Between 265 & 485 SQ/FT The 12-Gauge Garage

[Jack Olsen]

Well, I took it out and... it had a whole different feel. At low speeds, it worked great. But anywhere above about 50 mph, the downforce the wing was generating overwhelmed my cable system. I could try fastening all the cables down more securely, to reduce flex. But my fear is that if I'm having trouble with the downforce at 50 mph, then the forces I'll be dealing with at 100 mph (since downforce is increases exponentially, not linearly) are probably going to be too much for the kind of control cables I've set up.

Here's some video or the thing working fine at 20 mph, but struggling at higher speeds. Once it's up, it's easy to hold in position with the pedal. But that transition is going to take a lot of force in order to work.



Back to the drawing board.

 

[Aerocraftsman]

Is it possible to alter your pivot point. I think if it were closer to the mid point of the wing chord it would reduce the force required to move it under load.
Just a thought.
Mark

 

[jmauld]

Looks like a cool project for this: http://www.musclewires.com

 

[Jack Olsen]

Is it possible to alter your pivot point. I think if it were closer to the mid point of the wing chord it would reduce the force required to move it under load.
Just a thought.
Mark

That's a great point. I'm going to get started on version 1.1.

 

[Crown]

Am I seeing this correctly and are you pushing the wing up at the front? wouldn't it be easier to pull the wing down at the back at higher speeds? If you're able to get the pivot point more towards the front the wing would flatten out under the air pressure alone creating your low-drag mode at higher speeds.
With the current setup your wing is at "rest" at low speeds (you don't have to push the pedal to keep it down) and creating downforce in this mode. You have to use force (pushing your pedal) at higher speeds to flatten it out. In other words: when there is a lot of force on the wing at higher speeds because of the drag you have to use even more force now to flatten it out.
I think that's the problem right there: If the wing would be at "rest" at higher speeds and you would have to use some force to get it in downforce-mode at lower speeds it would be easier to do so. It's the same concept but the other way round.
Now we have to design a system that works the other way round but doesn't require holding your pedal when you have to change gears al the time...

Could you make a sketch of your current system?

 

[rmalkow2]

If I'm seeing the cables correctly in the video it appears you have them crossing in an X pattern within the air space under the wing. In addition to what has already been said two issues come to mind. First this would appear to make the load directions have to change maybe more than desired to perform the actuation on the wing. Side (X) load is converted to vertical. Second, the cables are exposed to a pretty turbulent air stream area over the back of the car body. I don't know if you can ever eliminate the cable dance by keeping them exposed this way. Can you move the actuating link portion of the system that runs from the wing down to the car body to make it vertical and close to (or hidden inside) each wing upright? Can you make this vertical link out of light bar stock or tubing with good pivot joints and then actuate it via a small cam an the cables? This could keep the cables mostly under the skin of the car and provide a solid vertical link to the wing.
Not sure all this makes sense but I think you'll picture it.

 

[Jack Olsen]

Am I seeing this correctly and are you pushing the wing up at the front? wouldn't it be easier to pull the wing down at the back at higher speeds?

It would be easier, maybe, but not safer. One thing that's different here than it would be for an airplane is that there is not a direct relationship between speed and the downforce I need. On my home track, I go from just under 100 mph to about 132 mph on the back straight. But the back straight isn't actually straight, and I would benefit from no drag (and less downforce) for the first part of that track section, but then need full downforce in the last part when I'm reaching top speed and the track is curving through a very fast sweeper.

Many systems for active aero do opt to control the trailing side of the wing, like you describe. I don't like that for a couple of reasons, at least for my application. The most important problem is that it eliminates a safe 'dead man' position, where if there's an operator or mechanical failure, that kind of wing is going to be pushed by the airflow into its low-downforce position, which is no good when you're entering a high speed corner. With the leading edge articulated up for low downforce (and low drag), the force of the oncoming air should push the wing back to its safer (high downforce) position if something goes wrong.

It also will take a lot of energy to push the trailing edge of the wing up and into its high-dowforce position, which you need to do at the car's top speed before you brake and go into a corner where you need maximum downforce. And that movement requires a pushing force, rather than pulling, which has always seemed more difficult to my way of thinking. (Although, I am definitely NOT an engineer, so my thinking on this sort of thing is limited.)

If I'm seeing the cables correctly in the video it appears you have them crossing in an X pattern within the air space under the wing. In addition to what has already been said two issues come to mind. First this would appear to make the load directions have to change maybe more than desired to perform the actuation on the wing. Side (X) load is converted to vertical. Second, the cables are exposed to a pretty turbulent air stream area over the back of the car body. I don't know if you can ever eliminate the cable dance by keeping them exposed this way. Can you move the actuating link portion of the system that runs from the wing down to the car body to make it vertical and close to (or hidden inside) each wing upright? Can you make this vertical link out of light bar stock or tubing with good pivot joints and then actuate it via a small cam an the cables? This could keep the cables mostly under the skin of the car and provide a solid vertical link to the wing.
Not sure all this makes sense but I think you'll picture it.

No, the cables you see crossing are not a part of this system. They are there because I don't have a better way to resist lateral forces for the two uprights, which have to bolt to the underside of the rear decklid on the car and need reinforcement to not collapse. I'm aware of the drag penalty from those cables, but it's the best I've been able to do so far.

There ARE cables going to the controls for the wing mechanism, but they come straight up the leading edge of the uprights.

Some of the DIY approaches to this problem use a motor to control the movement of the wing, and the motors made for automatic snap-up headlights are a popular choice. I was leery about this for a couple of reasons: cost, weight, and the risk of the motor failing when the wing is in its low-downforce position. But it might be that I have to rethink this.

Conversely, it might make sense to replace as much of my current cables with simpler connecting rods -- although finding a straight path through a car's chassis is tricky.

I could also find a way to increase the mechanical advantage at the control end -- trading travel of the pedal for more force on the 'wing' end of the chain.

 

[Jack Olsen]

Here are some illustrations of what I have.

This is the mechanism in the raised (pedal depressed, low-downforce) position, and then in the default, lowered (pedal released, high-downforce) position.

And this shows the cable that pulls so that the leading edge of the wing will be lifted. Looking at this, I immediately see that -- in addition to the benefit of moving the rear anchor closer to the fore-aft center of the wing -- I would also benefit from switching the cable side and the lifting side of that little see-saw part of the mechanism, so that the lifting point is closer to the leading edge of the wing, and not so close to the center.

You can also see how the (unpainted, silver) piece that attaches to the wing has two sides, so that the (painted black) upright is trapped between two pieces, with a stop on the lower and upper sides to limit the absolute range of movement.

And again, here's a video clip that shows the movement involved:



Click on the image for the video.

 

[lisiecki1]

You could probably gain quite a bit of force just changing the length of your levers.

Also, should you go electric, maybe you could put the motors on a simple momentary push button on your gear shift and wire through a warning light on the dash. At least that way you would have advance warning going hot into a corner with the light still on....

 

[Aerocraftsman]

Maybe a trim/booster tab would help. When you're trying to move the surface in a direction that causes a load, the tab can be linked in such a way to help to aerodynamically drive the surface in the desired direction and reduce the load. The nice thing about a system like that is that the faster you're going, the more effective it becomes since like you pointed out, it's not a linear increase in force.
Mark

 

[Crown]

It would be easier, maybe, but not safer. One thing that's different here than it would be for an airplane is that there is not a direct relationship between speed and the downforce I need. On my home track, I go from just under 100 mph to about 132 mph on the back straight. But the back straight isn't actually straight, and I would benefit from no drag (and less downforce) for the first part of that track section, but then need full downforce in the last part when I'm reaching top speed and the track is curving through a very fast sweeper.

Many systems for active aero do opt to control the trailing side of the wing, like you describe. I don't like that for a couple of reasons, at least for my application. The most important problem is that it eliminates a safe 'dead man' position, where if there's an operator or mechanical failure, that kind of wing is going to be pushed by the airflow into its low-downforce position, which is no good when you're entering a high speed corner. With the leading edge articulated up for low downforce (and low drag), the force of the oncoming air should push the wing back to its safer (high downforce) position if something goes wrong.

It also will take a lot of energy to push the trailing edge of the wing up and into its high-dowforce position, which you need to do at the car's top speed before you brake and go into a corner where you need maximum downforce. And that movement requires a pushing force, rather than pulling, which has always seemed more difficult to my way of thinking. (Although, I am definitely NOT an engineer, so my thinking on this sort of thing is limited.)

Good point, I haven't thought of that early this morning. It sure is nice to have an air brake as an emergency backup (and added traction).
Thanks for the added info and vid of the wing. It makes your concept easier to understand and I must say I'm impressed by your ingenuity.

I think it would be easier to operate if you'd pull the levers at the back instead of the front (reversing the levers actually) and connect the front of the lever to the front of the wing, further apart from the hinge at the back.

You could probably gain quite a bit of force just changing the length of your levers.

True, but the downside of that is probably a longer travel of the pedal or a shorter travel of the wing.

 

[EdT]

Interesting problem. I not an aeronautical engineer, so take these comments for what you think they're worth. As mentioned above, moving the pivot point closer to the middle of the wing is probably a good idea. There is a center of pressure for your wing in both the "flat position and the high drag position. I have no idea where it is, but it probably moves around a bit during the deflection of the wing due to the extreme reflex at the trailing edge. In order to maintain your desired fail safe system, I think the pivot point for the wing should always be behind the CP for both conditions AND the transitional positions between the extremes. In this way, no matter what fails the wing should revert to the high down force condition.
I am pretty sure your linkage could be far simpler and more direct. Moving the pivot point toward the CP will make the effort required to move the wing far less since the aerodynamic loads will tend to be self canceling. That said, with the pivot point anywhere behind the CP, the wing will not want to stay with the front facing forward unless it is restrained so it would be preferred if there was always some downforce coming from the wing. If it gets too flat, it will want to flip over the top and, of course, when it's in the max downforce position, it's trying to flip "under" the pivot points. I imagine the loads on the stops may be considerable. How much down force does the wing produce?
Just a few random thoughts. Perhaps some one who is an AE will chime in.

 

[VMX42]

Hey Jack,
It seems you are morphing from Garage Guru to a Mad Scientist...

I would be looking at a different method of actuation for your movable rear wing. The mechanical system simply won't have enough mechanical advantage to achieve what is required.

I suggest you look into some of the modern 12v Linear Actuators that are now commonly available.

http://www.servocity.com/assets/ima...hannel_and_tubing_with_lightweight_mounts.jpg

With a bit of research I am sure you will find one that will suit your requirements.

Good luck, really enjoying your projects,
Cheers,
Jeff

 

[kbuhagiar]

Hello Jack,

I find this latest engineering challenge fascinating.

Knowing how important it is to you, at what point does the additional vehicle weight (from installation of actuating hardware) supersede any aerodynamic advantage that may be attained? Or is that not a consideration at this point?

Just curious...thanks!

 

[Alexbn921]

I have to agree with most of the other posters there is no way you can overcome the downforce of the wing unless you move the pivot point to a more neutral position or use some sort of linear actuator. With that being said, my rc airplanes use a leading winglet to balance the force on the tail and decrease the load at the control stick. Maybe you could use the wind to your advantage by actuating a small airfoil to do most of the work.

 

[TwoBytes]

Hey Jack,

Have you thought about hydraulic activation?

I think you could still do it cheap and light, and you could still design it so it "fails safe".

On how to do it cheap... I reckon you could get a few lightweight motorbike rear brake master cylinders from a wrecker, and modify two of them to serve as "slaves"...
https://www.google.com.au/search?q=motorcycle+rear+brake+master+cylinder&tbm=isch

You could even tune the mechanical advantage by selecting master cylinders with different bore and stroke measurements.

I'm not sure what would be involved in using a master cylinder as a slave, I haven't thought it through completely, but it seems like it should be doable??? The other option might be a slave cylinder from a hydraulic motorbike clutch??

 

[TwoBytes]

Random question...

What is that square block thing called?

I'm not sure what I would use it for exactly, but I think I need one

I can find "vee blocks", and "webbed angle plates", but I'm not sure what to search EBay for to get one of those blocks.

I'm also wondering what that thing behind the square block is.

 

[Squankum]

Hey Jack,

Have you thought about hydraulic activation?

Yeah! What's one more oil line on that car?!

 

[Jack Olsen]

Those are mostly things you'd find in a machinist's shop. I bought them all used, and their function for me is as a big heavy object -- just as the picture illustrates. But the V-blocks are handy when you have to drill into a piece of pipe. The angle blocks are also great for clamping material at a 90° angle when I'm welding.

* * *

My plan with the wing is taking shape. I picked up a pair of salvaged Miata headlight pop-up motors from a wrecking yard today. I'm going to move the rear mount of the wing forward, and then have a motor down at the base of each upright, where the ducktail spoiler makes their drag-inducing bulk less of an issue. I'll use a push button to trigger the thing, with it re-setting to high-downforce when I release the button. The Miata motors have an 'in motion' indicating line, which I'll wire to a light on the dash. I'll also put a white decal on the leading edge of the wing which will allow easy confirmation that it's in one postition or the other whenever I check my mirrors.

I havent weighed the little motors yet, but that's the part that makes this a little agonizing.

The short throw on hydraulics (as well as the cost) makes me worry that my mechanical pieces would have to be more precise than my generally-sloppy workmanship allows.

Here's a video of a guy who used the same motors to turn his wing into an air brake. I emailed him, and he was very generous with his time and his suggestions.



Click on the image for video.

I'm still not 100% convinced this is going to make an actual difference in lap times. But my frugal self comforts itself with the idea that even though I'm out of pocket about $140 so far, the set of Bowden cables and the pair of 2-step motors will come in handy for a Halloween yard display (or something like that) if they don't earn their place on the 911.

 

[Outlawmws]

Well, they are certainly fast acting enough. (I'm surprised)

If these wing mods work out I can see, some fairings needed for those motors...

Jack, did you notice his strut wires relaxing, depending on the turn?

 

[DocRocket]

Interesting problem. I not an aeronautical engineer, so take these comments for what you think they're worth. As mentioned above, moving the pivot point closer to the middle of the wing is probably a good idea. There is a center of pressure for your wing in both the "flat position and the high drag position. I have no idea where it is, but it probably moves around a bit during the deflection of the wing due to the extreme reflex at the trailing edge. In order to maintain your desired fail safe system, I think the pivot point for the wing should always be behind the CP for both conditions AND the transitional positions between the extremes. In this way, no matter what fails the wing should revert to the high down force condition.
I am pretty sure your linkage could be far simpler and more direct. Moving the pivot point toward the CP will make the effort required to move the wing far less since the aerodynamic loads will tend to be self canceling. That said, with the pivot point anywhere behind the CP, the wing will not want to stay with the front facing forward unless it is restrained so it would be preferred if there was always some downforce coming from the wing. If it gets too flat, it will want to flip over the top and, of course, when it's in the max downforce position, it's trying to flip "under" the pivot points. I imagine the loads on the stops may be considerable. How much down force does the wing produce?
Just a few random thoughts. Perhaps some one who is an AE will chime in.

Cue the AE: At Low speeds (for AE, 130mph is low speed, less than Mach 0.3), the Aerodynamic center (where the moment/torque created by the forces on the wing is constant with angle of attack) is around the 1/4 chord mark, and this is true even for fairly high camber airfoils. If your wing has a 1 foot chord, the mean aerodynamic center is about 3 inches from the leading edge--and the forces are usually considered to act at that point. It will probably be easiest if you move the pivot point near the 1/4 chord location. I can see the argument that having the pivot towards the rear means the wing defaults to the high downforce setting, and I like the thinking here for this application, but you could move the pivot to the middle of the wing and get the same effect, and drop your actuator force in the process. Right now you need a force on the leading edge that is about 3/4 of your lift force when you actuate the movement. If you move the pivot to the center, your actuator force will be about 1/2 the lift force--and you still have the lift acting ahead of the pivot, so it will tend to push itself down without some action to hold it up.

 

[Mach5]

I may have missed the post, but you did confirm that the sanctioning body through which you race will permit active aero, right?

Sent from my SPH-L710 using Tapatalk

 

[Jack Olsen]

Well, they are certainly fast acting enough. (I'm surprised)

If these wing mods work out I can see, some fairings needed for those motors...

Jack, did you notice his strut wires relaxing, depending on the turn?

I'm surprised by how well they're working too, consideting the Miata owners manual suggests keeping them in the up position all winter to avoid problems with ice. He says he's able to hit the brakes at 130 mph and it flips up just fine.

And I did notice the wires. It could be that his uprights are bowing in, or that the wires are pretty slack to begin with. Hard to say.

Cue the AE: At Low speeds (for AE, 130mph is low speed, less than Mach 0.3), the Aerodynamic center (where the moment/torque created by the forces on the wing is constant with angle of attack) is around the 1/4 chord mark, and this is true even for fairly high camber airfoils. If your wing has a 1 foot chord, the mean aerodynamic center is about 3 inches from the leading edge--and the forces are usually considered to act at that point. It will probably be easiest if you move the pivot point near the 1/4 chord location. I can see the argument that having the pivot towards the rear means the wing defaults to the high downforce setting, and I like the thinking here for this application, but you could move the pivot to the middle of the wing and get the same effect, and drop your actuator force in the process. Right now you need a force on the leading edge that is about 3/4 of your lift force when you actuate the movement. If you move the pivot to the center, your actuator force will be about 1/2 the lift force--and you still have the lift acting ahead of the pivot, so it will tend to push itself down without some action to hold it up.

That's helpful to know. Thank you! I will definitely move the pivot point forward.

My airfoil is NACA 63-520 modified for additional camber with a third order equation.

In picking out an airfoil, it was important that it be thick (since I built the first version out of aluminum), and this one came out the best of the ones we tested.

And this was how its drag compared to the other ones I was looking at.

Here's a look at it with its Gurney flap (which was needed for my construction technique as much as anything). Theoretically, with my wing width and chord length, it should be generating over 400# of downforce. I'm sure the real number is substantially less.

And since I've got your ear, can you tell me what the correct way to measure angle of attack when an airfoil shape has been modified with additional camber? I'm not sure if measuring from the leading edge to the tip of the tail is correct with this shape.

And in truth, I'm still not at all clear on what the benefit, re: drag, would (or should) be if I'm able to reduce the angle of attack on the straights. But I figure it's at least worth trying.

I may have missed the post, but you did confirm that the sanctioning body through which you race will permit active aero, right?

I haven't gone club racing since my son was born, six years ago. If I go back to it with this car, I'll use the uprights and wing in a fixed position. But for now, I'm just trying to improve my lap time at my local track.

 

[DocRocket]

Usually it is measured from the leading edge to the trailing edge (where it was before you added the flap) as the "zero AOA" line. Some airfoil designers have used different conventions but the NACA series all use this convention. When you modified the camber, you moved the lift curve up and to the left. You will now stall it out at a lower AOA, but you have more lift at a lower AOA too.

The Lift Coefficient for a wing is lower than for an airfoil (infinite span wing), but your wing looks to have an aspect ratio (span/chord) of 8-10 or so, so the difference will not be too large, and if you have endplates, that will help even more.

Also, while 1,000,000 is not a bad estimate for your Reynold's Number, it would really be closer to 725,000 at 130 mph, so if the data is available at 500,000 or 750,000, you would be better off using that data. Airfoil data is normally not available at a large number of Reynold's numbers, and I am somewhat surprised you got any at 1,000,000, as it usually starts about 2,000,000.

Looking at the airfoil data, you should be able to cut the lift in half on the straights and cut the drag to 1/3 or so of what it is at max lift (might only be 1/2 after adding drag from the uprights and end plates). However, this is probably worth the effort as at this speed (130 mph ~190 ft/s) every 3 lb of drag is about 1 horsepower required (power required = drag(in lbs) * speed (in ft/s) / (550 lb*ft/s)/hp). So, if your wing is 400 lb of downforce, and taking into account some fudging for the uprights, etc, you probably have about 20-30 lb of drag, consuming about 10 horsepower, just for the wing. You could potentially cut that about in half, so that should be worth a little bit of top speed on the straights. You might also consider a notch at about 7 degrees AOA, as that would seem to be where the Lift/Drag is maximized. If you want to find the right location, find where Cl/Cd is highest (use excel or something else to calculate it based on the data in the curves, or you can draw a line from 0 CL and 0 CD to the point where it is tangent to the Cl v. Cd curve and that will be pretty much the right location. The endplates will also help prevent "induced drag" which is the drag induced by making lift on a finite wing--there is spillage over the tips, creating vortices, and adding drag while reducing lift--the airfoil curves only include "pressure drag" which is just due to the 2-D pressure distribution about the airfoil, creating both lift and drag, but this also includes the normal viscous drag.

Overall, depending on the exact aspect ratio, and the effectiveness of your end plates, you can probably count on about 90+% of the lift/down force you are calculating from the airfoil data, and about 150-200% of the drag. However, the trends will still hold pretty close.

 

[EdT]

Doc,
Glad you picked up on my cue. Just from looking at the pics at the lead in of the post, I was wondering if the wing is still "flying " at the high AOA or is is stalled which makes it act more like a flat plate with a lot of drag from the turbulence off the back side of the stalled wing. That is, is it still a wing or is it more like a whale tale that's not well fastened to the body. Not trying to hijack, just curious.

 

[DocRocket]

It's hard to tell just looking at things, because the AOA is measured from the wind hitting the "wing"--which is not parallel to the ground, because it is disturbed by the car. It is also not likely to be following the slope of the back window of the 911, it is expanding at some angle off the back of the roof. Jack posted some pictures of yarn tufts in the wind, but short of getting good, measurable photos of those tufts (minimal angles, etc), it is hard to tell exactly where the oncoming wind is at. One option would be to put an adjustable stop so you could alter the full "down" position, and do a few laps at each setting to find the one that feels the best. If you wanted to be more scientific about it, you could look at a straingauge setup to measure the compression in the uprights and maximize the downforce, but that might be a bit of overkill for the application (not if you were in NASCAR or on an IndyCar or F1 team, but Jack might find it overkill, or at least overcost). Could also use a multi input pitot probe to find the angle of the oncoming wind, this method provides AOA input to several fly by wire aircraft, but again, lots of $. Finding a way to put a camera up, square with some yarn tufts in front of the wing is lots cheaper.

 

[Squankum]

Is there an affordable way for him to put pressure sensors on the the surfaces of the wing, top and bottom side? Would that be useful data?

 

[DocRocket]

If you can find affordable, small pressure sensors, that are easily integrated to, for example a laptop computer, and will fit inside or on the wing without disturbing the flowfield, you could place a whole bunch of them on the upper and lower surface of the wing and generate a pressure distribution around the wing. From that, you could integrate to find the resultant force and break that up into lift and drag components. I suspect this would, overall be fairly expensive, and I am not aware, at this point, of any pressure sensors that would be appropriate here. The ones I have most often used are piezo electric and about the size of a penlight. There are smaller ones, but here, small costs more. Also, they generally are not that user friendly to just getting data into a spreadsheet or something like that, requiring some intermediate hardware and software to log the data. It is possible that some system like this could be built up for a reasonable cost, but I suspect that unless Jack recently won the Powerball, it would not be worth the cost to him, to get more data to support his efforts to take a couple more seconds off his lap time. I think that some experiments with the wing in different positions would give the data he needs, and not cost a fortune (although the gas cost might be significant ).

I have done a fair bit of this and experimental aerodynamics is an expensive business. Only exceeded in cost per useful data by computational aerodynamics.

 

[Jack Olsen]

Wow. Thanks very much for getting involved, DocRocket. I've done some testing with tufts, and -- while it's not as precise as it could be -- my incoming air is less than the 17-19° angle of the rear of the car, probably close to 11°.

Now, I've never had what I consider to be a reliable way to measure my actual AOA because of the camber worked into my airfoil shape. So when I did some 'poor man's windtunnel' testing, which involved 100-mph straight line runs in opposite directions (to correct for local winds) with my on-board data logger recording changes in front and rear ride height. (I used ride-height sensors from a Lincoln Continental air ride system, which I got for five bucks on Ebay, since I haven't won that Powerball yet.) I could see ride-height differences in 4° changes to the AOA:

As it happened, there was a problem in the way I calculated the front-end data for that graphic, and the corrected numbers ended up not showing the front lifting up as the back went down. Using the run with no wing and no splitter as a baseline, the following happened:

When I put on just the front splitter, the front end runs .41" lower, and the rear runs .04" higher.

With both the rear wing and the front splitter, and the wing set at 12 degrees, front end runs .24" lower, and the rear runs .01" lower.

With both the rear wing and the front splitter, and the wing set at 16 degrees, front end runs .09" lower, and the rear runs .17" lower.

With both the rear wing and the front splitter, and the wing set at 20 degrees, front end runs .08" lower, and the rear runs .26" lower.

I also did some yarn tests to look at adhesion:

Looking at this, it makes me think I'm overestimating my AOA, since there's still so much adhesion at the 20° setting. I would guess that I would still see downforce imporvements even after I'd lost adhesion on the rear 1/4-1/3 of the wing. But again, I haven't had a clear idea of how to measure AOA with my camber-curved wing.

 

[DocRocket]

So, if your airflow is about 10 degrees below horizontal over the back window, you should be able (based on the airfoil plots above) to run the wing at about 24 degrees for max lift/downforce. I would stay with the 20 degrees. You also don't (I think) want to push too close to the max CL, because it is usually right before stall. The last thing I would think you would want is a stalled wing in the middle of a corner (no downforce, just drag). I am a little dubious of the curve for the NACA 63-520 in it's post max-Cl behavior. That is not what Cl v. Alpha curves usually show, usually it is a fairly sharp decline in Cl after the max, not that kind of gentle curve. Along with the rest of the curve, it makes me think that is a curve fit to the real data, using some kind of quadratic or cubic function. Notice how some of the other airfoils are a little less smooth in the whole curve? I think those are just connecting the dots in the data, the 63-520 is, I think a curve fit. In any case, I would think that you (in the 20 degree AOA picture) are operating at about 10 degrees AOA if you are measuring from the chord line (leading edge to where the trailing edge would be without the flap) to the oncoming wind. At that point you have 90+% of the lift you are ever going to get, and you are fairly far away from stalling the wing. Part of the problem in stalling the wing is that you are in turbulent flow behind the 911, so there could be some differences in the wind direction from instant to instant, but the more critical part is that you may find yourself dealing with the flow from another car interfering with the flow over yours: That M3, or 'Vette, or Ferrari you are trying to pass is going to have an impact on the flow over your 911, so staying a little bit shy of the limit here will keep you from instantaneously unloading the rear end in the turn. A little bit of extra downflow will just push you closer to max Cl and give you slightly more grip in the turn, rather than causing the car to lose 400lb of downforce and the attendant lateral grip and taking an impromptu tour of the infield in a 1080 spin.

The pictures at 20degrees are showing a bit of separated flow that appears to be starting at the back 25% of the suction surface of the wing--this is pretty typical high AOA behavior for an airfoil, and is pre-stall behavior. At max CL, I would expect the separated flow to start at the point the surface starts to curve away from the flow, about at the 1/3 from the front line for this airfoil. At a higher AOA, you will have stall and the flow will be separated over essentially the whole suction surface.

 

[Jack Olsen]

That's helpful. I'll aim for 10°, measuring leading edge to trailing tip (absent the Gurney flap) and -- if I get some time, look at ride height changes if I take a little away from that or add a little.

I dug up an old set of uprights I'd previously stopped using, and pulled out the cables and started from scratch with some new top pieces to go between the wing's feet and the top of the uprights. My goal is to have a pivot point about 1/4 of the way from the leading edge of the wing, and then a rigid support farther back that will be raised and lowered by a simple step motor.

Unfortunately, this means letting go of my scheme to control the leading edge of the wing, and it makes things slightly more precarious when it comes to the wing 'defaulting' to the correct position in the event of a failure. However, the new control isn't a cable -- so the chance of a snapped or slack cable is gone. And I'll have a pair of independent feedback mechanisms to alert me to a positional problem with the wing.

As I'm working it out now, the new system will use a button on the steering wheel. When the button is depressed and held down, the wing will go to a low drag setting. If my finger slips off the button, it will go back to high downforce. If I touch the brake pedal, it will also send the wing back to high downforce. I'll have a pretty big set of LED lights in the cabin that will confirm when the wing is changing position, and I've also got an idea for a simple visual reinforcement that will let me know the wing has returned to the correct position when I'm at the end of a straight.

Anything that reduces the simplicity and default-to-a-deadman-position makes me nervous. And I would not encourage anyone else out there to experiment with this sort of thing. I'm only working on it to test out at my home track, which has no walls to hit if something goes wrong, and is also a track I frequently drive in the same car without my front splitter and rear wing. So I think I'm in a safer-than-usual position with this testing. But again: I wouldn't encourage anyone to try this kind of thing without being certain they're not going to hurt themselves.

To make four identical top-of-the-upright pieces, I used my tried and tested method of cutting an oversized version of the piece with a band saw, then using a more-precisely-sized wooden buck and my table router to shape the final piece. I adhere the aluminum to the wood piece and use a trim bit to duplicate the shape from the wood to the aluminum.

Pieces like this go everywhere. It's important to have a face shield and also long sleeves.

Here are the pieces. The inner ones are slightly thinner than the outside ones.

This is the arm that will raise and lower the trailing end of the wing.

The motors that control the wing (there is one on each side) are from a Miata's pop-up headlight system. I've tucked them behind the spoiler, where I think they'll have a minimal impact on the car's overall aero profile.

The control arms will be made with the same carbon fiber tubing you can see above the motor as a crosspiece between the two uprights. But the vertical ones will have Heim type end pieces. I've also got a relay on order that I think will be able to give me the ability to always have the motors want to go back to the high downforce setting except when I'm specifically holding down the button to move the rear edge of the wing down.

 

[DocRocket]

One thing to keep in mind when you put the pivot at the 1/4 chord point, is that there will be a moment/torque from the airfoil itself. It will likely be about 5% or so of the lift*chord, but in this setup the motor is the only thing countering it. It will tend to bring the trailing edge down in your set-up (you can get a better idea of the numbers if you can find moment coefficient data for your airfoil, it is usually around -0.05 or so, but may be a bit higher with your degree of camber). Just remember that the motor and shaft must counter that moment ALL THE TIME. The idea of the aerodynamic center is that the moment does not change with angle of attack, only the forces (lift and drag), but they are considered to act through the aerodynamic center. Therefore all the serious support can be there (1/4 chord or so), but the rest of the support must still deal with the moment, which could result (for you) in dealing with about (it looks like) a 0.05*400lbs*1ft torque, or about 20 ft*lbs, which doesn't sound like a lot unless you try to counter it with a relatively short moment arm, in which case you have put 100's of lbs on that motor. Just be aware while you are setting this up, I don't have a good feel for how much force pop-up headlights take, but I doubt it is more than 5-10 lbs, and you are going to be asking for at least 27-30 lbs out of the motor (with the max lever arm you can deliver on your wing).

 

[Jack Olsen]

Am I right in assuming I can divide those (approximate) numbers in half for the two motors?

Of course, I've got no specs at all for the motors. This is junk-bin engineering. But they've been working for awhile in this guy's setup, which is using the wing as an air brake instead of changing its angle for drag reduction. I'll post the video link again:



(Click on the image for video.)

Now, he's got a somewhat smaller wing -- 11"x58" compared to my 11"x68". And his top speed in the video is only 122 mph, compared to my 134 mph. And his motors might be doing less work, since it looks like they are able to utilize the leading half of the wing's downforce as an aid when he brakes at high speed (and the leading edge of the wing comes down). And then when he's done braking, the car is going slower (and the wing is generating less force) so the 'hard' job of pushing the nose back up is done when there is less force needed to do it.

 

[Jack Olsen]

Lunch break progress today.

First off, I realized that my motors were going to generate way too much travel. So I shortened the control arms on the units, which will also give them a mechanical advantage when it comes to doing their job.

DocRocket's point about the motor bearing at least a portion of the downforce when the wing was fixed in either position also got me to re-orient the control arm slightly so that they're in line with the force they'll be carrying. This way, there isn't the same kind of leveraged pressure on the gear in the motors. At least, it's reduced to its lowest amount. Right? In this picture the wing is in the 'up position.' In the down position, the arm is oriented 180° in the other direction.



I don't have the carbon fiber or Heim end pieces yet, but I cut a couple of pieces of L stock and attached a motorcycle battery. Here's some video:



(Click on the image for the video.)

There is an 18° net change in the angle of the rear wing when the motors move it.

 

[sean Buick 76]

Very cool!!! keep us posted and lets see some video of that car flying around a corner at 100 MPH please???

 

[flybefree]

Mr. Olsen you are in to something interesting...cool project!

v/r

Shaun

 

[DocRocket]

a) Yes, you can divide the total forces by 2 since there are two motors (and they are arranged symmetrically, etc)

b) I like the thought on putting the motors where they use 180 degrees of travel and end up with loads just pushing on the motor shaft--should minimize the torque on the gears, still need to deal with the loads on the bearings, but those were there anyway you set this up.

Yes, the force in the rods is fixed and your load into the motor is determined by the lever arm. OTOH, so is the load they can deliver, so there may be a tradeoff there.

 

[1949 caddyman]

Could you use some type of an up & down mechanical stop to take some of the load?

 

[Outlawmws]

The shorter those motor arm are, the easier they move something, and the load on those gears DOES go down. Also the load on the motors mount should go down in the parked positions. good and good!

As long as the down position does not go so far as to get you into lift on the tail, you should be golden.

 

[Jack Olsen]

Thanks!

I should probably take each motor apart, photograph and re-lube the gears. That way I can check after each event and confirm whether I'm actively tearing the things apart.

I tested out a red vinyl stripe on the very underside of the wing. It's the simpler of my two feedback tricks. When the red color is visible in my rear view mirror, I know the wing is in the low-downforce position. There will also be a dash light -- but lights can fail.