NSX aerodynamics

[MvM]

I’d expect that a car’s Cd would improve by simply lowering it an inch. According to the Bosch Automotive Handbook, lowering a standard car by 30mm without any other changes should decrease the Cd by about 5%. The actual results will depend on the individual model of car being tested, but Mercedes, Volkswagen, etc. do lower all of their “efficiency” models to decrease wind resistance.

The bottom of a car will probably always have more wind resistance than the top simply because of the spinning wheels jutting down into the airflow. When you lower the car, some of the air that would have flowed past the rough underbody will flow over smoother top instead.

Do you (or anyone else) know if it makes any difference if you lower your car all around the same distance of if you lower the front more than the rear.

Apart from giving the car a bit more agressive stance, in my perhaps simple view, lowering the front more than the rear (say: 1.25" in front, 0.8" in the rear) would lower air pressure under the car at the rear.
And at the same time, it would position the rear wing higher above the car relative to the nose so it might create more downforc as well.

 

[greenberet]

Here you go, from the second book Ty B recommended, "Race Car Aerodynamics" by Joseph Katz:



The first book Ty B recommended is even better and if you can read German, get the original. "Aerodynamics of Road Vehicles" is a 1998 translation of Wolf-Heinrich Hucho's book "Aerodynamik des Automobils". The latest edition in German is from 2005, is more up to date than the English translation and I agree with Ty B's assessment, it's the bible for road car aerodynamics.

 

[MvM]

Here you go, from the second book Ty B recommended, "Race Car Aerodynamics" by Joseph Katz:



The first book Ty B recommended is even better and if you can read German, get the original. "Aerodynamics of Road Vehicles" is a 1998 translation of Wolf-Heinrich Hucho's book "Aerodynamik des Automobils". The latest edition in German is from 2005, is more up to date than the English translation and I agree with Ty B's assessment, it's the bible for road car aerodynamics.

I can read German ok, but in Geman, but I am much more familiar with English when it comes to technical terms.

Thanks for the picture, it does not explain everything though.

Am I correct in assuming that the Alfa angle is negative (- minus) if the nose of the car is lower than the rear.
If that is correct then you would indeed increase downforce if you lower the front more than the rear. Having the TEIN RA suspension on my car allows me to do exactly this :smile:

 

[greenberet]

Am I correct in assuming that the Alfa angle is negative (- minus) if the nose of the car is lower than the rear.

Yes.

The graph doesn't state how the downforce was distributed front to rear, however, so be careful. With a downwards rake dialled into the suspension, the underbody will probably generate downforce over the font axle. If the rear wing is in cleaner air, maybe that will generate some more downforce over the rear axle. How the overall balance would really be impacted for an NSX, God only knows.

 

[MvM]

Yes.

The graph doesn't state how the downforce was distributed front to rear, however, so be careful. With a downwards rake dialled into the suspension, the underbody will probably generate downforce over the font axle. If the rear wing is in cleaner air, maybe that will generate some more downforce over the rear axle. How the overall balance would really be impacted for an NSX, God only knows.

I understand what you mean.

Would be nice to have something to measure you road clearance at the front and rear wheels while driving.
That would be a way to measure your downforce as your speed increases. Especially if you have a free road (like in Germany) where you can (legally) get to very high speeds.

 

[Mac Attack]

Macattack, I'd have to disagree with you. The NSX-R flow is clearly diverging at the rear of the car and the streamline widens post wing. The NSX shows the airflow converging at the back of the car and the streamline has much more continuity after passing the wing. This to me explains where the NSX-R gains Cd compared to the NSX; more downforce too.

Ty,

I agree. When I meant the NSX-R wing is more effective, I meant from a downforce perspective. Looking at the streamlines, like you concluded, the NSX-R wake is larger than the standard NSX (meaning higher Cd like you stated). Obviously, the tradeoff for some increased rear downforce was an increase in drag.

I used to have a link to an official Honda page that gave the technological highlights of the NSX-R. It was like a 6 page thing that detailed Honda's "tweaking" in suspension, aero, engine, weight, etc, and it gave specific values on the impact their aero changes made to lift and drag coefficients over the standard NSX. I can't find it now :frown:

This picture looks fabricated (as in not a real picture of a smoke trail or CFD) and thus irrelevant. I think i've seen that picture before without the blue lines around the car.

Perhaps, but it should be obvious to most folks participating in this thread that these streamlines look accurate enough to be realistic. Just thought I would throw it in there with the other pics since I haven't run across them before on this site :wink:



Like greenberet said, lowering a car "should" decrease the drag slightly in the real world. In fact, commercial vehicle manufacturers use front airdams that extend pretty low in the front to balance the amount of airflow going through a vehicle's radiator with the total airflow underneath. Having a low front airdam will increase the "frontal area," but the tradeoff in not having the air get disturbed underneath the vehicle is worth it from a drag perspective. Of course, that has to be balanced with brake cooling, airflow through the radiator/engine compartment, and how far down you can extend the dam before it breaking all the time during normal consumer driving.


When I used to race, besides the cool IR thermography cameras for tire temp (http://www.nsxprime.com/forums/showthread.php?t=142671), we also instrumented our coilovers with LVDT's, or linear velocity displacement transducers to measure their extension as a function of time over driving. With this, you could kinda get an idea of downforce or lift generated on a straight flat section of a course. F=kx, so knowing your spring rate and displacement from the LVDT's, you could measure the change in load. Pretty simple, besides giving you important info on what your coilovers are doing anyways. This is what I recommended to Driving Ambition here (also discussing aero too): http://www.nsxprime.com/forums/showthread.php?t=134440

Oh, I don't encourage increasing the rake of our vehicles because it will drastically change the static/dynamic weight distribution and make it prone to oversteer.

Dave

 

[goldNSX]

I used to have a link to an official Honda page that gave the technological highlights of the NSX-R. It was like a 6 page thing that detailed Honda's "tweaking" in suspension, aero, engine, weight, etc, and it gave specific values on the impact their aero changes made to lift and drag coefficients over the standard NSX. I can't find it now :frown:

This one? http://world.honda.com/NSX/technology/t3.html
This info dates back some years and I'm not sure how long Honda will provide us with the info on their website.

 

[Gansan]

For posterity's sake, I printed each web page as a pdf for safe keeping. You can't trust websites to stay around forever.

 

[goldNSX]

For posterity's sake, I printed each web page as a pdf for safe keeping. You can't trust websites to stay around forever.

Do you offer it as a download? :wink:

 

[Gansan]

Do you offer it as a download? :wink:

https://files.me.com/gkawano/d0rnyo

There you go!

 

[goldNSX]

Thanks!

 

[Ty B]

A german car magazine had measured some values far back. Anyone with the numbers?


Yes and no. If you use simple wings at the front or the rear then yes, downforce will increase drag. But if you work on the underbody and speed up the air with a diffusor (well more F360 like than the infinitesimal one on the NSX) you gain downforce for free. A lot of the NSX-R NA2 improvement was done on the underbody, esp. the front part, esp. compared to the 91 model. It helped the air under the car to the rear part of the car. It's said that it's more important to control the air flow in the front half of the underbody than in the rear one. Have a look at this article: http://autospeed.com/cms/A_110872/article.html where a guy tried to reduce the drag by convering the rear underbody. The front of this car is a brilliant work of aerodynamics but the rear 'looks' unfinished. The rear of the NSX underbody also looks 'unfinished' or like a 'compromise' due to therminal requirements (enigne overheating?).

Flushing the wheels with the fender is not so critical and could be a drawback, sep. if the wheel itself allows for a lot of turbulences within the fender which results in increased drag. Have a look at the wheels in the article, they're Pizza-like. This is aerodynamically optimal. From my memory I have in mind that with larger rims it took longer to reach 150 mph+.

It's been several years since I had my CFD course, so the details are all now a bit fuzzy. I would still argue that lowering any car an inch, by itself, does not improve the overall Cd for that car. In most cases removing the mirrors will drop as much Fd as lowering a car by an inch.

This infers a reduction in area, not a change in Cd. Could there be some loss of turbulance from masking the tires as well as reducing the frontal area, perhaps. The largest gain is the reduction of frontal area from masking the tires. A far larger drop in drag can be gained by the placement of deflectors upstream of the rear tires and the use of an air dam and sideskirts.

The rear diffuser does not speed up airflow. A properly designed diffuser slows down air flow. The trick is keeping the flow attached while you increase flow area and increase pressure. You want to fill the rear void with equal pressure to the surroundings. This eliminates the turbulance and drag from the collapsing slipstreams off an abrupt rear end.

This does not happen for free. There is still drag created, but less overall than the abrupt disruption.

 

[goldNSX]

It's been several years since I had my CFD course, so the details are all now a bit fuzzy. I would still argue that lowering any car an inch, by itself, does not improve the overall Cd for that car. In most cases removing the mirrors will drop as much Fd as lowering a car by an inch.

This infers a reduction in area, not a change in Cd. Could there be some loss of turbulance from masking the tires as well as reducing the frontal area, perhaps. The largest gain is the reduction of frontal area from masking the tires. A far larger drop in drag can be gained by the placement of deflectors upstream of the rear tires and the use of an air dam and sideskirts.

The rear diffuser does not speed up airflow. A properly designed diffuser slows down air flow. The trick is keeping the flow attached while you increase flow area and increase pressure. You want to fill the rear void with equal pressure to the surroundings. This eliminates the turbulance and drag from the collapsing slipstreams off an abrupt rear end.

This does not happen for free. There is still drag created, but less overall than the abrupt disruption.

I understood a rear diffusor in the way you described but from a different point of view. You're right about the air being slowed down in the diffusor. As the slowed down air exits the diffusor it will be 'thrilled' or sucked out/speed up by the air not passing the underbody at the rear. This will speed the slowed down air in the diffusor up and this air speed the one in the underbody up too which finally results in downforce on the chassis. Did I explain it complicated enough? Am I right or completely wrong? :wink:

 

[Gansan]

Yeah, you guys are describing the same thing from different points of view. The diffuser helps the underbody air smoothly decelerate into the wake. Another way of looking at it is it lets the low pressure from the wake behind the car better reach under the car and help evacuate the air down there, which makes the underbody air flow faster. Faster is lower pressure.

I think also there is a coanda effect of the underbody air getting pulled up as the diffuser volume increases going back. That change in the direction of the air also adds downforce. You can see that as a spike in low pressure right where the diffuser starts and the air turns upward.

 

[greenberet]

A far larger drop in drag can be gained by the placement of deflectors upstream of the rear tires and the use of an air dam and sideskirts.

I agree that wheel spoilers and side skirts most likely decrease drag, but the latest 2005 edition of “Aerodynamics of Road Vehicles” (in German) takes a different view of the front air dam. According to the book (pages 236 and 521) the wind resistance of cars with rough underbodies goes down if you increase the depth of the front air dam until the point that the air dam is almost the lowest point of the underbody. If you make the air dam deeper than that, overall wind resistance will start increasing again even before the frontal area is impacted although front lift will keep going down. It also states that if the car has a smooth underbody, a deeper front air dam will increase drag, period.

That’s surely a generalization that doesn’t apply to every car ever made, but to decrease the aerodynamic drag of an NSX, it’s probably a better idea to make the bottom of the car smooth rather than fit a deeper front air dam.

In "Aerodynamics of Road Vehicles" as well as in "Race Car Aerodynamics", there’s an interesting chart regarding the downforce and drag created by diffusers with different angles. As would be expected, downforce increases the steeper the diffuser as long as the flow stays attached. Drag is minimized at around 3-5° (5° is what Honda used for the first generation Insight) and it goes up again at steeper diffuser angles. As goldNSX said, a properly designed underbody and diffuser can get you downforce effectively for free and at diffuser angles of 3-5°, Mother Nature actually pays you well to take it. The diffuser speeds up the flow in the throat of the car's underbody and at the same time, decreases the effective base area at the rear of the car.

 

[CL65 Captain]

Don't forget there is a difference when it comes to induced drag which is a by product of lift (or downforce in this case) and parasitic drag. Total drag is the sum of these two which are inverse.

 

[Ty B]

Sigh, I wasted my youth! I have to rely on the imperfect google translations, or the work of some bored author. :biggrin:

I've been working with "The Handbook of Hydraulic Resistance", by Idlechick and "Fan Engineering". In an effort to normalize air flow, in 2 Id diameters past a non uniform disruption, I need average flow and temp for a control feedback loop. Turns out stainless sifting mesh can be used at a minimul pressure drop irregardless of Reynolds number. For the flow parameters, A large wire diameter causes complete turbulance and provides perfectly normalized temp readings within 10 wire diameters. I luv it when one thing can be used to solve 3 or more needs.

The ole high wing vs low wing discussion has reared its ugly head. It depends on how intricate the rear design is. A completely flat bottom with no other devices yields no down force, just food for thought.

Taken by itself, an air dam has limitations and a maximum gain is always the compromise. Same for high speed vortex tunnels leading to the rear diffuser. I understand Ferrari has gone in circles on how to control downforce so that straight line speed is not lost. Hence their moveable devices at the front that control the high speed air running down the tunnels.

The diffuser is filling the low pressure drag zone in the void at the rear of the car. Instead of focusing on downforce, look at it from the perspective of pressure. More drag is being shed with a diffuser equalizing the rear low pressure zone than the diffuser is generating. The diffuser is still generating drag, yet equalizing the pressure void at the rear of the car offsets the drag that was generated from the slipstream going turbulent. It's a net gain, but not free. The definition of a diffuser is a device that slows flow and raises pressure. It just so happens a diffuser mounted under the back of a vehicle traps enough high speed flow to generate rear downforce while it partially equalizes the trailing rear low pressure zone resulting from the turbulence being generated and negates some of that turbulence. One thing and 2 positive outcomes, nice.

3-5 degrees ensures that attached flow is not lost and limits the overall diffuser flow. Only flowing enough air to partially fill the vehicles rear turbulent low pressure zone. Upwards of 15 degrees can be safely used while maintaining attached flow in the diffuser. I think the reason that they only used 3 to 5 was because they couldn't get more front downforce cleanly. Any more and Honda would have lost the balance, front and rear, they were aiming for.

Also, the rear underbody layout of the NSX limits the run of the diffuser. Following Greenberet's aero experiments, I can appreciate the head scratching you've had to do. :smile:

 

[Hrant]

Don't forget there is a difference when it comes to induced drag which is a by product of lift (or downforce in this case) and parasitic drag. Total drag is the sum of these two which are inverse.

This graph looks very much much like pricing in microeconomics. Where the dotted lines are variable and fixed cost while the undotted line is the average cost :biggrin: What is missing is the demand curve :tongue:

Now, can someone explain all this engineering gobbledygook in simpler English so the lay-folks we can appreciate what you are really/actually saying :wink:

 

[SFNSXguy]

This graph looks very much much like pricing in microeconomics. Where the dotted lines are variable and fixed cost while the undotted line is the average cost :biggrin: What is missing is the demand curve :tongue:

Now, can someone explain all this engineering gobbledygook in simpler English so the lay-folks we can appreciate what you are really/actually saying :wink:

"Leave your road car bodywork alone." :biggrin:

 

[Ty B]

Don't forget there is a difference when it comes to induced drag which is a by product of lift (or downforce in this case) and parasitic drag. Total drag is the sum of these two which are inverse.

But are they in ground effect?:smile:

 

[greenberet]

Now, can someone explain all this engineering gobbledygook in simpler English so the lay-folks we can appreciate what you are really/actually saying :wink:

Unfortunately, the terms used when discussing aerodynamic drag aren’t standardized. In the induced vs. parasitic drag chart above, I assume the author defined parasitic drag as the combination of viscous drag and form drag.

Definitions

Viscous drag (also known as skin friction drag): The drag created when air flows over a surface instead of flowing in a free stream. The first layer of air molecules over the car’s surface is clinging to it, practically standing still. The air molecules just above those move a little bit faster. Only several millimeters above the car’s bodywork is the air at its normal free stream speed. That layer of air moving below the free stream speed is called the boundary layer. When the air sticks to the surface of the vehicle instead of moving at its free stream speed, drag is created.

Form drag (sometimes called pressure drag): This is the drag caused by the basic shape of the vehicle itself. Cars do not have particularly aerodynamic shapes compared to airplanes or raindrops. The blunt front ends, sharp transitions from the windshield to the side windows, wheels jutting out into the airflow below, wheel cutouts along the sides, etc. all cause drag.

Induced drag (also known as vortex drag): The drag caused when air is put into a spin between areas of high and low pressure. Wings keep airplanes aloft by creating areas of low pressure on top and areas of high pressure beneath. At the wing tips, the high pressure air below tries to spill up into the low pressure zone above. As it does that, it is put into a spin – those are the trailing vortices you can sometimes see. To put the air into a spin energy is required and that’s drag on the airplane. The winglets modern airplanes have are air dams that try to prevent the high pressure air below from rotating up into the low pressure zone above. The endplates on race car rear wings serve the same function. If you could make the winglets or endplates infinitely large, you could probably completely get rid of the induced drag. However, that’s firstly not possible and secondly you’d then have infinitely high skin friction drag.

Interpretation of the induced vs. parasitic drag chart above

The induced vs. parasitic drag chart above is for an airplane wing. Airplane wings have very aerodynamic shapes so they have little form drag but the higher the speed, the higher the skin friction drag will be. The induced drag is directly proportional to the wing’s lift coefficient – how much lift it generates per mph. The airplane weighs a certain amount and at high speeds the wing only has to generate a small amount of lift per mph to keep it afloat. At low speeds, however, the wing has to generate a lot of lift per mph and that causes tightly spinning vortices at the wingtips. The chart shows that at a certain airspeed you minimize the sum of skin friction and induced drag. Above or below that airspeed, overall drag increases.

Conclusions

In general, you want to reduce overall drag. The winglets on the tips of modern airplane wings reduce overall drag because they decrease the induced drag more than they increase the skin friction drag. Properly designed rear diffusers on cars do the same thing, decreasing overall drag just as Ty B explained above. And in addition to decreasing overall drag, they can help you generate net downforce at the same time.

Recommendations

For an NSX, I personally think the aerodynamically best thing to do would be to give it standard 2002+ bodywork, lower the car as far as is acceptable for you, put wheel spoilers in front of all four wheels, and build a completely smooth underbody with a diffuser that is angled upwards 3-5° and starts as far forward as possible. The car would still be perfectly practical but drag and lift would both be reduced. If you need more downforce over the front or rear axle, you will need to start sacrificing straight line speed with a steeper rear diffuser, a bigger (perhaps 2002+ NSX-R) rear wing, a front splitter, etc. All those things are likely to increase drag and make your car slower in a straight line but they may make it faster in turns.

 

[shawn110975]

I was told by a very smart man to make the car 50% less drag is to

" remove the windsheild, and rear window/ and engine glass and wear a helmet " the drag will be reduced by 50% LOL

 

[greenberet]

A completely flat bottom with no other devices yields no down force, just food for thought.

Yes, but a rough underbody like an NSX’s that slows down the air beneath it yields lift.

A smooth underbody without a properly shaped inlet or rear diffuser doesn’t create downforce in and of itself, but getting rid of that lift is already a worthy achievement. And if you add a diffuser you should be able to reduce drag further and start generating downforce.

 

[Hrant]

Recommendations

For an NSX, I personally think the aerodynamically best thing to do would be to give it standard 2002+ bodywork, lower the car as far as is acceptable for you, put wheel spoilers in front of all four wheels, and build a completely smooth underbody with a diffuser that is angled upwards 3-5° and starts as far forward as possible. The car would still be perfectly practical but drag and lift would both be reduced. If you need more downforce over the front or rear axle, you will need to start sacrificing straight line speed with a steeper rear diffuser, a bigger (perhaps 2002+ NSX-R) rear wing, a front splitter, etc. All those things are likely to increase drag and make your car slower in a straight line but they may make it faster in turns.

Appreciate the conclusion/recommendation. I figured as much too in reading all the whiz kid posts :wink:

I have the 02 NSX-R underbody parts, as well as the rear diffuser. Also a NA1 ProCar vented hood. The front lip is a marginal improvement with the GT lookalike - more esthetic if you ask me, and the car is lowered about 3/4"-1".

I sure can use a bit more downforce in the rear to keep up with the pimping weekend boy racers but a larger wing is not an option. I like to keep it as stealth as possible :tongue:

 

[Ty B]

Yes, but a rough underbody like an NSX’s that slows down the air beneath it yields lift.

A smooth underbody without a properly shaped inlet or rear diffuser doesn’t create downforce in and of itself, but getting rid of that lift is already a worthy achievement. And if you add a diffuser you should be able to reduce drag further and start generating downforce.

Shhhhh, I was listening to all the head scratching. Very nice summation, that would have taken a couple hours to put together only after I dug out my John Anderson texts. Nice over all discussion as well.

Improving the airflow, without screwing it up, takes a series of pre-thought out changes. You can not change one item at a time with the aim of making progessive improvements. Single changes to individual pieces often end up conflicting with each other. It is best to package together a group of alterations and then fine tune the balance.

Can you design it once, slap it on and be perfect? No. Can you make changes to a grouped set and end up with a finished package in less iterations than item by item changes? Yep. Can you eyeball it? Nope. Can you take multiple manometer readings and crunch the data, make adjustments, re-test and yeild an aimed for target? Yep. Is it fun to do? Depends on your personality type.

The data honda provided, as pointed out by Gold, is amazingly revealing in that they had to add so much front aero to gain the 40/60 aero balance for the NSX-R. The NSX it seems is a plowing pig at speed. Wind vaning with a huge static margin to keep us mere mortals safe. Wich change made the largest contribution? Venting the radiator flow over the top of the car, or sealing in the front assembly with a flat panel?

As to lowering the car, how much is reasonable? Keep in mind the roads and transitions you are going to deal with. I hate scraping, speed bumps, driveways, parking building transitions; the front overhang is limiting in stock 4x4 mode as it is. I would hate to bottom out at speed on some mountain road with shear cliffs or shear drop offs and no guard rails; going ultra fast on the open freeway is so frowned upon around here as well. Some flexible side skirt extensions gain the same results without giving up ground clearance.