Blog

The Veloseter N undertray does a good job of covering the mechanical bits that cause drag under the car. It doesn’t have diffusers in the wheel wells (like a Porsche or Tesla), but there’s a small dam in front of each tire to reduce drag. There are holes for cooling and access, and the general design is one of convenience rather than performance.

The top-side is more fashion than function. The OEM front fascia has lots of vents, both real and fake, and some of the bodywork is angled to hold a small amount of downforce. As factory aero goes, it’s pretty decent, but there’s also a lot of room for improvement.

I’m going to make my own splitter, but before I do that, I’ll take a look at what’s available for purchase.

Aftermarket front lips

Front lips that are lower and/or longer than the OEM lip are a popular modification. They can reduce drag by reducing the amount of air going under the car, and hold a small amount of pressure on top, making some front downforce.

But if the front lip isn’t integrated into the undertray, it might increase drag due to flow separation between the lip and undertray. In general, I don’t think front lips add much performance and are largely poseur junk.

Aftermarket splitters

Functional front aero starts with a splitter. Splitters reduce drag by keeping some air from going under the car, but the real function of a splitter is to make downforce. I wrote about common splitter mistakes in Your Splitter Sucks!, and if you haven’t read that, it’s required reading.

With that in mind, I’ll review some of the splitters I’ve seen online, bash them mercilessly, and then show you something a fuckload better.

Aero Blitz splitter

The Aero Blitz splitter is half way between a front lip and a splitter. You can see it’s pretty small. There’s some surface area to hold pressure on top of the blade, but it’s not going to be a very useful amount. Anything that fastens with double-sided tape and self-tapping screws isn’t meant to make downforce.

The shape of the Blitz splitter follows the shape of the front fascia and gives it a “Batarang” shape. Most of the aftermarket splitters do the same, and I don’t like it. Functionally, it just doesn’t make sense. But then again, anyone buying this cares that it looks like a splitter, not that it functions as one.

Aerotekk splitter

Aerotekk has done a decent job making a product that looks great and is functional, at a reasonable price. This is a good solution for someone looking for a little bit of front downforce.

The mounting points look OEM, but if you read the description, it says that it doesn’t come with any brackets, and that mounting the splitter is up to you (or as they say, a professional). So is this just a flat an alumalite blade mounted flat against the undertray?

I don’t see any height adjustment, but a clever person could find a way to do that when they figure out the mounting brackets. With height adjustment you’d also need an air dam to fill the front gap.

You might also install splitter diffusers, but this would take away from the rigidity (alumalite isn’t that strong). I don’t see splitter support rods, nor what they attach to. Anyway, an OK starting point for DIY modifications.

APR splitter

The APR carbon fiber splitter is about twice the price of Aerotekk’s, has the same cutouts for the wheel deflectors, but looks to have a different mounting solution. It’s also dead flat, so it’s not doing much to create suction.

The splitter has some thickening and rounding of the front edge, but could be larger. The shape of the front lip has the bat-wing shape that follows the contour of the bodywork. Whatever.

It looks like there might be some height adjustment via the splitter rods, but the mounting of said rods is sus. Do they adjust for height and if so, how much? Are they really just fastened to the plastic radiator grille? How much weight can they hold? All said, I’m skeptical this does anything more than look like a splitter.

Street Aero splitter

The Street Aero front splitter is made from 6mm alumalite, and a consequence of the material is that it has a sharp leading edge, which results in flow separation. The splitter doesn’t seem to have any height adjustment, and appears to have negative rake (pointed upwards)!

The shape is the bat wing I’ve come to expect out of any of these appearance-first designs, but this one has larger tire cutouts for no good reason, and there’s no oil filter access, so I guess you have to remove the splitter every time you change the oil?

Ventus splitter

Ventus Autoworks appears to have the same, flat, flimsy, do-nothing splitter as everyone else. I’m not sure of the construction, but it claims to have an outer layer of carbon. (Actual carbon or vinyl, it’s hard to tell.) The splitter attaches via splitter rods that are equally suspect as the APR ones.

You can get the APR splitter with little vertical winglets on the end, which probably act as an air fence. The Veloster N already has air fences in the fascia, but maybe more is better.

I will say it looks good, and blends in with the style of the car. Even the winglets are pretty tasteful when you see the whole package. But does the splitter do anything for performance? Since it’s not much lower and there’s nothing happening underneath, it can’t be doing much.

DIY splitter

None of the aftermarket splitters do what I want, so I’ll make my own. What I’m looking for is this:

• Height adjustment – Splitters make downforce via suction, which is highly dependent on height. For street driving, I want more ground clearance, but for track driving, I want it lower. I also want to be able to tune the amount of front downforce by raising and lowering the splitter. It’s just like having a wing that’s adjustable for angle.
• Attached flow – Air separates on sharp edges, so the leading edge needs to be as large and rounded as possible. Trailing edges need to break cleanly.
• Suction – A flat splitter doesn’t create a lot of suction. I want to accelerate the air under the splitter by expanding air behind it. This can be accomplished using splitter diffusers, but because the Veloster is so high off the ground, I’ll curve the entire undertray upwards.

Splitters gain downforce linearly as they get closer to the ground, and at around 3” the suction goes up exponentially. At about 1.5” from the ground, downforce decreases sharply. So ideally you want your splitter in that 1.5” to 3” window. But this is a street car on stock (soft) suspension, and so I can’t go that low.

Ground clearance on the VN is about 6” on the front fascia, and I can reduce that by a couple inches, but any more than that will give me problems when street driving. I used an aluminum z-plate to attach the middle of the splitter and lower it slightly.

A splitter needs to be stiff enough not to deflect at speed. You don’t need to be able to stand on the lip like I see people doing, because most of the downforce is made behind the lip. I used 6mm (1/4″) Meranti plywood for most of the undertray, I added a layer of 9mm on the front. A much easier solution is to simply laminate two pieces of 1/4” together as a continuous curve.

Rather than using splitter diffusers, I’ve curved the entire rear of the splitter upwards at 12 degrees. This will expand air at the rear, which accelerates the air in front, maximizing suction across the entire splitter blade.

Note that for wind tunnel testing, I tried a flat splitter, and curved splitters with 5-degree and 12-degree curves. Curvature wins, big time.

The problem is, what to do with the diffused air? A proper race car would dump the air into the wheel wells and use vents behind the tires and on top of the fenders to extract the air. I don’t (and won’t) have either of these vents, and consequently the splitter can’t make as much downforce. That’s the consequence of having a dual-duty car.

You might be wondering what happens to the air that goes under the car, since it isn’t being extracted properly. Me too. I’m guessing that the air just diffuses under the car creating more drag. (Although this splitter actually reduced drag.) But I will always trade front downforce for drag, so I’m not very concerned by this. I also don’t have a flat belly pan and rear diffuser, so I’m not concerned with keeping air attached underneath.

As a tire rolls forward, it compresses the air underneath, shooting a jet of air sideways on either side. This phenomenon is called tire squirt, and it can have downstream effects on underbody aero. The faster you go, the more pressure on the face of the tire, the faster the tire rolls, the more compression, and hence the more tire squirt.

To reduce tire squirt I’ve added strakes in front of the tires. These spin a vortex and hit the tires at and angle, sending lower energy air sideways across the face of the tire. This reduces tire squirt and aids in extraction beneath the splitter blade.

To make the strakes I cut an arc into the splitter and slid rectangular pieces of thin plywood in the slots. I then trimmed the top side (the side facing the engine), and then fiberglassed over the exposed bottom strakes.

To support the front of the splitter, and provide height adjustment, most people use splitter rods. I made my own from elevator bolts, clevis pins, and turnbuckles. The first hurdle was where to mount the splitter rods to the frame or bumper. Splitter rods that mount to the bumper plastic are poseur bullshit.

I found out that a 4” square U-bolt is a near perfect fit. You have to bend them out about 1/8”, and then they slip right around the front bumper. At first I bought 5/16 U-bolts, but then switched to 3/8”, it’s way stronger.

I placed the splitter rod supports 19.5” apart, so that they line up with a couple holes in the front grill. You’ll have to cut the webbing out to enlarge the hole. There are coupler nuts on the ends of the bottom of the U-bolt, which is where the splitter rods attach.

I originally made two splitter rods, and then decided to make three. I don’t know that three is really necessary, but I tend to overbuild things.

The splitter rods attach to the splitter using elevator bolts. Mark a hole with a small drill bit, recess the bottom facing using a Forstner bit, and then drill a hole large enough to install the elevator bolt from the bottom. This makes a nice flush fit, with lots of support from underneath.

The elevator bolt attaches to a clevis on the top side. Make sure to match the thread pitch on both, as the elevator bolts are typically a coarse thread and the clevis joint is usually a fine pitch. You can find the coarse thread clevis from McMaster Carr and everything else from Lowe’s or whatever local hardware store you use (but avoid Home Depot for their shady business practices). Fender washers between the clevis and elevator bolt is a good finishing touch.

The splitter is adjustable for height via turnbuckles and in the lowest position sits about 2-3” lower than the front lip, and so I need to cover the gap with something. Most people use race plastic or garden edging. It doesn’t really matter what you use as long as it blocks the air from going between the top of the blade the bodywork.

The splitter is 69” wide, and for aesthetics, I made the airdam go around the bumper fascia on the sides of the car. There’s no performance gained from that, and a 67” splitter with an airdam only on the front section would have been a lot less work. I made 5-degree and flat splitters 67” wide and they worked great.

This is a street car, and so I didn’t do a lot of the downforce-producing tricks I would have done on a track car. For example, other than the strakes under the splitter, I’m not doing anything to kick air sideways, or extract air from the wheel wells. There are no vents behind the fenders or on top of them.

If I was building another dual-duty splitter, I wouldn’t use the z-plate, it restricts the vertical height adjustment. Instead I’d bolt a piano hinge to the undertray. This would allow a full range of motion from stock height to resting on the floor. I’d make the splitter slightly thicker so I could use two splitter rods instead of three, and then it would be easier to adjust the height.

For a track car, I’d make a different splitter, wider, with some treatment on the ends to kick air up and sideways. And I’d make it longer, maybe 5-6”. It would be adjustable for height and I guess it could be used on the street, but would also need a lot more rear downforce to balance out.

But even as is, this is a good splitter. I tested this splitter in the A2 wind tunnel and it reduced drag while making good downforce. At 100 mph it made 165 lbs of downforce and reduced power consumption by 3.1 hp. This is good, but can be improved upon.

I’ve written in the past about aero balance, and how I like to add more rear downforce than front, so that the car can be tuned to oversteer in slow corners and push in the fastest corner, and be neutral for most other corners. But on a FWD car, that doesn’t actually mean you add twice as much rear downforce, you have to take into account the weight of the front and rear of the car. Most FWD cars have roughly 2/3 of the weight on the front tires, so adding say 220 lbs of front downforce adds 10% more front grip, but adding 220 lbs rear downforce would be 20% more rear grip.

I still need to tune the aero balance on the car, so I can’t say exactly how much rear aero is required to balance the front. But this is why wings are adjustable for angle, and why splitters should be adjustable for height and angle as well. And yet none of the manufactured splitters are adjustable for either height or angle, and why you should DIY your own.

I recently got an older AiM Smarty Cam, and it came with an AiM Solo DL and GPS-08 external antenna. There are a lot of AiM products and it’s confusing to know how to put them together, so I contacted Peter Krause who patiently walked me through what I’d need.

It’s really easy for the Veloster, or really any other car with OBD2. The 7-pin port on the Solo DL connects to the car’s OBD2 port. A 5-pin male to male cable goes between the Solo DL and the Smarty Cam’s EXT port. That’s it. Or rather, if I had any of the cables, then that would be it.

The camera’s 4-pin GPS port isn’t used in this configuration, and so I’ll be getting rid of the GPS-08 antenna, which will more than pay for the cables. These are widely available; I paid about $50 each.

In this configuration, the Smarty Cam gets power and GPS info for speed and lap timing, as well as the trigger to turn on and recharge from the Solo DL. The Solo DL in turn gets its power and extra channel information from the car’s OBD2. So essentially the car powers both devices, and you don’t need any other cables. Sweet.

I’ll also need to load the latest track maps into both devices. This means downloading the latest version of Race Studio 2 and uploading the track maps into the Solo DL. And then downloading the latest version of Race Studio 3 and uploading the track maps to the Smarty Cam. Weird, but OK.

With all that done, I should be able to get quite a lot of extra information, in addition to the GPS data I’m use to.

I like my AiM Solo mounted on the windshield, as far left and low as I can put it. Most road courses go clockwise, and so the Solo is out of view for the majority of corners. I’m addicted to the delta timer, and this is a way to impose limits on screen time.

Running the wires was easy and oddly fulfilling. I pulled the rubber moulding away from the door jamb and tucked the wires into the slot below. The wires need to go in and out of this slot, so I used extant gaps in the interior panels for that. Going aft, I ran the wires back towards the seat belt anchor and popped them out under the driver’s seat and under both seat rails. The wires then go up to the camera.

I had a lot of ideas on how to mount the camera, and this is the one time Facebook read my mind and sent me the correct ad. There’s a product made specifically for cars called a Lapole, and it’s compatible with a lot of cameras, but doesn’t mention the Smarty Cam. I could probably adapt it to work, but at $100, I shouldn’t have to.

So I bought a heavy-duty adjustable spring-tension curtain rod for $20 and added a cam-lock mount with a 1/4-20 thread. This was one of those items I knew I had somewhere collecting dust, and it was going to take me 1 minute or 1 hour to find it. Thankfully just 1 minute.

There’s a lot of tension on the spring-loaded pole, so I cut a piece of plywood on top to use as a clamping caul, so that I don’t ruin the headliner. I was less concerned with the floor and simply ran the pole down to the carpet.

The camera mount seems quite stable, and I’m chuffed it all came out so nicely. I’ll make some tethers next and find a better way to bring the wires up the pole, but good enough for a quick project. I haven’t taken any video yet, but I’ll update this post when I get it on track.

On Monday I’m going to the A2 Wind Tunnel in North Carolina. This will be my first trip to a wind tunnel, and I hope to get many questions answered. Like most things in life, I’m certain I’ll come home with twice as many questions as answers, and I’ll be making regular visits thereafter.

The tunnel is a 10-hour drive, and expensive, so I want to be well prepared for the test, with lots of parts that come on and off quickly. But for sure I’ll need a second set of hands as well.

I asked around online to see if anyone could come help me, and while I got a few tentative offers, nothing solid. Then I messaged AJ Hartman for advice on how to not fuck up, and he said he will meet me there and help with parts swapping and data analysis! No. Fucking. Way. I had to dick punch myself to see if I was dreaming. Ouch; this is for realz.

There are so many things I’d like to test, but at $600 per hour, I’m already looking at a $2400 bill! Every minute I stand around or waste time is $10, so I have to be efficient about collecting data and do all the analysis at home.

So what am I testing?

Three splitters

Because touring cars are front limited in downforce, front aero is the most important thing you can do for performance. The more downforce you can get in front, the more you can add in the rear. It’s easy to add more rear downforce using greater wing angle, a bigger wing, a Gurney flap, etc. But getting more front downforce requires a lot of work.

Splitters are the easiest way to make front downforce, and many people add a splitter beneath the factory lip, or build an air dam. Racing rules often limit the length of the splitter lip, as well as how far rearwards it can go (front axle being normal).

But how effective is a splitter over the OEM undertray? That’s one of the things that I wanted to test out, and so I built a garden-variety splitter. Like 90% of the splitters I’ve seen, it’s dead flat, with sharp edges and no optimization. Kinda like your splitter.

Some racing rules allow flat splitters with some variance to the angle. This is smart, because adjusting the splitter height usually results in some small change in the angle of attack, and so “flat” would no longer apply. But what if you started with a warped piece of wood and attached that level to the ground (or slightly upward)?

GLTC rules allow splitters +/- 5 degrees. SCCC Time Trials allows +3 degrees for and aft (which I’m taking to mean 6 degrees of curve, wink-wink). A bit of upwards slope in the rear should help diffuse some air, but how much will this improve performance. And what about rounding the front edge, chamfering the rear edge, how much does that help? Well I’m about to find out, because I built a splitter like that as well.

And then there’s my 3D splitter, it has 12 degrees of sweep upwards in the rear, strakes underneath, and other minor tricks.

I’m not testing splitter length. I just have too many things to do and didn’t have time to make an adjustable length splitter. The splitters do not extend width-wise past the front fascia, and so I can’t test things like tire blockers, air curtains, end plates, and that kind of thing, either. I also don’t have fender vents or any proper sideways extraction from under the car, so none of the splitters are optimized for extraction.

However, all three of these splitters are the exact same dimensions, and so I’ll be able to test exactly what happens with flat, a mild curve, and an optimal curve.

Five wings and two spoilers

I just got through saying how much more important front aero is, and I’m testing twice as many rear aero options. I like wings, what can I say?

All the wings will be tested with and without a 1/2” Gurney flap. I also made sure all of them can use the same end plates, a rectangular street sign, as is my wont. I’ve brought some other end plates to test as well.

S1223 (586)

My Selig S1223 wing measures 53.3 x 11, or 586 square inches. This is probably the last wing I build based on an aviation airfoil. After finding Enrico Benzing’s wings and what the FSAE kids are doing with those and the MSHD airfoils, all my future wings will probably be car-specific designs.

I built this wing thinking it would be the primary wing on my Veloster. But data may prove otherwise.

MSHD (495)

The MSHD airfoil looks a lot like a S1223 withe more camber. I built this wing using foam and fiberglass, it measures 63.5 x 7.8, or about 495 square inches. The 3D design is specifically for a Miata roofline, so there’s a center section that’s 6 degrees offset from the sides. That’s not ideal for the Veloster, but since this wing is wider than the others, it will benefit from twisting the ends down, even if the angle is isn’t exact for the Veloster.

I built this wing thinking it could be a sellable product, custom built to any roofline. But I have as much desire to do product support as I do to wash the dishes, and so like everything else I build, it’s for me or given away.

9 Lives Racing Big Wang (506)

9 Lives Racing has a great wing, it’s essentially Enrico Benzing’s Be 123-125 airfoil with a slight modification (thicker trailing edge to mount a Gurney flap). I expect it to be quite good, and it won’t surprise me if it’s best in the test. The one I’m bringing is 55”, and just over 500 square inches in area.

I have lots of end plates for this wing, so I might test the vanilla rectangular plate vs the 9LR CFD vs one I designed. I don’t expect to see much difference, but I don’t mind being proven wrong.

Procar Innovations (484)

Procar Innovations has an aluminum wing that looks similar to a TCR wing, which is another Benzing wing, the Be 183-176. PCI shortened the chord and made the wing less aggressive, and in Benzing coordinates this is around 183-155 (18% thickness, 15 degrees of camber). It measures 55 x 8.8, or 484 square inches.

Thanks to Kevin Zhu and Jakob Boedenauer, I’ll be able to pick this one up down the street from the wind tunnel. They use this wing in GLTC, and so it’ll be really interesting to see how three similar sized wings stack up in the 500 square inch class.

MIC 249 vs spoiler

This is a made in China wing that measures 53.3 x 4.6, or just under 250 square inches. If you race GLTC and want a wing in the “free” category, this is it. Justin Lee and I tested one vs a 9LR wing, and the small wing lost. But the story wasn’t over, because I didn’t test this size wing vs a spoiler of the same dimensions.

So I also made a 250 square inch spoiler, and I’ll finally be able to get the data on which is better in this size. It’ll also be interesting to see if a wing with half the area makes half the downforce.

OEM spoiler vs base

Another test I’ll run is the OEM spoiler vs the roof extension on the base model and Turbo R Spec. The OEM spoiler reportedly cancels lift, and measures out at about zero degrees across it. The R Spec roof kicks up a few degrees compared to the roofline but is still pointing downwards at around 5 degrees.

Diffuser

One of the things I’ve always wondered about is a diffuser without a flat bottom. I’ve seen videos, CFD from numerous manufactures, and countless track cars with a diffuser and no flat bottom.

So I made a diffuser just for this test. It installs into the hitch mount and is only wind-tunnel spec, meaning not really roadworthy. It’s as large as most rules allow, and has a few tricks that I’m not going to reveal just yet. If it works, maybe I’ll build a real one.

Hood vents

Veloster’s run hot, and so I got a spare hood with an extractor vent. I’m fairly certain this will help cooling, and I enlarged the vent further to increase the effect. I also built a cover for it so that I can A/B test not only the cooling properties, but the aerodynamic properties.

I expect the hood vents will add a small amount of downforce and drag. After I test the temperature on track, I’ll do a full write up.

Canards

At the last minute (the morning after posting this article), I decided I needed to answer a question about canards. So I stumbled down to the barn at 4:30am and took a tracing I’d made and transferred that to an aluminum street sign.

I then cut tabs into the mounting side of the canard and alternately bent one up and one down. These would serve as temporary brackets so that I can tape them into place. This is wind-tunnel spec for sure, and if these work, I’ll need to make better ones that mount properly.

The canard design is pretty clever, in my opinion. On the top canard, I use a flat spot by the headlight to gain more surface area, increasing the high pressure side. On the bottom canard, I follow the lines of the air curtain vent, directing more air into the vent. I went for a fairly low angle of attack, hoping to avoid massive flow separations on the back side. We shall see….

Open windows

I’ve always wondered how open windows effects a car with and without aero, and so I’ll test the car with both open and closed windows with the full aero kit, and then again with no aero. I also brought rain guards to see if reducing the open window area will help.

Packing

To transport the parts I bought a 60” aluminum cargo carrier and a waterproof cargo bag, but I needn’t have. As I started loading all the large parts inside the hatch, I realized I could probably get everything inside. The people at the wind tunnel are going to laugh as I roll up like the Beverly Hillbillies, trunk packed to overflowing. Just one more reason to love the VN, so much junk in the trunk.

Most people enjoy the feeling of oversteer, it’s usually accompanied by a loud whoop and a “yeah,” “woooo,” “ha ha ha ha,” etc. It doesn’t matter if you’re on four wheels, two wheels, or no wheels, there is just something innately human that enjoys sliding sideways and yelling with glee.

Conversely, nobody likes the feeling of understeer, it often immediately evokes inward frustration rather than outward joy. Understeer is usually accompanied with words like “turn already,” “goddamnit,” “stop pushing,” “pig,” “fuck,” or more eloquently, “goddamnit, stop pushing and turn already you fucking pig!”

And yet, understeer is usually faster. Why?

• Traction sensing is easier
• You can brake later and accelerate earlier
• Mistakes require fewer corrections
• People suck at driving

Traction sensing

One reason understeer is faster for most drivers is that it’s easier to get to the limit of traction by pushing the front tires. Understeer is safe, and going over the limit doesn’t cause any drama. Conversely, exceeding the limits of rear traction can result in a spin.

You might even say that the reason most people find it easier to sense front traction is that most people are afraid of spinning. As much as I don’t want to admit it, I also fall into that camp.

My personal style of driving is to grip the steering wheel hard (probably too hard) and sense the feedback through the wheel. I don’t saw at the wheel, though, I’m very deliberate and economical with steering.

My style of driving requires a more geometric racing line, and returns a higher min speed. And while this may not be the fastest way to navigate a corner, I do so at the limit of traction. As Ross Bentley says, a person driving off the ideal line but on the limit of traction will be faster than a person driving the ideal line and driving under the limit.

Brakes and acceleration

A car that is unstable is easier to turn. Understeer is inherently more stable, which is exactly why it’s harder to turn. But this stability pays off in braking and acceleration.

An oversteering car can be difficult to control when threshold braking, and especially when trailbraking. It takes skill to keep the back from coming around. An understeering car is more stable, and allows later threshold braking and more aggressive trailbraking.

The same is true on acceleration; oversteer requires deft throttle modulation to keep from spinning, while understeer allows you to mash the gas as soon as the car is pointed correctly. In my experience, I get on the gas earlier with an oversteering car, but I get to full throttle earlier with an understeering car.

Mistake recovery

A good driver makes a lot of small mistakes and corrects them instantly. They are constantly losing and catching grip with skill and grace. A car that understeers is very easy to correct; if the car starts to lose traction, just wait until the car turns before adding more throttle. An oversteering car takes more skill to correct; sawing at the wheel and dancing on the pedals repeatedly until the car gets back into line.

An important factor here is the stability index of the car, which in layman’s terms is exactly what it sounds like, but in reality is so much math and calculations that it will make your head spin. Let’s not do math here and talk generalizations.

Mid-engine and rear-engine cars with a lot of weight on the rear tend to be very pointy, swapping ends if you are careless and lift the throttle. A car like this without modern traction control systems suffers no amateurs in the cockpit; you need skill to drive one quickly. Conversely, front engine cars tend to be much more forgiving. If you look around, how many people drift with mid- or rear-engine cars? Not many, because it’s much easier to control a car that doesn’t snap oversteer.

This isn’t just about mechanical stability and grip, but aerodynamic forces as well. Cars that have a lot of rear drag or lateral surface area (like a LeMans prototype shark fin) move the center of pressure rearwards. This in turn makes the car more stable at speed. The distance between the center of gravity and the center of pressure is called static margin, and cars with a greater static margin are more stable, and require fewer corrections to mistakes. They are also harder to turn, and so this is again an understeer vs oversteer situation, where more rear aero is easier to drive and typically faster.

People suck at driving

You can add all these factors together and conclude that understeer is faster simply because most people suck at driving. Driving a pointy car at the limit demands a lot of skill. Driving an understeering car is child’s play. That’s why virtually every car off the showroom floor comes with understeer from the factory.

Manufacturers know that people suck at driving. You’d go broke selling cars to people that requires the skill to regularly catch a spin. I’m an enthusiast and decent driver, but I’m no Max Verstappen, and I never will be (not in skill, nor in attitude; I would have sacrificed one win to let Perez take second in the Drivers Championship last year). I didn’t cut my teeth racing karts, or spend every waking moment being a douchebag racing cars.

Otherworldly drivers like Max can do things I’ll never be able to do, and driving an oversteering car faster than a understeering car is just not in my wheelhouse. I’m quite happy with my conservative, front-biased driving style. By using that style, and driving at the limit of front traction, I regularly beat the pants off people overdriving their cars under the limit.

Oversteer mythology

If understeer is faster for non-pro drivers, why do so many non-pro drivers believe that oversteer is faster?

• If oversteer was faster, the 1000 hp drift missile at your track day would set FTD. But instead it’s a Miata with too much wing on scrub SM 7.5 tires.
• If oversteer was faster, F1 cars would rotate through the majority of corners. But they don’t. You see more F1 cars push the front than the rear.
• If oversteer was faster, people would put wider tires on the front than the rear. But they don’t. High powered cars run a staggered setup, or maybe square – but you won’t see a RWD car in a gorilla stance.
• If oversteer was faster, you’d see cars with more front aero than rear. But most race cars are split about 1/3 front and 2/3 rear downforce.
• If oversteer was faster, Spec Miatas would have more oversteer than a stock Miata. But SMs have a 63% front roll couple, compared to a stock Miata’s 59%.
• If oversteer was faster… well, you get the idea. It ain’t.

If you think oversteer is faster than understeer, I want to see your data.

Driver 61

Driver 61 is a YouTube channel with great content. In one of his latest videos, he delves into understeer vs oversteer, and why Alonso loves understeer. He goes on to examine various drivers and what they prefer, and shows this list.

• Oversteer drivers: Max Verstappen, Sebastian Vettel, Charles Leclerc, Michael Schumacher, Lando Norris, Ayrton Senna.
• Understeer drivers: Lewis Hamilton, Fernando Alonso, Sergio Perez, Jenson Button, Nigel Mansel, Alain Prost.

It’s an illuminating video, and worth the watch. What’s interesting is that Driver 61 himself was originally an oversteer guy, and more recently has transition to an understeer guy. My twin brother Ian is the same, originally very rear happy, but transitioning more to understeer sadness.

Ian vs Mario

I wrote an article called Twin Studies, which is an in-depth analysis of my identical twin brother and I driving the same cars very differently. In that article we drive a NA and NB Miata, and a Veloster N, back to back.

Ian has an oversteer driving style, a la Verstappen and Schumacher. He holds the wheel lightly, thumbs on top, and pivots the car on entry. I have a more understeer style, like Alonso and Hamilton. I grip the wheel tighter, and sense traction through the front wheels.

Both styles are fast, but I know the track better, and so I was faster on this day. But not by much, and he might have eventually beat me. But he did transition more and more to my style of driving throughout the day ….

Gran Turismo Test

In this article I’ve thrown a lot of theory around, so it’s time to get the gloves on and actually do some (virtual) driving. You should do this as well, it’s very illuminating data.

In Gran Turismo (or any other game where you can easily change front and rear grip), try out this experiment: get a Miata or similar RWD car with near 50/50 weight balance and mismatch the tires.

I used a bone stock ND Miata and put Sport Hard (SH) tires on the front and Sport Soft (SS) tires on the rear. Set up like this, the car understeered quite a bit. I ran ten laps at Tsukuba, then swapped the tires so that the car would now oversteer and did another 10 laps. In both cases, the car has the same amount of grip.

I started with the understeer setup and then switched to the oversteer. This gave the oversteer setup a bit of an advantage, because it’s been a while since I raced GT7 and I would have to hit my marks right away on the understeering car. I did OK, with a best lap of 1:09.695. What’s surprising is that my outlap, which isn’t shown in the lap summary, was a 1:10.3. Meaning that the understeering car was easy to get up to speed on immediately.

Next I swapped the tires and tried the oversteer setup. Whoah! Big difference. I almost spun twice on my outlap and barely managed a 1:13. Unlike the understeer setup, I needed to learn how to drive this. Eventually I figured it out and got down to a best lap of 1:11.028. It felt like quite an accomplishment to string together that many corners and catch the slides each time.

Then I put Sports Hard (SH) tires front and rear, just to see what would happen. This is using the worst tire on both ends, and isn’t really part of the oversteer-understeer experiment. But you can see that a balanced car is the fastest. Even with tires that have the least grip I did a 1:09.347.

But to get back to the two ends of the spectrum, the understeer setup went 1.3 seconds faster than the oversteer setup, and was 1.6 seconds faster on average. In both cases, the car had the same amount of grip, and so you can see that understeer is a lot faster. At least for this driver.

I also included the optimal lap time for each, and you can see that the understeering car’s best lap was about 1/10th of a second off the optimal lap. On the other hand, the oversteering car’s optimal lap was over half a second off. This shows how sloppy my best lap was, and how difficult it is to put together a clean lap in an oversteering car.

Driving impressions

The way to drive an understeering car fast is to make it oversteer on corner entry. In the tight hairpins, turns 1, 4, and 8, I found myself diving towards the apex on the brakes, releasing slowly, and sharpening my steering through the whole corner. I didn’t always brake traction on the rear and rotate the car, but I kept the weight on the nose for as long as I could. I did something similar in the 90-degree T5, but using lift-throttle oversteer to help turn the car, then got right back on the gas again.

Conversely, through T9 I used understeer to get to full throttle early. I ran it in tight until the front set, and then stood on the throttle to the exit, purposely pushing the front tires the whole way.

I found that to drive the oversteering car quickly, I had to adopt a more “momentum” style of driving: lighter and longer braking, optimize mid-corner speed, and get on the gas early using maintenance throttle. Unlike the understeering car, which I could mash the accelerator as soon as the car was turned, on the oversteering car, feathering the gas was important.

With less grip on the rear, I needed to shift weight to balance traction, and so T5 was on throttle (lightly) the whole time. I also tried holding a taller gear, which helped keep the rear from stepping out, but I eventually figured out how to gradually get on the gas in the correct gear. I did a lot less trailbraking because the back would come around, so I rookied my way into straight-line braking and releasing the brakes quicker.

The other major difference in driving the two different tire setups was the number of steering corrections required. In the understeering car, if I went in too hot or misjudged my turn in, it was a simple matter of waiting for the car to turn before adding throttle. There was very little drama, if maybe a little cursing (“turn you pig,” kinda stuff).

On the other hand, the oversteering car was a fucking handful! I made more steering corrections in two corners than I did over an entire lap in the understeering car. It was a lot of work, but it was also more fun.

Finally, driving the balanced car was the most fun of all, and it was the fastest as well. I used some strategies from both styles of driving, and was finally able to drift all four wheels. For example, in T1 I trailbraked heavily, and in T5 I used LTO to turn the car, both understeering tactics. But I also found myself getting on the gas earlier and feathering the throttle out of most turns. The final turn was completely different: a four-wheel drift through the entire corner. You can really only do that with balanced traction.

Wet track

Next I made it rain in Japan and tried the same builds on a wet Tsukuba. This time I only did 5 laps though. No more shitty screen shots of my TV, just the data this time:

The delta between understeer and oversteer grew a little in the rain. I put down one good lap on the oversteer setup, but the the optimal lap was two seconds faster on the understeer setup, and so it had way more potential in it.

What’s interesting is that the SH/SH combo, which was the fastest in the dry, became the slowest in the wet! So when the track is wet, you’re better off with more traction at either end than you are with balanced traction and less grip.

Conclusion

If your car has criminal understeer, you want to fix that for sure, because you’ll go faster on a dry track with a balanced setup. To do that, check out Emil’s video on the subject:

But don’t make the mistake of going too far in the opposite direction and making a car that oversteers, because that’s even worse than understeer. The ideal setup is balanced traction, and with that, you’ll go fastest on a dry track, even with tires that have less grip. But if the track is wet, or there’s mixed conditions, you’d be smart to dial in some understeer.

Real-world testing?

My brother and I have been talking about doing this Gran Turismo tire test in the real world at Pineview. It’s easy to get an empty track session, and there are so many slow corners that we’ll spend the whole time breaking the tires loose. I just need to bring two sets of different tires and make this happen.

Unfortunately both of my Miatas are out of commission right now, but if I get one up and running this summer, I’ll fill this space with the Gran Turismo exercise done IRL. I might also be able to borrow a Miata and try this, and I’m planning to meet Emil Dogaru (from the video above) at Summit Point later this year and maybe do this test with him.

Subscribe to get updates or watch this space.

If I was to grade the average splitter I see on the average track car, I’d give it a D+. I know that grades are supposed to average out to a C, but the average person sucks at aero. I see better splitters on Miatas than other cars, and those generally fall in the C+ range.

• Grade F: Flimsy splitters or front lips, usually bought on Amazon or eBay, but some reputable aero companies also produce shit. These look like splitters, but that’s where the similarity ends.
• Grade D: A misunderstanding of how splitters work results in splitters that are optimized for the top surface, not the bottom. They are are often too high or too low, and universally have sharp leading edges and are flat. Most have splitters wider than the front fascia, but do nothing with the exposed ends. Some might have a spill board on the outer edge, which is at least something, but that does nothing for the underside.
• Grade C: A splitter gains downforce through suction, which is achieved primarily through proximity to the ground and using diffusers and/or curvature to accelerate the air underneath. A splitter without height adjustment is like a wing without angle adjustment. A splitter without any diffusion is as useless as a flat wing.
• Grade B: With great suction comes great responsibility. Because the air under the splitter is low pressure, everything around it is higher pressure and wants to get in there. You need to extract air using hood vents and fender vents, and kick air upwards and outwards wherever possible.
• Grade A: At this point the splitter is making a lot of downforce, but there are still minor problems that occur around the wheels. Air hitting the rotating tires compresses and squirts out under the splitter, and air intrudes through the wheel spokes as well. Mitigating losses around the wheels and tires is the last frontier.

Grade A

Let me show you what a solid Grade-A splitter looks like. It’s not surprising that it’s on a Miata, because we do aero better. We have to; our cars suck. It’s also not surprising my teammate Alyssa Merrill made this. We bounce ideas back and forth, and while I think I’m pretty clever, she’s got a lot of great ideas of her own.

First, you have to use the right material. Alumalite is an automatic grade C, not because of the material properties, but every single person using Alumalite leaves a sharp leading edge. Despite the fact that Alumalite curves nicely, pretty much all of them are flat, as well.

Alyssa chose to use my favorite splitter material, Meranti BS6566 plywood. It’s very light (32 lbs/cu-ft), stiff, easy to work with, boil tested for delamination, and x-rayed for voids. It’s expensive, but you get what you pay for. She’s doubled the thickness of the lip to about an inch and done a good job of rounding the underside of the leading edge, so that air stays attached along the bottom surface.

The location of the radiator opening makes a difference to how the pressure side of the splitter functions, and has an effect on drag and cooling. I’ve seen a lot of radiator openings just barely off the splitter. When you put a hole there (for radiator or brake ducts), you lose the high pressure air bubble on top of the splitter, and with it, some of the downforce you’re trying to make. You might be thinking that a lower opening is better for cooling, but with the radiator opening down low, the air has to take a hard upward turn inside your radiator duct. Air can’t change directions at more than 12 degrees, and because the duct is much steeper than that, air separates from the surface of the duct causing drag and losing cooling efficiency.

Alyssa made the radiator opening as high as is practical while retaining the center bumper support. The higher opening holds a larger/taller high-pressure air bubble on top of the splitter, making more downforce. In addition, this provides a straighter route to the radiator, increasing cooling efficiency and decreasing drag.

All the air that goes through the radiator has to go someplace, and the worst place for that is under the car. The best place to send that air is upwards, because it creates downforce. Some people will use louvered hood vents, while others will use just a front wicker and a gaping hole behind. Louvers smooth the air over the hood, and result in less downstream losses on the wing. But since most cars are front-limited with respect to downforce, a gaping hole is a better solution for a track-only car. For a street car, louvers look better for sure.

Having a splitter without any height adjustment is akin to a rear wing without any angle adjustment. You see that kind of junk on OEM body kits, and you can do better. With a proper wing, tuning the high-speed handling of your car is as simple as changing wing angle.

It’s like that with splitters as well. Most of the downforce comes via interaction with the ground, so setting the height is critical. I’ve seen a lot of splitters that adjust for height initially, but then once you mount the airdam or fascia, the height is fixed. This is a C grade. At least there was some height adjustment at first, and that’s an acceptable situation if you can later modify it.

People getting a B grade with a fixed height splitter do so because they understand aero maps and have set their splitter height for maximum downforce, and thereafter do all of their adjustments via wing angle. But having both ends adjustable for downforce is obviously even better.

A-grade splitters are adjustable after the fact. Alyssa used turnbuckles to support the splitter, which are sturdy and make height adjustment a simple twist of the wrist. The turnbuckles are behind the airdam, which makes height adjustment more difficult, but it also completely hides them, for less drag. This is some trick shit.

What’s even more trick is that the turnbuckles fasten with pins, and so removing the splitter entirely is super easy. Just yank four pins and the splitter drops off the front.

Another area to address is the width of the splitter, and what to do with the ends. Under most racing/TT rules, splitters can be the width of the car or the front tires. In most cases, this results in some unused material on the outside edge of the splitter.

You get a D for doing nothing with the exposed end. A flat splitter with a flat exposed end is just additional drag and an invitation to catch on something. You get a C for putting a spill board on the outside, because while this can reduce drag, there are so many better uses of that real estate. Higher grades are achieved by putting a vertical dam on the outside edge (front tire spat if you will), which builds local pressure on top of the blade, creating more downforce. Also, air spilling off that will direct air sideways, extracting air beneath the splitter, making more downforce.

Alyssa chose to do something similar, but pulled the bottom edge of the airdam outwards to meet the end of the splitter. This won’t make quite as much local pressure on top, but should be better at sideways extraction and have less drag. Hard to say which is better, both work.

Let me address another huge mistake: flat splitters. Most people understand wings, and and flat splitters are like flat wings; meaningless. You need some curvature in the rear of the splitter to diffuse the air. When air expands behind a restriction, it accelerates the air in front of it, lowering the pressure. This suction is the purpose of a splitter.

Time attack cars often curve the entire rear of splitter blade upwards, essentially making a splitter diffuser the full width of the car. Miatas have historically used the rear subframe as mounting points for the back of the splitter, and the consequence of that is that Miata splitters are flat. To get suction, Miatas need to use splitter diffusers.

Splitter diffusers operate on the same principle as rear diffusers; they expand air in the tunnel. The diffuser itself doesn’t create the downforce, but it helps air expand and creates downforce in front of it.

Alyssa and I are designing splitter diffusers and will make them for sale sometime this summer. Here’s your first sneak peak.

These splitter diffusers have one large tunnel that diffuses air inboard of the wheels. This is an area of loss to begin with, and also an easy place to extract air from. Most splitters diffusers have only the one main tunnel.

But you’ll notice our splitter diffuser has additional area outside the main diffuser, with strakes. The purpose of this area is to mitigate tire squirt. As a tire rolls forward, it compresses the high-velocity air underneath it, squirting air out the sides. Some of this air goes under the splitter, which decreases the suction and downforce. By adding a diffuser in this area, it slows the air down, hitting the tires with less velocity.

The strakes spin a vortex off the trailing edge. Because vortexes take energy to create, the air hitting the tire has less energy, which reduces the amount of tire squirt. The strakes are also angled just-so, to achieve the maximum amount of extraction from under the splitter and send it outboard, protecting the main tunnel of the splitter diffuser.

Recall that most of the air is dumped inside the wheel well. Air here is extracted via vents on top of the fender, and behind the wheel. By extracting air upwards, you create an opposing force downwards. And by sending air sideways, you mitigate losses from air intruding underneath. Fender vents are important to make the most out of splitter diffusers.

A+ splitter

There’s an old saying I think attributed to fighter aircraft: 90% of the cost is the last 10% of performance. The same applies to car aerodynamics. While this splitter is a solid A, to make this into an A+ requires a lot of labor and cost, for little return.

First, I’d cut out the center bumper support completely, it isn’t necessary for chassis rigidity, and I can fashion a custom bumper after ducting the radiator. Removing that center support allows a slightly higher radiator intake, with the benefits I already discussed. I’d put a 100mm radius on the edge of the radiator duct, and reduce the opening size slightly.

I’d also tilt the radiator forward to reduce the frontal area, and duct the radiator exit straight upwards through an extractor vent to create more downforce. You might call this more engine bay management than splitter improvement, but it’s all related to creating more front downforce, and goes hand-in-hand with designing the splitter and ducting.

I’d also make the splitter slightly wider. Alyssa’s car is for HPDEs and is unbound by racing rules, so an extra couple inches on the end isn’t a problem. With that extra space I’d place place a simple piece of angle aluminum. This end-wicker, if you will, holds pressure on top of the blade and locates a higher stagnation point behind. This pressure differential creates more downforce, and it’ll still kick air sideways for extraction.

With a clean sheet, I’d also add curvature to the entire undertray. I’d laminate two pieces of 6mm plywood together, rather than using a single piece of 12mm. This would allow the rear edge to sweep upwards, creating more suction under the middle of the splitter blade.

Alyssa’s car doesn’t have a flat bottom, so it’s not a huge problem to diffuse some air into the center of the car. There will be some additional downstream drag due to local flow separations around the suspension bits, exhaust, transmission etc, but this is just a little bit of extra drag on things that were already creating drag to begin with.

I’d also put some height differential between the center and sides of the splitter. Remember early last year when all the F1 cars were porpoising? That was a cyclic gain and loss of downforce via ground effect, and it’s the natural consequence of trying to make as much downforce as possible, losing that downforce, and watching the car rebound on its springs. Once the car is back at a taller ride height, it again gains more downforce, repeating the cycle.

Many cars mitigate that effect by having a higher center tunnel, which gives two different heights of the splitter. This way if the lower part of the splitter goes under 1.5″ height, you don’t suddenly lose all of the downforce. The higher center portion also feeds the underside aero, which isn’t a factor on Alyssa’s car, which has neither a center tunnel nor a flat bottom.

So on her car I’d make the sides of the splitter slightly higher, rather than the middle. This will have the same effect of balancing ground effect losses, and will also protect the splitter blade, because the sides will ground out before the middle when cornering.

Finally, I might address some of the aerodynamic losses that come through the wheel spokes. By covering the bottom half of the wheel with caketin covers and spinning a vortex outboard, I’d reduce losses and redirect air down and out, with a minor gain in splitter suction.

But like I started with, modifying her splitter to go from A to A+ is a 90% effort for 10% gain situation. Addressing other parts of the car to the 90% threshold is time better spent.