86 Cup and Aero Options

If I didn’t have Miatas, I’d probably have a FRS/BRZ/GT86; similar philosophy, but in a coupe with better aero. My brother is thinking about a new car, and a 86 is on his list, so I thought I’d take this post out of the can and finish it. One of the reasons for owning this car would be to run it in the 86 Cup time trials, which seems like a cool series with a great rule set.

The 86 Cup rules has a Stock class, which allows you to take any modifications from an approved list. This is probably how I’d race one, but it doesn’t allow you to do much with aero. Since this is a website based on aero, let’s take a look at the aero options for the non-stock classes.

The rules allow you to modify the car, taking different amounts of points for each modification. The point total puts you into one of four different classes, which can result in different cars, with very individual builds. In the Street class, you’re allowed up to 4 points, in the Modified you get 7 points, and the Unlimited class is, well, unlimited.

.25Level 1 FrontAftermarket bumper, no splitter
.75Level 2 FrontSplitter, spats, canards, everything
.125Front spatsNot unless you increase track
.25Vented hoodWould help with splitter
.125FendersAftermarket or vented
.125Cut rear bumperPoor man’s diffuser
.25Level 1 DiffuserStarts behind rear axle
.5Level 2 DiffuserStarts in front of rear axle (but how far?)
0Underbody panelsAny year OEM underbody panels
.125Vortex generatorsPlease no
0OEM spoilerAny year OK
.25SpoilerAny non-OEM spoiler
1Level 1 wingUp to 55″ wide, 8″ stands
1.5Level 2 wingUp to 57.5″ wide, 10″ stands
1.75Level 3 wingAnything goes, I guess
Aero options in 86 Cup

That’s a lot of options to choose from. How do I think they stack up?

Front end

The aftermarket has turned out some interesting BRZ/FRS bumpers, but without an undertray and splitter (and hood vents if you’re doing those), I’m not sure any of them are better than stock. I guess if you have .25 points lying around, then why not.

A better choice is a splitter, it costs .75 points, but then you can use any bumper you want, wheel spats, and canards as well! Data from my testing showed that a splitter is very worthwhile, and so I’d take this option with side spats and maybe canards.


If you study the YouTube videos from Kyle.Engineers, you’ll see that a diffuser doesn’t do that much unless you’re optimizing the underbody. That means first add a splitter, then side skirts, and then a flat bottom (or Venturi), and after all that, add a diffuser. Indeed, a flat bottom alone is superior to a diffuser, so I wouldn’t bother with a diffuser unless I’d done everything forward of that first, and still had points. Cutting the bumper also helps, but it’s such a negligible gain that I’d only do that if I had points to spare.

Spoilers and wings

The spoiler and wing options are interesting.

  • A spoiler costs zero points for OEM and .25 points for anything else. I’ve got a pretty wild imagination when it comes to spoilers, and I’d probably make my own adjustable spoiler if I was in the Street class. With more points to spend, or unlimited, I’d jump right up to Level 3 wing
  • There are two Level 1 wings that are whitelisted (SARD LSR and BRZ tS), and I’ve never seen the data on them. These might be a bargain, or a waste of points. From what I’ve seen, they need bigger end plates for sure.
  • The level 2 wing can be 2.5” wider and mounted 2″ higher than Level 1. I don’t know how much that would change the performance, but .5 points seems like a lot to spend on a couple inches.
  • If you need more rear downforce, Level 3 is obviously the way to go. It’s only .25 points more than a Level 2 wing, with no restrictions on width, height, or set-back distance. Go big.


A full aero package (level 2 front, hood and fender vents, level 2 diffuser, level 3 wing) is 3.5 points. That same number of points would also get you a Super 200 tire (A052, for example) on a 8.5″ wide wheel, or a built engine with forced induction.

It’s hard to say what would be the best build. On some tracks, power, on others, mechanical grip or aero. I wrote about that once before, and while some tracks favor one or the other, it would be difficult to modify your car between races. But it does appear that the 86 Cup rules strike a good balance, and it would take me a few hundred simulations to figure out which options I’d run. But at first blush the splitter and spoiler only cost 1 point, and seem like a good combo for a Street-class build.

Verus Engineering Aero Packages

Verus Engineering makes a lot of GT86 aero parts and if you dig around on their website you can find information packets, which gives you in-depth information on their aero parts. These informative packets also group some aero parts together in different “Ventus packages” that are supposed to work well together.

In a way I want to congratulate Verus on providing the customer such valuable data, but honestly… it borders on smoke and mirrors. The only thing I need is the drag and lift values, and after you wade through 29 pages of text, images, and tables, guess what they don’t give you? Drag and lift values.

It took a couple emails back and forth with their engineering department, but I finally got what I wanted. They claimed that drag and lift was confusing to most customers, so they no longer provide that. Well it’s not confusing to this customer, it’s the only two values I give a shit about.

Anyway, rant off, here’s how those Ventus Packages stack up.

  • Ventus 1 – This package has canards (dive planes), underbody panels over the transmission and rear suspension, and a diffuser. It’s interesting that they chose to use canards, which are very inefficient, together with underbody aero, which is totally unrelated. This would be .625 points in the 86 Cup, and have Cd = 0.372, Cl = 0.013.
  • Ventus 2 – This package adds a splitter and ducktail spoiler to the first package. People often think splitters and spoilers add drag, but here’s another case where you get more downforce and less drag. This would cost 1.25 points in the 86 Cup, with Cd = 0.355, Cl = -0.195.
  • Ventus 3 – This package adds endplates on the splitter, side splitters (side skirts), and a high-efficiency rear wing. It’s interesting that they chose to use both a spoiler and a wing on this package, I guess you’d have to take both for points. It looks like a Level 2 wing, and so this package would cost 2.25 points, plus whatever side skirts cost (not listed in the 86 Cup rules). Cd = 0.441, Cl = -0.499
  • Ventus 4 – This package exchanges the spoiler and high-efficiency wing for a bigger wing. Same 2.25 points as the Ventus 3 package, with Cd = 0.465, Cl = -0.633

Personally, I would not have done the packages in this way, because you’d get more benefit by going front to back. Meaning, do the splitter first, and the diffuser only after you’ve sealed up the sides and cleaned up the underbody. But Verus is trying to sell things to street-driving customers, and the splitter is not curb-friendly, and diffusers look cooler, so starting at the back makes financial sense. Likewise, canards are the last thing I’d do, but maybe the first thing a typical customer buys? Lame, but I guess it’s not what you can build, it’s what you can sell.

Verus Simulations

So you might be wondering which of those packages is fastest? Let’s run them through OptimumLap and find out. I usually factor in more drag and lift to simulate open windows, but this time I’ll use the Verus data as is. I’ve made the tire grip 1.2g across all configurations. I’ll simulate laps at Watkins Glen, NYST, and the 2010 Solo Nationals autocross track, to see what happens at different speeds.

Note that I’m a bit skeptical of Verus’s CFD values. The downforce (negative lift) is less than I’d expect. A GT86 with a splitter and a wing should be similar to my Miata with the same, which was measured at .41 drag and -1.2 coefficient of lift (with windows open, mind you). I don’t run a diffuser, flat bottom, canards, etc., and got way better values than Verus did. I don’t know why there’s this discrepancy, but anyway, I’ve added my car’s aero as the final entry in the table.

Ventus 1.372.013.6252:21.591:42.0462.52
Ventus 2.355-.1951.252:20.631:41.4562.32
Ventus 3.441-.4992.252:21.281:41.1062.07
Ventus 4.465-.6332.252:21.341:40.9061.96
Simulated lap times based on different builds

Based on OptimumLap simulations, the Ventus packages keep going faster as you add more of their doo-dads. This holds true at autocross and medium-speed tracks like NY Safety Track. However, at Watkins Glen, notice that the bone stock configuration actually went faster than Ventus 1, and the Ventus 2 package with a spoiler instead of a wing was the overall fastest. Watkins Glen is a fast track, and if you don’t have a lot of power, drag reduction is super important. Finally, the true potential of the aero packages might be closer to my car, which flat out stomps the CFD-generated numbers. YMMV.

My builds

After reading the rules and running some simulations, this is how I’d make a car for the 86 Cup.

Street class (4 points) – Splitter (.75), vented hood (.25), spoiler (.25), 9” wheels (1.25), GT Radial SX2 245/40-17 (-1). That’s 2 points, giving me 2 points for suspension or mild engine mods.

Modified class (7 points) – Splitter (.75), vented hood (.25), level 3 wing (1.75), 9” wheels (1.25), NT01 (1.0). Thats 5 points so far, and gives me 2 points for suspension or engine stuff.

Unlimited class (any) – All the aero, supercharger, wide tires and slicks, fully adjustable suspension, etc. I mean, duh, all of it.

Stock – (0 points) – But I’d probably race in the stock class, which allows front camber plates and camber bolts, rear lower control arms, oil cooler, cat-back exhaust, base trim spoiler, and updating/backdating to any OEM parts. I’d ride on 17×8 RPF1s with 225 SX2 and call it race ready. Sell my Miatas and get into this? Tempting.

Miata Suspension Calculator

I’m pretty green when it comes to car setup, and my understanding on things like front roll couple is notsomuch. But my teammate Alyssa sent me a link to the FatCat Suspension Calculator, and after playing with the tutorial for a while, I learned a thing or two.


  • Bounce frequency – Lower values are comfortable, higher values less so. A 1990 Miata has a bounce frequency of 1.15 front, 1.01 rear. Fat Cat suggests that higher than 1.7 hz is not suitable for the street. That’s about 360 lbs front and 275 rear.
  • Roll stiffness – Determines how much body roll the car has in a corner. You can change roll stiffness with sway bars, and it won’t affect bounce frequency.
  • Front roll couple (FRC) – A low FRC makes a car easier to steer, but stable at speed, and susceptible to oversteer. A high FRC makes the car more stable at speed, but harder to steer, and tends to understeer. Mazda engineers felt that about 59% FRC was a good balance. Fat Cat says Miatas are biased a bit towards oversteer, and they use a higher FRC.
Front Roll Couple Notes
Less than 58%Oversteer
59%1990 Mazda Miata
63%Track and high speed setup
Front Roll Couple values

My Street Car

My street car has a Racing Beat hollow front sway bar, stock rear bar, and Tein Street Advance shocks with 7k/6 springs, which are the same shocks the calculator uses as an example. Luckily I did these all at the same time, but if I had done the shocks or sway bars individually, it would have ruined the handing.

ConfigurationBounce StiffnessFRCResult
Stock sways and shocks1.14, 1.0187258.9%Stock
RB front sway onlyStock129572.5%Understeer
Tein Street Advance only1.77, 1.88191050.5%Oversteer
RB front, stock rear, Tein shocks1.77, 1.88233459.6%Stock-ish
It’s easy to change FRC for better, and for worse.

Adding just the front sway bar would have made the car boring and push everywhere. Adding just the shocks, the car would have been loose mess, wanting to swap ends. But put all the parts on at the same time, and the car’s balance is very much like a stock Miata.

My Race Car

My 1994 race car is on Spec Miata suspension. When I bought it, the previous owner had disconnected the rear sway bar. That’s a front roll couple of 71.2%, which is a lot, but I didn’t know any better and anyway the car seemed to handle well like that.

But then I added a wing, which resulted in way too much understeer. This was how the car was set up during aero testing at WGI, and if you read about that, you know I also had some unfavorable negative chassis rake. All things combined, the car was an understeering pig.

Before racing at Mid Ohio last year, I reconnected the rear sway bar. This has brought the FRC back to 63.1%, which makes the car steer easier at low speed. And the wing adds progressively more rear bias, so the car transitions to understeer at high speed. The end result is a car that handles really well at all speeds. And that’s one of the cool things about aero, you tune a car so that it’s loose at low speed, and tight at high speed, just like you want it (or at least how I want it).

Getting Back 20 Points in Champcar

Champcar charges 20 points per aftermarket sway bar. I’m using a Spec Miata front sway bar. Is it worth getting 20 points back and using the stock sway bars?

Reverting the car to stock sway bars front and rear would reduce the roll stiffness by about 13%, but the FRC would be only 1.5% more. I don’t know if those changes are things I’d feel or not. However, there’s another option on stock sway bars: changing the spring rate to 850 lb front and 400 lb rear would bring the roll stiffness and FRC very close to where they are now. The only drawback is that bounce frequency would go up in this configuration, from 2.36/1.85 hz to 2.60/2.05 hz. Finally, I could leave the front spring as is, increase the rear spring to 400 lb and remove the rear sway bar, which would have the same FRC as a Spec Miata, but 9% less roll stiffness. In any case, there are a lot of options, and I can get 20 points back if I want to.

Spec Miata 750/325, 24mm/15mm sways295263.1%
As above, no rear sway bar275071.2%
Spec Miata, stock sway bars258164.6%
Stock sway bars, 850/400# springs296362.9%
No rear sway, 400# rear spring264563.1%
Sway bar options on the race car

NB Miata Rear Suspension

I was reading one of Keith Tanner’s books on Miatas, and he states that NB Miatas have slightly different suspension geometry than an NA, and consequently different motion ratios. On equal springs, an NB has softer rear suspension than an NA. So if you have an NB, multiply your rear spring rate by .87 to get the equivalent to an NA spring on the FatCat calculator.

With that knowledge I ran some numbers on my team mechanic Clayton’s 1999 NB. It has 318 lb front, 233 lb rear springs, and Flyin Miata sway bars 1″ front and 5/8″ rear. The previous owner had disconnected the rear sway bar for some reason, giving a FRC of 71.4%. That’s a lot of front bias. Connecting the rear sway bar brings FRC to 58.8%, which is almost identical to a 1990 Miata. With this suspension setup, bounce frequency is just below the 1.7 hz comfort threshold, and roll stiffness is 2.4x more than a 1990 Miata. Sweet.

DIY Modifications to a Cheap eBay Wing, Part Deux

You can buy a cheap wing on eBay, Amazon or whatever for as little as $50, or a double decker for $75. I’ve previously written about the testing I did on this wing. Let me be clear about this: it was a piece of junk right out of the box. At a minimum you must make new endplates (see previous blog post) and get the wing higher.

The mighty Mophon wing.

I made those and other modifications and it performed better than I expected. The 53″ wing created almost as much downforce as the 60″ 9LR wing! But also a lot more drag, which negatively affected front downforce. You have to think of aero as a system, and as a system, the 9LR wing with OEM hard top had a Cl of -1.01 and a Cd of .48 for an aero efficiency of 2.11. Whereas the eBay wing in the same configuration had a Cl of -0.56 and a Cd of .55, for an aero efficiency of 0.90.

While the eBay wing system was worse than the 9LR, the drag and lift values showed that my car would go faster with the wing than without it. Even my very slippery fastback without a wing was slower in a race simulation than the eBay wing with a hard top. Even cheap wings can be good!

Just the same, the eBay wing has been sitting unused. I got tired of looking at it and decided it was time for some new development efforts. Before I get to that, I’ll recap all of what I did previously, and then get onto version 2.0.

Wing stands

The janky mounts that came with the wing might work on a hatchback, but are far too low to be efficient on a Miata. You need to get the wing higher, where the air is less turbulent. I made my wing stands out of a square piece of 1/4″ aluminum plate, 12″x 12″, which I simply cut at an angle to make two matching sides.

I bolted the stands through the trunk gutter with three bolts per side. Note that the trunk gutter isn’t parallel or plumb, meaning that once the stands are bolted in, I had to pull on them and bend them to be near vertical. This bends the trunk gutter into shape.

Pro tip: If your wing stands are slightly sprung (meaning there is tension between them, and you have to force them into a vertical position), it makes the wing stands stiffer. So don’t fret about getting them vertical, just pull them into position about 40″ apart at the top when you mount the wing.

End plates

The end plates were the next things to go, not only because they were too small, but because they didn’t allow me to set the convergent gap and secondary wing angle correctly.

Slots for upper wing adjustment didn’t allow for correct gap and angle. End plates way too small.

Convergent what? Basically, the gap between the two wings must accelerate air as it passes through. In order to do that, the gap at the front edge has to be larger than the gap at the back. The factory end plates have two arcs cut into the end plates, and pivoting the upper wing through this arc is supposed to allow you to set the angle and gap distance. But if you do that correctly, then there’s no overlap.

The top wing should overlap the lower wing at about 4% of the chord, which is about 0.3″ in this case. I measured that distance and drilled a single front hole to use as a pivot, and then three holes further back to provide a range of adjustment. This gives me the proper overlap and gap, and I can set the upper wing angle easily.

Upper wing modifications

The upper wing is flexible, and if you set the proper overlap and gap to the lower wing, the upper wing will vibrate and hit the lower wing. So I added a stop in the middle of the lower wing, which will support the upper wing as it bends. The stop is just a piece of aluminum with a couple rivets to hold it in place. You could duct tape a piece of rubber tubing and it would work (did that for the first race).

I also riveted a Gurney flap to the upper wing. This stiffened up the upper wing a little, but based on the test results, I think it was too large, and part of why the wing had some much drag.

So this is where the 2.0 developments being. The first thing I did was cut the Gurney flap down to half the height, as shown below.

Small Gurney flap. Is it necessary on the upper wing?

However, I’m not 100% sure the upper wing needs a Gurney flap. The gap between the wings is supposed to accelerate air, and that should be enough to keep air from separating at the trailing edge. Also, the trailing edge of the upper wing is round and thick, and just looking at the camber across the chord makes me think the wicker isn’t helping. At some point after taking this picture, I removed the wicker altogether.

Curving the underside of the main wing

The underside of the main wing is flat and has two mounting grooves that span the length of the wing. This conveniently allows you to mount the wing at any width, and makes the main wing a good deal stiffer. The compromise is less than optimal flow under the wing, not only because of the flat shape, but the grooves.

The profile of the wing is much better without the large flat spot along the bottom. .

I could have simply filled the grooves, but if you’ve been following my blog then you understand I like to go a bit overboard. So I bought a piece of pine baseboard at Home Depot, glued this to the bottom of the wing with epoxy, and sanded it to a rounded shape. If you decide to do the same thing, you need to know exactly how far apart your wing stands are before you do this, as there’s no adjusting the width afterwards.

Pine baseboard, glued and sanded.

Note for Grid Life Touring Cup (GLTC): I was looking at the rules for Grid Life Touring Cup, and noticed they don’t count points for wings that measure 250 square inches or less. The lower wing measures 244 square inches, so you could use this wing without the 3% penalty! You can buy the lower wing alone for about $50.

Finishing touches

All that was left was to sand and paint it, and I think it looks pretty cool. You can still see some of the wood grain beneath, which is either a shitty paint job, or a DIY detail to appreciate, depending on your point of view.

Before shaping the underside, the wing and end plates weighed 7.6 lbs. After shaping the underside, the ensemble weighed 8 lbs. Pine boards are light. I’m not 100% happy with the end plates, they need to be larger, but I can continue to mess with that.

Test results

To see how the wing performed I took it up to my usual testing grounds, Pineview Run. Now this is not the ideal place to test aero, because at lower speed, aero is not as pronounced. I don’t have Man and Machine Consulting’s telemetry equipment, so lap times is the best I can do.

Those excuses aside, my teammate Evan tried the 60” 9LR wing back to back with my new modified wing and somehow the double wing went faster. We did nothing to control variables such as wind speed and temperature, so let’s not call this a conclusive test, but for sure it’s an improvement on the 1.0 version of this wing. Will I race with it? I might use this wing at a Lemons race, or on a short track like Pineview, but for any real racing, I’ll be using the 9LR wing for sure.

DIY Wing End Plates

A wing without end plates allows the low-pressure air below the wing to bleed over into the high-pressure air on top of the wing. This interaction creates vortices, which reduce downforce and create drag. The middle of the wing still works well, but you get progressively less downforce and more drag at the ends. For a quick video on why a wing needs endplates, see this video by Kyle.Engineers.

End plates separate the flow between the top and bottom of the wing, effectively reducing drag and increasing downforce. The end plate has to be large enough to keep these two pressure zones from colliding.

In the following image, notice how different wing shapes have similar high-pressure areas above the wing, but very different low pressure shapes below the wing. Indeed, the shape of the end plate should be similar to the pressure zone shape. Note that the low pressure side (suction) is more important than the high pressure side, and so end plates must extend further below the wing.

End plate shape should match pressure zone shape. Image is upside down so that it relates to car wings. Image from Race Car Aerodynamics, buy the book.

Take a look at the wing shapes above:

  • The first one (on the left) is a wing with a NACA profile around 4410. (4 degrees of camber, max camber at 40% of the chord length, thickness of 10% of chord length). It’s like a less aggressive 9 Lives Racing wing.
  • The second one looks like a skateboard deck. I’ve seen a lot of DIY wings in 24 Hours of Lemons (skateboard decks, snowboards, and just a piece of curved wood), and I love the spirit. Mostly I don’t see them with end plates. Do it!
  • The third one is a symmetrical airfoil. It doesn’t make a great wing for a car, but is good for stanchions and other places where you need to hold something up with little drag.
Inside the race cars of future past
Early days with no end plates and symmetrical air foil. We don’t use this shape now except for wing stands or other braces.

DIY single-element end plate

I use aluminum sheet for end plates (recycled street signs are a good source of aluminum), but you could use any sheet metal, carbon fiber, plywood, alumalite, etc. The endplate needs to be relatively stiff and light.

Different racing organizations have rules on end-plate size, and for simplicity, you can make a rectangle or square of whatever the maximum size is. Some people will cut a notch on the top trailing edge to create a trailing vortex, and you can shape the bottom to match your pressure zone (see first image).

My single-wing endplate, with a wicker

Pro tip: Lay a straight edge across the chord of the main wing, and use that same angle for the top of the end plate. This makes it easy to set and adjust your wing angle using the top of the end plate.

If you look at F1 end plates you’ll notice slots above and below the wing, a leading-edge slat, strakes along the sides, and a gurney flap at the trailing edge. Yes, all of this on the end plate! All of these tricks further improve end plate functionality, but are a bit overkill on a street-based car with an off-the-shelf wing. I personally wouldn’t bother with these modifications, but it’s good for conversation.

Bite-size tech: Red Bull RB12 rear wing endplates
These end plates are overkill on a Miata, but what the heck, let’s talk about it.

End plates for dual wings

Last summer I raced in the 24 Hours of Lemons race at Thompson, and saw some good aero, and a lot of bad. Lemons cars have wings largely for looks, it doesn’t really matter that some of them were a slab of plywood set at an angle. Among these quasi-aero devices were a lot of cheap eBay/Amazon wings that would have worked, but were done poorly.

Case in point: on one orange Chevy Lumina (winner of the IOE), the wing was on backwards. I enquired about this, and apparently the wing came pre-assembled with the pointy part of the wing facing forward! That’s just dumb from the “factory” but shame on the team for not correcting it. Or maybe it was intentional? This is Lemons, it’s hard to tell.

Lemons Pittsburgh: Lemons Adds Pitt Race for 2019, Moves NJMP ...
Winner of 24 Hours of Lemons “Index of Effluency”. The wing is assembled backwards, with the trailing edge pointing forwards. Love those end plates.

At the race I saw a lot of dual wings with absolutely ridiculous end plates that had big holes or cutouts on the underside. As you saw from the first image, the underside of the wing is what matters! Moreover, they had the upper wing mounted so far away from the main wing that it defeated the purpose of a dual wing setup.

Take a look at these disasters, with the meat of the endplate at the back of the wing, or a cutout below that would let the pressure zones collide. It would be easy to correct the function of these wings by building your own end plates.

These end plates do almost nothing for the low pressure side of the wing.
GT Wing Spoiler 52Inch Universal Lightweight Aluminum Rear Spoiler Wing Adjustable Angel Double Deck Racing Spoiler BGW Drift JDM Drift Black
Designed by fucktards. No adjustability of the upper wing, too large of a gap between the wings, and the end plate is facing the wrong way.

DIY dual-element end plates

So if you have a crappy dual-element wing with crappy end plates, and you want to make it work better, build your own end plates. Again, let’s start by looking at the pressure zone below the wing.

This image is from Competition Car Aerodynamics. Buy this book.

Notice that the low-pressure zone extends below the wing by almost the length of the chord of the main wing. Meaning, if you have a 10″ chord wing, you’re going to need at least a 10″ deep end plate. Also notice that the low pressure zone extends in front of the wing, but not much at the trailing edge.

In Competition Car Aerodynamics, McBeath examines what happens with end plates of different sizes. At first he uses no end plate (ep0), and then end plates of increasing size. The larger the end plate, the more downforce and less drag.

End plates of different sizes on a dual-element wing.
End PlateDownforce% IncreaseDrag% Decrease
ep0 (none)769.2Equal194.8Equal
ep1 (minimal)786.72.3%188.33.5%
ep2 (medium)873.413.6%183.86%
ep2 (large)900.117%178.19.4%
Bigger end plate means more downforce and less drag.

OK, so if bigger is better, how big is too big? There is a height at which end plates start creating more drag, and a diminishing return on downforce. But I don’t want to give away all the secrets, so please buy the books on my Resources page and learn yourself some aero.

Make ’em

Here’s how I’d DIY myself end plates:

  • Start with a 12″ x 12″ piece of sheet metal. Use a street sign if you’re Lemons, otherwise plain aluminum will do.
  • Put most of the surface area at the front and below the wing (as pictured in the drawing, above).
  • Lay a straight edge across the chord of the main wing, and use that same angle for the top of the end plate. This will help you set and adjust your wing angle.
  • After mounting the main wing as above, mock up where you want the holes for the secondary wing. I would put a single mounting hole in front that acts as a pivot and drill two or three holes at the rear. Don’t exceed 35 degrees. I don’t trust adjustment slots because they can shift out of whack, and so I go with holes instead.
  • Make the gap between the wings about a half inch in height, and overlap the upper wing on top of the lower wing by about a quarter inch. This should create a convergent gap between the wings, meaning the front opening is larger than the rear. This will accelerate the air going through the gap.
  • Set the lower wing angle almost flat (zero degrees). Most wings will have the highest lift-drag ratio in this vicinity.
  • Start the upper wing at 25 degrees and if you need more downforce, use the 35 degree hole. Don’t exceed 35 degrees with the upper wing. If you still need more downforce, rake the entire wing a few degrees.
Three adjustment holes on the upper wing, the main wing is adjusted by the mounting bracket.

Excuses, excuses…. Ow!

This website is full of data that supports why cars (Miatas) should have aero. Nevertheless, I meet people all the time that say they don’t want to use aero. Fine, don’t. But don’t justify that with bullshit excuses or claims you can’t back up with data. That just gets you a dick punch for being a moron.

What are the most common excuses I hear for not using aero?

Evan’s car has evolving aero, built to a budget.

“My car doesn’t make enough power to use aero”

This is the dumbest one. I wrote about this previously, and it’s just plain wrong. Miatas have terrible aero, and the normal aero things one does to a Miata reduce drag (freeing up power) and add downforce (increasing grip). There’s simply no downside to using aero on a Miata.

  • Drag reduction – Airdam, splitter, flat bottom, low spoiler: all of these reduce drag. Explain to me how the fuck it takes more power to use aero?
  • Wing drag – A single-element wing has about as much drag as your side mirrors. Move the mirrors inboard and shut up. Or just ignore the mirror drag and stand agape as you gain 2-3 seconds per lap.

“I’m not a good enough driver to use aero”

This is my second favorite saying. As if the self-admitted “not-good enough driver” has aerodynamic knowledge, but not driving skill? The fact is, adding grip and stability is not going to make you a worse driver, it’s going to make you a safer and better driver.

  • Understeer is safe – A wing will generate more and more downforce at speed, shifting the balance of the car to the rear, causing the car to understeer the faster you go. This is especially good for people who claim they aren’t good drivers.
  • Easy to tune – Not everyone is schooled in setting up suspension. Conversely, changing the angle of your wing is child’s play.
  • Braking – If you look at data, you’ll see that aero helps braking more than it does for cornering. Drivers who identify as “not good” will benefit greatly from better braking.
  • Stability is good – A stable car is easier to drive at high speed. Cars without aero suffer from aerodynamic lift, which makes the car unstable at high speed. Downforce makes a car easier to drive, and better for noobs. Lack of aero, and instability, require a better driver. Not the other way around.

Obviously a pro driver will get more out of a car that has a lot of downforce, but a rookie driver will also benefit greatly from aero. Saying that there’s some skill level requirement to a car that has less drag and lift is fucking asinine.

“I don’t like the way aero looks”

If you’re choosing your performance based on looks then you’re not serious about performance and shouldn’t be reading my website. For a street car, you have a leg to stand on, but for a race car, notsomuch.

My brother is one of these “I don’t like the way aero looks” people. It’s infuriating because he races a Yaris. Somehow he thinks a splitter and wing are going to ruin its wonderful lines?!? Dude, this is not a Giugiaro masterpiece! Aero all the things, brother.

I finally got Ian to throw a $50 wing on the Yaris and he was able to flat-foot Thunderhill Turn 1 for the first time ever. So at least he’s been partially converted. I’d like to see it with a splitter, flat bottom, barge boards, and diffuser, but I’ll take this small victory for now.

“Aero costs too much”

OK, I don’t really hear people say this one so much as see it. People will dump money into every aspect of performance except aero. They’ll run RE71Rs and change them every 7 hours, Or they’ll do an engine swap or turbo. All the grip and power things cost money, and then keep costing money. On the other hand, once you do aero, you’re done. Aero is way cheaper for the same performance.

  • Tires – Teams that run RE71Rs and the like for endurance racing are spending 3x more on tires. Sure, they are getting a second or two a lap, but also paying for tires (and mounting) 3-4 times more often.
  • Power – Everyone wants more power, but for Miatas without fuel cells, you’re limited by the fuel you can carry. Some motor swaps, like Ecotecs, are pretty frugal on fuel, but more power means more gas. Gas costs money, pit stops cost time, do the math.
  • Aero – Do it once, forget about it. I’ve done a lot of OptimumLap simulations on different tracks, and while some tracks favor one or the other, grip, power, and aero often balance each other out. Over time, aero is simply the cheapest way to get the same lap times.

What’s your excuse for not using aero?

If your racing series doesn’t allow aero, obviously you can’t. But if you’re not using aero for some other reason, tell me why. But be careful… and cover your junk.

Tracks That Favor Power vs Grip vs Aero

Some race tracks are power tracks, and some are grip tracks. At the extreme ends of the spectrum, compare a salt flat to a skid pad. On the salt flat, there’s no cornering, it’s all about power. Likewise, on a skid pad, it doesn’t matter how much power you have, it’s all about grip.

leaf-skidpad | InsideEVs Photos
Some tracks are all about grip.

However, you also have to consider the aerodynamic factors of drag and lift. On a salt flat, drag could be a huge factor. On a large diameter skid pad, aerodynamic lift (downforce), could also weigh heavily (pun intended).

Most race tracks are neither salt flats nor skid pads, but somewhere in between. Some favor grip, some power, and some aero. I thought it would be interesting to run a bunch of simulations in OptimumLap to find out which tracks favor power, grip, or aero.

For the simulator car I used a Miata (surprise!) with 100 hp, 2400 lbs, tires at 1.0g grip, andvCd 0.5, Cl 0.5. These are very low, round numbers, meant only to illustrate the differences. Moreover, the low power and grip values create larger gaps in the data, which makes it more obvious which tracks favor which attributes.

I ran this base car on 12 different race tracks that range from an autocross course to the fastest closed-course race tracks in the USA. There’s a good mix of tracks in OptimumLap, but I also added the map my brother made of New York Safety Track, because it’s local and awesome.

NY Track Info
New York Safety Track (NYST). Dumb name, great track.

After the initial base simulation run, I then added 10% grip and re-ran the group. The lap times dropped across the board (especially at the autocross course), and I logged the data.

I then removed the additional grip and added 10% more power. So at this point, we’re back to the base car, this time with 110 hp instead of 100. Lap times fell again, but not always in the same order.

Finally, I removed the extra power and changed the coefficient of lift from -0.5 to +0.1. This is not a drastic change, one could easily get this from an airdam and small spoiler. I left the drag (Cd) the same, as it probably wouldn’t change much (it would actually have less drag, but I chose to concentrate only on downforce).

The table below shows the base lap time, and then what happens when I added more power, more grip, and minor aero. I ordered the table by tracks that favor power at the top, in the middle are tracks that favor aero, and at the bottom of the table are the tracks that favor grip.

Watkins Glen157.30155.33154.97154.97Power and Aero
Summit Pt93.9392.7692.6992.35Aero
Big Willow109.46107.98108.18107.11Aero
Lime Rock71.0670.0170.2269.60Aero
Thunderhill 142.71140.60140.74140.59Aero and Grip
Start by looking at the Base column, which is the lap time before I modified any values, and then look across for the time in bold, which is the fastest lap.

Race tracks with long straights favor power, and so it’s not surprising to see VIR and Mid-Ohio at the top of the list. Likewise, an autocross course, or a smaller technical track like Waterford, are grip tracks. Finally, some tracks are fast and flowing, and suit aero.

Not every track was definitively a power, grip, or aero track; some tracks are a mix of two. Watkins Glen favors both power and aero, while Thunderhill (3 mile) favors grip and aero.

I then re-ran the simulations using a Miata in a higher state of tune. I left the weight the same, but gave it 120 whp and 1.2g grip, to see if the relationships would change at all. Nope, they stayed the same. Astute readers will cite that a high-power car like a Corvette rarely benefits from more power at any track. But this isn’t a site about Corvettes….

Note: As I stated earlier, the coefficient of lift value (Cl) for the aero car is 0.1, which is pretty low. A Miata with an airdam, splitter, and wing has a Cl of 1.0, which is way better. If I re-ran all the simulations using that value, then it would show that every track is an aero track, and that wouldn’t tell me much about the nature of the track. So I purposely chose a low Cl value that would illustrate which tracks are more favorable to power and grip, and put the aero into equal perspective.

DIY Side Spats

I’ve been de-aeroing my street car this year. What?

Last year I was experimenting with aero and I made an airdam and splitter on my 1993 street car. The airdam was a vertical 3″ extension under the stock lip, and the splitter was simply the forward part of the undertray. I built an adjustable spoiler to balance it out.

September 2019. Small airdam, splitter, adjustable spoiler (the black plastic part can be moved up or down into three different heights).

It went pretty well until I stuffed it into the weeds. The splitter dug in hard, bending the mounting brackets, the undertray got shoved underneath the car, and I sheared some bolts and plastic. It wasn’t pretty.

Oops. Shoved the splitter/undertray back and broke my side spats.

So like I said, I’m de-aeroing my car. My hope is, that if I go off track into the weeds again, the car just glides over them, rather than the splitter ruining my day.

The front end is now the popular R-package front lip. The brake ducts aren’t really functional, so I covered them up. Underneath, I still use a full undertray so I can duct the radiator and whatnot, but the front edge terminates right at the R-package lip.

One of the things I don’t like about the NA front end is that it doesn’t have enough tire coverage, and so I built some front spats to deflect air away from the tires.

R-package front lip with plywood undertray and covered brake duct. From this angle, it looks like the wheel spats cover the whole tire, but they don’t.

Last time I made spats out of extra HDPE plastic I had lying around from making the airdam. This time I decided to use metal, so I could pre-form the shape. It was pretty easy to mock up in cardboard, then bend it over and get the shape right.

I fastened the bottom of the spats to the undertray, sandwiching them between the undertray and the front lip. At the top I folder over the extra material, drilled a hole, and used the standard fender mounting bolt.

Angle aluminum wickers on each side.

Finally I added some 1/4″ angle aluminum to the edge of the spats. These required a bit of effort to form. I had to clamp it in a vise as I bent the arc slowly. If I hadn’t clamped it down, the whole piece of aluminum would have come out bent in more than one dimension.

The purpose of the angle aluminum is to act as a wicker (Gurney flap), and fool the air into thinking the spats are wider than they are. As air moves around the front end, it encounters this raised edge which deflects air around it, effectively making the spats wider.

Wicker helps deflect air around the tire.

I suppose the angle aluminum also makes the outer edge stiffer, but I’m not sure if that was an issue, these are pretty strong and very light. I don’t know if they can handle any serious grass cutting, but at least it’ll just be the spats I lose next time, and not the whole shebang.

DIY Gas Can Mods and Speed Test

At our last race we had two different fuel jugs, VP and Hunsaker. Both jugs hold about 5.5 gallons. The Hunsakers have unleaded fuel nozzles and pour a lot slower, but they have a convenient venting system that keeps you from getting gasoline on your race shoes.

I was curious how fast each can could dump its contents, so I filled them up with about 5 gallons of water and did a little test. But first, let’s see the contenders and how I modified them.


My Hunsaker jugs have the small diameter nozzles that fit a standard unleaded restrictor plate. This has been useful for the Lemons team I race with, since their car (minivan), is street legal. The small nozzle’s .59” inside diameter (ID) restricts the amount of fuel that can flow, and makes fueling take a lot longer.

Unleaded nozzle on the left, DIY on the right

But my race car doesn’t have the unleaded restrictor plate, and so it’s kind of silly to use the Hunsaker cans as is. However, if I remove the nozzles, the large diameter tubing on the Hunsaker is too large to fit in my filler tank neck. So I made new nozzles that were as large as I could fit inside the clear tubing. I bought 1.25” aluminum tubing which has a 1.12” ID.

VP Jugs

We typically use the VP jugs for fueling because the hose is a larger diameter and they dump faster. The VP caps have a screw-in plug, and if you remove it, you can thread in a brass adapter that fits a 1” tube. Add a piece of tubing and a hose clamp and you’re done. The inside diameter of the brass adapter is 3/4”.

At the last Lemons race, my teammate Dieter modified one of my VP jugs with a Hunskaker-like air vent. This seemed to work pretty well, as no fuel escaped the vent or got on our shoes. I bored out the vent tube a bit to give it freer breathing, it looks like this:

Poor-man’s Hunskaer

Speed test

First up is the standard Hunsaker with unleaded nozzle: 44.4 seconds. That’s with the small air valve popped open. My teammate Pat always removes the entire vent cap from the Hunsaker jugs, instead of just opening the vent. I always thought that was silly, but I wanted to see what the difference was in dump time: 38 seconds. Huh, he wasn’t wrong about that.

Next up is the standard VP jug with homemade nozzle: 25.4 seconds. Having seen the benefit of a larger vent hole, I opened up the vent on the VP can. This can also has the “poor-man’s” Hunsaker vent, which is basically a long piece of tubing on the vent hole so you can invert the can and no gas drips out. This one dumps at 22.5 seconds, better still!

Finally, I tested the Hunsaker fuel jug with an aluminum 1-1/4” tube in place of the unleaded nozzle. Having already seen the benefit of the unscrewing the vent cap, I did that: 10 seconds. Woot!

Fuel Can ConfigurationIDSeconds
Hunsaker, unleaded nozzle, small vent.59″44.4
Hunsaker, unleaded nozzle, unscrew cap.59″38.0
VP with standard air vent.75″25.4
VP with larger vent and tubing.75″22.5
Hunsaker with 1-1/4″ aluminum nozzle1.12″10
Final results

I was actually unprepared for how fast the last configuration dumped the fuel, and could be off by a second when I fumbled to time it. It doesn’t really matter anyway, the point is to use the largest diameter nozzle you can, and don’t neglect the size of the air vent.

I Don’t Want a 1.8 Swap, Part Deux

I previously blogged about all the reasons I wasn’t going to swap a 1.8 engine into my 1.6 Miata. I’ll sum it up with the phrase “I’m not building a better car, I’m building a better Miata.” The goal is to accentuate everything that made the 1.6 different, and more Miata, than the later 1.8s. Part 2 of this blog is my progress report. Where am I at?

  • Higher compression ratio – Done. I bought a spare head on eBay so I could do all the work on the bench. I got the head decked .035″ for $50, and the engine should have about 10.3:1 compression now. I also put in Supertech +1mm oversize intake valves. I didn’t do the exhaust valves because I was trying to save a bit of money, and my understanding is the intake-to-exhaust valve diameter ratio is still good. I went with OEM valve guide seals after reading multiple cases of how much the Supertech valve seals leak. I did a simple DIY porting job, mostly just knocking down casting flash and blending. My teammate Evan Merrill’s landlord is a retired machinist and does heads in his spare time. He does top-notch work at a very reasonable price.
  • Hotter cams – Done. Kelford 203-B cam, 264/264 degree camshafts with 9mm lift. This is a medium cam with a lumpy, but streetable idle. I’m still using stock valve springs and retainers, nothing radical. Note that there is a normally-aspirated version of this cam, which has a bit less duration on the exhaust cam. I got the forced-induction cam with an eye on the future.
  • Higher redline – Done. The 1.6 has a shorter stroke, and so it’s safe to rev it a little faster. 7000 RPM on a 1.8 is the same piston speed as 7125 RPM on a 1.6. It’s easy to set it anywhere I want with the Megasquirt PNP2. I have it set to retard the timing at 7200 rpm, which is stock, and cut fuel at 7500. Also worth noting, I got a Speedy EFI variable TPS sensor.
  • Lighter flywheel and clutch – Done. Fidanza 7-lb flywheel and Exedy stage 1 clutch. The flywheel feels great when rev matching, but I’m having a bit of buyer’s remorse on the clutch. It works great, I’m just thinking I should have gone for a heavier clutch in case I go forced induction in the future. This clutch is good for 145 ft-lbs, which is fine for anything normally aspirated.
  • Less weight – Done. I removed the A/C (needed to be recharged and messed with my idle), and added lightness here and there. My already svelte car is down to 2100 lbs, and that’s with an OEM hard top, Hard Dog roll bar, full carpet and interior, tow hitch, and half tank of gas.
  • Shorter final drive – Punt…. After a lot of number crunching and simulations, I decided to go with a 4.1. It’s an hour drive to Pineview Run, and some of it is on the interstate, and I just didn’t want the engine buzz. As a side benefit, the speedometer is now accurate. However, if I was hardcore about sticking to my principles, I would have shortened the gear ratio. If a 4.625 or 4.778 comes up at the right price, I still might do it.

Another place I punted on the 1.6 theme was retaining the 6-inch ring gear. These are known to fail under stock power, and only people hamstrung by STS rules actually wants one. So I swapped in a 7″ ring gear and Torsen from a later model, which required new axles and drive shaft.

I took the car to Rick Gifford for tuning. He sees a lot of Miatas and Megasquirts and has been known to work a bit of magic. He has a Land and Sea dyno (DYNOmite) mounted flush with the floor, it’s a neat drive-up experience. The dyno measures low numbers, like a Mustang in terms of power readings. I multiply by 112% to get a Dynojet number.

On Rick’s dyno, I got 129.7 hp. Power maxes out at 6800 RPM, which is only 300 RPM more than stock, so the cam didn’t make it a peaky powerband at all. The engine made 109 ft-lbs of torque at 5300 rpm, without any big dips, and a fairly flat curve. Check it out.

Apples to Apples to Oranges

Let’s compare my 1.6 build to other Miatas and engine swaps. A stock 1.6 engine would put out about out about 83 hp at the wheels on Rick’s dyno. My 130 hp is 156% more power. This is about 145 hp and 122 torque if measured on a Dynojet. Woot!

On the same day we dynoed Clayton’s NB1 with a gutted cat, Cobalt exhaust, and cold-air intake. It made 117 hp and 108 torque. That’s about 131 hp, 121 torque on a Dynojet. Of course it could be further improved with a header and standalone ECU, but even using the stock ECU, this is good power, and that’s why people swap 1.8s instead of tuning 1.6s.

But if you’re going to do a 1.8 swap, may as well use the later VVT engine for a bit more grunt. Coincidentally, right around the time I was getting my engine work done, Napp Motorsports was filming a multi-part YouTube series with their two Miatas: the first a VVT swap and the second, a K-swap. Stefan Napp also takes their cars to Rick Gifford for tuning, so this gives me a chance to do an apples-to-apples comparison.

Some background on the Napp cars. Dylan’s car started as a NA6, and, not satisfied, he did a NA8 swap with an exhintake cam. It dynoed 102 hp on Rick’s dyno, and in Dynojet terms that’s more like 114 hp. That’s not a very strong NA8, but it was choked by a 1.6 wiring harness including the AFM flapper valve. In the end, Dylan decided a VVT swap would be better.

But Stefan Napp didn’t just swap Dylan’s motor, he decked the head .040″, did some port polishing, and got a good valve job. The engine now has 11:1 compression, and exhales through a custom fabricated exhaust system. It has a standalone ECU, and on 93 pump gas it put out 135 hp and 126 ft-lbs of torque. In Dynojet numbers that’s a strong 151 hp and 141 torque. (Side note, it only made +5 hp on E98. Booo, corn.)

When you compare my 1.6 to the 1.8 VVT, the bigger motor makes 5 more hp and 17 ft-lbs more torque. The 1.6 has always been gutless in terms of torque, and right there you see it.

Stefan’s personal Miata has a K24Z3 swap, which is a 2354cc 4-cylinder used in an Acura TSX and other cars. This is the easiest way to get a Honda engine in a Miata. It initially made 175 hp on the dyno, but after further tuning, Rick managed 197 hp, and 150 torque. Wow.

EngineHp / TqDynojetHp/ Liter
1.6 – Stock83 / 6693 / 7458.2
1.6 – Mine, with cams, etc.130 / 109145 / 12290.7
Dylan’s NA810211462.1
NB1 – Clayton117 / 108131 / 122171.2
NB2 – Dylan135 / 126151 / 14182.1
K24Z3 – Stefan197 / 150220 / 16893.4
Hp/Tq is from Rick’s dyno. I multiplied by 112% to get Dynojet-like numbers. The Hp/Liter values are based on the Dynojet number.

In terms of efficiency (hp/liter), my little 1.6 kicks ass on the VVT motor, but gets beat by the Honda engine. Not that Hp/Liter is important, but it’s a good measure of the state of tune. Dylan is planning future mods to hit the elusive 100 hp/liter, so his 1.8 may yet surpass my 1.6.

Where’s the power come from? What did it cost?

I have a spreadsheet where I’ve entered data from various dyno charts I’ve found on the web. By looking at the parts that were used and the total output, I can make a fair guess at which parts produce how much power. Based on that data, this is what made what, and how much it costs.

DIY cowl intake5%105%$40
ECU, tuning15%120%$1400
Cams, cam gears15%135%$700
+1 mm intake valves + seals, valve job4%139%$500
DIY port and polish2% 141%$0
.035″ head shave6%147%$50
Raceland header4%151%$170
Hi-flow cat1%152%$90
Cat-back exhaust4%156%$350
My best guess as to where the power came from, and what it cost me.

By running a standalone ECU (MS PNP2), I was able to get rid of the restrictive flapper valve (AFM). This is really the first thing anyone should do to a 1.6, and it makes all the other mods tunable. But it’s not cheap, and when you add in $600 for dyno tuning, this comes to nearly $100 per 1% power. That’s my ceiling for my “worth it” calculation, and I like to be closer to $50 per 1% (or $50 per hp, if you think that way). But the ECU makes all the other things better, so you have to start there.

The cams added about the same amount of power as the ECU, and were a big bang for the buck. I probably didn’t need the adjustable cam gears, but I decked the head and thought it was a good idea to get the cam timing back to stock. I still haven’t played with the cam timing, so there’s probably a bit more power there.

Keep in mind that these individual power gains are all guesstimates, I certainly wasn’t going to dyno and tune each part as I put it on. In all I spent roughly $3300 to get 56% more power. That’s $63-70 per hp, depending on the dyno.

As you can see, modifying a 1.6 engine isn’t the best use of money, but almost nothing car related is. Common wisdom is to just drop in a VVT and be done with it. I can’t argue with that, especially after driving Dylan’s car (future blog post). But if you have a low-milage 1.6 and feel like it’s dumb to swap a perfectly good motor, or if you just want to buck the motor-swap trend, then tuning a 1.6 can be pretty rewarding.

I feel like the project has been a total success. I’ve improved everything that made the early cars better Miatas. In terms of power to weight ratio, the car is about 16.1 lbs/hp, with me in it. That’s about the same as a new ND2 Miata, and in that light, a downright bargain.

But it’s no rocket ship. When I hit a long straightaway, I’m still patting the center console giving her the giddy-up: “C’mon baby, let’s GO!” An underpowered is still part of the Miata charm, after all.

I got to track test both of the Napp Motorsports Miatas back to back to back with my NA6. I’ll report on that after their video comes out.

Pushing the Play Button

I started a new business with my wife, we built a mobile app for stepmomz. App development is going slower than I’d like, and so I have some time to write again.

Here are a few things you can look forward to in the next few posts:

  • Updates on my tuned 1.6 (AKA “I don’t want a 1.8 swap”).
  • Trackday tire test: four Miatas, four AIM Solos, 10 wheel/tire combinations. A7, R7, RC1, RE71R, RR, RS4, VR1, Z214, lots of fun, lots of data.
  • Three-way track test of Napp Motorsport’s K24Z3 KMiata, VVT swapped NA, and my NA6.
  • More tire tests at faster tracks.