Presented by Nine Lives Racing
1 2 3 ... 9
Robbie
Robbie UltimaDork
3/26/19 2:12 p.m.

Always on the prowl for good ideas, I spent some time recently thinking about aero. That brought be to airfoiltools.com where you can see the coefficient of lift and drag and various other properties for many different airfoils profiles. Some googling brought me back to old GRM posts about how profile 'e423' could be a good one for lower speed motorsports since it produces relatively high coefficient of lift (or downforce) at low reynolds numbers (low speed and small wing area in air). Here is what the profile looks like:

You would make it upside down for a downforce application. 

Unfortunately, that profile is pretty complicated and would be really time consuming to make out of common or cheap materials. I thought about making a template out of metal and then using a router to cutout about a hundred pieces and then gluing them all together to make a giant plywood wing, but holy moly not sure if I have that patience. I also worried that a bit of tear out at the pointy end could really affect the performance of the wing, and would be hard to avoid. 

Then I learned that some folks are making airfoils for home-built wind turbines and such from sections of PVC pipe. Basically you cut the tube length-wise so you end up with a curved section like a slat from some window blinds, and then round the front edge and sharpen the back edge. airfoiltools also has some of those profiles:

 

and using the comparison feature, the section of pipe wing actually outperforms the e423 wing at low reynolds numbers (the pipe wing is blue). More lift at angle of attack 10-17 degrees:

AND less drag too! 

So, a section of pipe wing actually seems much more feasible to create than hundreds of carefully profiled pieces, mostly because the thickness of the wing doesn't vary much. A piece of flat stock (i.e. plywood) could be used. 

Here's where things get interesting:

The pipe-section profile I showed above specifies a thickness of 5% of the radius of the curve, and a chord length of radius/0.963. Let's say that I picked up about 16 pieces of 3/4 inch ply that are 16 inches wide and 8 ft long out of the trash, because I did (I think they were shelves in someone's basement and they got wet, so out they came and into the trash). If thickness is 3/4inch, radius is 20x that, or 15 inches. and then the chord length is 15.6 inches. I since the chord length doesn't take the curve into account, I'll need slightly more than that to get there. Someone else I'm sure can do the math on this, but 16 inches will be quite close.

If I use a router to round-over the front edge of the plywood, and then use a table saw to cut the back edge sharp (like maybe a steep angle like 60 degrees or something), then I just need to bend the plywood into a 15 inch radius. Enter an idea from another recent thread on here, and cut slots with the table saw:

Say the outside edge is where you want your 15 inch radius measured from (the airfoiltools profile say the middle, but this is for quick and dirty). So a full circle would have a circumference of 94.2 inches. A full circle with a 14.25 inch radius would have an 89.5 inch circumference. about 5% less. so the inside needs to have 5% less width than the outside. 16 inches minus 5% is 15.2 inches, or about 15 and 3/16 (or 1/4). So if the table saw blade is 1/8th inch thick, there would need to be 6 slots cut.

Now finally, we need something to bend this around. Tires from suv/trucks etc generally have almost 30 inch diameter.  So, cut the slots, fill them with wood glue, ratchet strap the plywood to the outside of a few tires, wait for the whole thing to dry, and BOOM! you have a nearly-free wing that should actually perform pretty well. 

What am I missing here?

Also, don't worry. I very much plan to do this and I intend to test the results as well and keep this thread updated. 

GameboyRMH
GameboyRMH MegaDork
3/26/19 2:23 p.m.

I'm wondering why you don't use an actual section of pipe for this to save a ton of fabrication work - PVC would probably snap, but what about ABS or aluminum? Or something light and weak as a mold or core for wet CF construction?

Robbie
Robbie UltimaDork
3/26/19 2:26 p.m.
GameboyRMH said:

I'm wondering why you don't use an actual section of pipe for this to save a ton of fabrication work - PVC would probably snap, but what about ABS or aluminum? Or something light and weak as a mold or core for wet CF construction?

I don't know where to get a 7-8ft long tube with a 30 inch diameter (other than maybe cardboard concrete forms). Do you have some ideas?

Keith Tanner
Keith Tanner MegaDork
3/26/19 2:36 p.m.

Why not go old school with spars, ribs and a doped canvas or fiberglass cover? You can make the ribs from your free plywood.

KyAllroad (Jeremy)
KyAllroad (Jeremy) UltimaDork
3/26/19 2:37 p.m.
GameboyRMH
GameboyRMH MegaDork
3/26/19 2:40 p.m.
Keith Tanner said:

Why not go old school with spars, ribs and a doped canvas or fiberglass cover? You can make the ribs from your free plywood.

Haha my first thought was actually a WW1-era airplane wing.

Been looking for giant ABS pipe...easy to find in corrugated form, apparently doesn't exist with regular solid walls...

chaparral
chaparral Dork
3/26/19 2:44 p.m.

You can build really nice wings out of plywood. Bingelis has the entire technique written out in "Sportplane Construction Techniques"

Build a pair of spars, one nearly vertical at the front to handle most of the force, one with a beveled edge forming the trailing edge. 

Build a rib to the profile you want (convex is best), and make enough of them to have two per foot of span. Cut out the shapes of the spars and the TE.

Glue the spars to the ribs. 

Cover with cloth (thin polyester comes closest to real Dacron), glue and sew so that it isn't able to flap in the breeze, shrink with an iron.

Attach to car using endplates - bonus points for adjustable AoA. 

stafford1500
stafford1500 HalfDork
3/26/19 3:19 p.m.

The issue you may run into is strength of the glued concave surface. Simple thin curved shapes were the fore-runners of modern airfoils and do work at low Reynolds numbers. For a ~15" chord and typical autocross speeds you need to be designing to a pretty distinct Reynolds number. Also, the angle of attack that sort of wing works in is very narrow, due to the thin leading edge. Find the right AOA and you are good. Outside of a small AOA window and the downforce and drag will be BAD, stalled on either the top or bottom surface.

Any downforce you add is better than what you have now. Drag will not be a significant issue at autocross speeds.

chaparral
chaparral Dork
3/26/19 3:39 p.m.

stafford1500,

On a car wing the concave surface would end up on the top - might be less critical than the bottom surface which needs to be complex its whole length to avoid having to sew it together there...

sleepyhead
sleepyhead Mod Squad
3/26/19 4:55 p.m.

two things to keep in mind...

using the plywood like a plate like that means you've got a very short "beam"... and that beam is much less effective after having hollowed it out and re-glued it back together.

second, that section (cambered plate) has a very small leading edge radius... which as stafford has alluded to, will significantly reduce the window it can operate in.

you probably want to use the plywood as spars running the length of the span, at 25% and 60-75% of the "chord" (front to back length).

Don't worry about the "back edge"... most of your performance will be generated at the front 50% of the airfoil

I have other thoughts... but it's pretty late here

Robbie
Robbie UltimaDork
3/26/19 5:02 p.m.

Posting this table for my reference later. Assuming hypotenuse of 16 inches, rear of wing is raised y inches for x angle of attack.

  • 8 degrees is 2.23 inches
  • 10 degrees is 2.78 (+.55)
  • 12 degrees is 3.33 (+.55)
  • 14 degrees is 3.87 (+.54)
  • 16 degrees is 4.41 (+.54)
  • 18 degrees is 4.94 (+.53)
  • 20 degrees is 5.47 (+.53)

So basically every 1/4 inch I raise the rear of the wing gives another degree of AoA. 

bustedplug
bustedplug New Reader
3/26/19 5:46 p.m.

If I understand correctly it seems as if you want a 1/4 of a 30 inch pipe and you have a table saw and free plywood. I think something like a 6-7.5 degree bevel on about a 2 .25 inch rip and glue is the easy answer. drum or barrel making calculators would have better maths.

Appleseed
Appleseed MegaDork
3/26/19 7:36 p.m.
GameboyRMH said:
Keith Tanner said:

Why not go old school with spars, ribs and a doped canvas or fiberglass cover? You can make the ribs from your free plywood.

Haha my first thought was actually a WW1-era airplane wing.

Been looking for giant ABS pipe...easy to find in corrugated form, apparently doesn't exist with regular solid walls...

Unsercambered airfoils are extremely draggy. Extremely. 

I would use an inverted Clark Y. How thick is the plywood you are planning on using? If it's thick enough you can create the top and bottom surfaces elongated wedge of the trailing edge. Stack the ply to rough out the leading edge and route/ beltsand to shape. 

buzzboy
buzzboy HalfDork
3/26/19 8:42 p.m.

I've got a bunch of 55gallon drums laying around. Time to build a wing

Appleseed
Appleseed MegaDork
3/26/19 8:53 p.m.

Blue foam insulation with a fiberglass skin builds wonders.

Robbie
Robbie UltimaDork
3/26/19 11:05 p.m.
bustedplug said:

If I understand correctly it seems as if you want a 1/4 of a 30 inch pipe and you have a table saw and free plywood. I think something like a 6-7.5 degree bevel on about a 2 .25 inch rip and glue is the easy answer. drum or barrel making calculators would have better maths.

I never thought about a barrel, but yes this would approximate what I plan to do. I'll just be cutting slits instead of cutting all the way through the plywood however.

Robbie
Robbie UltimaDork
3/26/19 11:28 p.m.
buzzboy said:

I've got a bunch of 55gallon drums laying around. Time to build a wing

A 55 gallon drum wing would look sick on a "rat" style race car.

Take two 12-16 inch slices of the drum (maybe 4 if you want to double the width) weld their leading edges to opposite sides of a piece of metal roll cage tube, and weld the trailing edges directly together. From the side it would look like a large comma. 

Cut the bottom in half for semicircle end plates.

Boom. drumwing.

Robbie
Robbie UltimaDork
3/26/19 11:33 p.m.

I think the traditional wing construction (spars, foam, fiberglass, etc) can of course make a really nice wing to almost any profile you want. 

But they honestly look like they would take a long time to build. I think I could have this plywood wing glued up and drying in a couple of hours.

SVreX
SVreX MegaDork
3/27/19 1:00 a.m.

Mr Joshua built a traditional wing for the Mumpkin with plywood ribs and skinned it with formica.  For lateral strength, he drilled holes in the ribs and ran a large dowel through the ribs side to side.  It had scrap aluminum for end plates.  It was incredibly strong and lightweight.

It could be built in a day.

The formica gave a smooth surface which was VERY strong. 

Fold the formica in half, glue to the ribs, and glue the 2 edges of the formica to themselves at the knife edge.  Make sure that your overall perimeter of the profile is less than the available width of formica.

stafford1500
stafford1500 HalfDork
3/27/19 4:17 a.m.

Robbie,

I think you are headed down the right path for the first iteration of added downforce. Aero can be a simple thing, or it can be as complicated as you want to make it. Your simple/easy wing section will get you most of what you want. More refinement gets you more, but there is definitely a diminishing return to effort involved. You will get some experience with aero the old fashioned way: build it, see what it does, refine it.

One thing most people do not picture when building aero bits (specifically wings) is that the flow around a car is very much a 3D issue. The data you found is for 2D sections and assumes there is no finite span or end plate effect, all with no lateral flow components. The 2D results always overestimate the 3D results. The reason I mentioned the thin leading edge is that the 3D flow may drive the ends or the middle of your cambered plate into stall while the rest is still trying to work at your design AOA. To correct this issue the wing section would need one of two things: a varying section across the span or a larger radius leading edge to allow it to operate across a wider AOA. The larger radius generally leads to more traditional airfoil shapes, while the varying section is a pain to design/build.

The follow on to where you are at now is twisted/curved sections, and potentially multiple sections, leading to cascades like you see on Amod cars and F1 front wings.

I hope to see your efforts turned into something for the Challenge or other projects here on the board. If you have any questions post them up or PM me directly.

sleepyhead
sleepyhead Mod Squad
3/27/19 5:07 a.m.

one refinement to add... is it looks like the comparison you pulled between the e423 and the 'cambered plate' was at Re = 50,000

based on the assumption you're building a 16" chord wing, and using it for challenge, with a "min speed" of 20mph and a "max speed" of 60mph... your Re range is ~250,000 to ~750,000

Re = (.002377*V.fps*chord)/(3.737e-7)

where V is in ft/sec, and chord is in feet.  .002377 is "standard day air density at sea level in slugs/ft^3" and 3.737e-7 is "standard day air viscosity at sea level in slugs/(ft*sec)"

SVreX
SVreX MegaDork
3/27/19 5:33 a.m.

In reply to sleepyhead :

Can you translate that please?

I think you are saying for autocross he will need a much bigger wing to have any significant impact. Am I close?

Treat me like a 5 year old (because I’m not an engineer)

sleepyhead
sleepyhead Mod Squad
3/27/19 5:49 a.m.

In reply to SVreX :

Ok... so... the relationship of how big an object is, how fast it is going, and the density and viscosity of the fluid/gas it is moving through "at that moment" will determine how that fluid/gas will... interact with the object.  "Reynolds Number" commonly shortened to "Re" is the "unitless representation of where you are in that relationship"

I'd have to pull out my Boundary Layer or Fluid Flows of Nature book... but, "back of the napkin, to non-engineers" is that around Re = 1,000,000 == turbulent (which doesn't mean what you think it means), and is considered "normal" or "large".  Re = 100,000 to 1,000,000 is considered (iirc) Laminar... and "Low".  Below 100,000 is super-low-ish... and isn't really applicable here.

the closer you get to 1million, the more the wings act like "real airplane wings"... outside of like "Airliners"

The end result is that... in the "low Re" regime... the flow can more easily "detach" from the surface... or as stafford would tend to say... it can be harder to convince the air to follow the shape and "do what you want it to".  Over 1million, it's easier.  Also, Drag tends to be lower at lower Re's.

edit:

I think you are saying for autocross he will need a much bigger wing to have any significant impact. Am I close?

Uh, not really.  I'm kind of saying the data he was using in the first post isn't valid to make comparisons with, because it's effectively "slower" than would be seen for autocross.

SVreX
SVreX MegaDork
3/27/19 6:02 a.m.

In reply to sleepyhead :

Gotcha.  Thanks.

SVreX
SVreX MegaDork
3/27/19 6:05 a.m.

In reply to sleepyhead :

...although it did make me wonder what 5 year old is familiar with Reynolds Number and The Boundary Layer or Fluid Flows of Nature!!  cheekycheeky

1 2 3 ... 9
Our Preferred Partners
OEIiqbJ44S8cfPGr9fdweXJJuW8NjsDg51I2Vzy9e73zocgaQHHWufGri0aViSmv