earlier post, I used factors from Van's specifications to change those numbers, step by step, to match Jim's airplane.

First I compensated for the lighter weight (1440 vs 1650). Then I compensated for the lack of a nose wheel. That gave me a model that can be compared to Jim's plane (lower the drags for faster TAS, match the HP). Then I built the model for Jim's plane going 193.4 mph at 8000' as cited by Paul in two earlier postings. I got an L/D of 13.5 (14.0) vs CAFE's 12.25 and a speed for L/D max of 94.4596 (92.78) instead of CAFE's 106 mph. It's a very substantial reduction in induced drag. I used Paul's estimate of slightly increased parasite drag.

From there, I computed the Oswald factor for Jim's plane to compare to the CAFE plane, both ways. The numbers for my math model of Jim's plane differ depending on which version of CAFE numbers I start with. CAFE's Oswald with the 4% correction above is .817 and Jim's is .8048. Oops! This is one of the cases for which the triangle tool was designed; it comes up with an unreasonable result and compels the re-examination of the input data. This suggests that the orginal CAFE data is better.

The original CAFE Oswald is .8496 versus Jim's .865. In other words, the increase in wingspan from 23' to 26' explains a lot of the reduction in induced drag, but there is also another 1.8% improvement in Oswald due to the shape of the tips. That's a more reasonable result since Paul's design is based on well documented research.

Paul said the Oswald is .91. Although my version is lower, it still represents a significant improvement. Paul is using more sophisticated ways of computing this. I am not knowledgeable enough to argue that one method is better, but the basic method appears to be well accepted: Oswald = Weight_squared / ( q x span_squared x Pi ) / Drag_induced.

I am using that basic formula, often cited on this forum and elsewhere. It's embedded in the triangle tool if anyone wants to review it. To put Jim's plane into the tool, change the span to 26, the two weights to 1440, the VLD to 92.78 and the vertical to horizontal ratio to 14.0.

You can also change the altitude to 8000 and the speed to 193.4. That speed is before the recent drag improvements that took Jim over 200 mph. The prop efficiency, SFC and fuel flow don't matter in this particular analysis but you can change them iteratively to see "what if". If anyone wants to review how I derived the VLD and Drag_min for Jim's plane, PM or email me and I will share the spreadsheet with explanation. It would be very beneficial if we could get those two numbers on Jim's plane in real-life, but for now, this is what we have.

Congrats to both Paul and Jim. I just play with numbers; they actually experiment. Thank you!

PS - I showed Jim's tips (the photo) to a very well known professional aerodynamicist whose recent successes would be recognized if I named them. He immediately said that he liked those tips. FWIW.

Same tips as when I flew with Jim a year plus ago...must just be weird picture angle.

BTW, it's Jim Smith, not Jim Wright.

- Another Jim

Thanks

fehdxl said:

Same tips as when I flew with Jim a year plus ago...must just be weird picture angle.

BTW, it's Jim Smith, not Jim Wright.

- Another Jim

Click to expand...

Correction made. Thanks. I must have had their pioneering spirit on my mind .

dihedral change?

I think what is causing the appearance of forward sweep on the trailing edge may be a change in the trailing edge line upward, if perhaps the new tips are lofted with a straight extension of the upper surface, and the lower surface tapers up to meet the upper at the tip? I'm just guessing, but it is a possible explanation.

A small comment on the formula that H. Evans gave for Oswald Factor is not strictly correct. The formula given is what we call "span efficiency", e and this value can be computed at any given operating condition (particular value of q and actual induced drag). If you test or analyze an airplane over a range of lift coefficient (range of speed, altitude, or weight, or some combination), you will find that the span efficiency varies over the operating range. How much it varies depends on the particular wing design. For a rectangular wing with no twist, it varies A LOT.

Oswald Efficiency Factor, E_o is something a little bit different. It is a single constant number that applies over the entire operating range. It is essentially a best curve fit of the formula CD= CD_o + CL^2/(pi x AR x E_o) to the full airplane performance polar (CD vs. CL) One way to find it is to make a plot of CD vs. CL^2 for the whole airplane polar (whole operating range). This plot will look roughly like a straight line, except right near stall. If you fit a best line fit through the data, the Y-intersept will be CD_o and the slope of the line will be 1/(pi x AR x E_o) and you can solve for E_o, the Oswald Efficiency Factor.

I can't say if this factoid has any bearing on the actual analysis that has been reported here -- its just a matter of convention and custom within the industry to distinguish very specific meanings. But it may explain why there is some variation in the reported value of E_o, if it is being computed at just one or two test points. The span efficiency, e, does vary quite a bit. The bigger problem of course is that the non-induced drag is also not constant, because of changes in flow separation points, effects of pressure gradients on skin friction drag, Reynold number effects, etc. And so it becomes rather difficult to separate true induced drag from other lift dependent drag. (part of the subject of my Thesis, by the way) The Oswald curve fit lumps it all together.

Thanks, Steve for the explanation

I did not know/understand the distinction and I thank you.

Just to clarify, though, I re-computed the speed for L/D max and took the efficiency factor only at that point. That should simplify the review of what I did. That said, each change in either parasite or induced drag causes a change in that speed. My "tool" only works at that speed where the two kinds of drag are equal. A reverse version of it finds that speed when the drag factors (FPE and induced base - as I named it) are known. It seems like the more I figure out the more questions there are.

WINGTIPS

I would challenge anyone to read the article in the Feb 1978 Sport Aviation about Steve Wittmans Tailwind. Then tell me who you think designed these wingtips. The Wittman wingtip design was a result of discussions between Wittman and August Raspet that dated back to 1954.

jrs14855 said:

I would challenge anyone to read the article in the Feb 1978 Sport Aviation about Steve Wittmans Tailwind. Then tell me who you think designed these wingtips. The Wittman wingtip design was a result of discussions between Wittman and August Raspet that dated back to 1954.

Click to expand...

We should never confuse "design" with "inventing". The Wrights could be said to have invented the airplane, but many have "designed" airplanes based upon the mounds of design data which has been forthcoming. I based the design of the wingtips on my plane on an NACA report on the best LE sweep, which was 55?. Later, in tests of Jim's plane which showed that his speed decreased remarkably at higher density altitude due to his low, low, 4.8 AR, I drew up the plans for these tips and convinced him to try his hand at making them from foam. He hasn't had one moment of regret!

I want some of these tips! Any plans to manufacture them for sale?

NeilMcLeod said:

I want some of these tips! Any plans to manufacture them for sale?

Click to expand...

Keep in mind that extending the wing span will decrease your roll rate and lower your maximum G. But if you want much better speed at altitude and with higher gross weight, for long cross-country flights, I'm sure you will be pleased with the additional speed you will get. You can make them in two different styles; like Jims with the total sweep, or as on Steve Wittman's, with the LE swept back to 25% chord, and the TE swept forward to that point. I can't say for sure that they will both be the same, but the reason they work is that the bring the wing closer to the ideal elliptical lift distribution. If you will contact Jim perhaps he can tell you how he made them.

Smith, Jim
[email protected]

Thank you. I understand the basic aerodynamics and the trade offs. Nowadays this old glider racer wants to go high, fast and efficient.