The Middle Child

There is a specific kind of car that the world decided to forget on purpose, and the 1989 Mazda Savanna RX-7 — the FC3S, if you want to be proper about it — is the patron saint of the type. Everyone remembers what came after: the FD, all curves and menace, the one that ends up on the bedroom wall. And everyone romanticizes what came before: the first-generation wedge, the plucky underdog that beat Porsches outright at Spa and looked like a Japanese answer to a question Stuttgart thought only it was allowed to ask. The FC is the middle child. It borrowed its shape, unapologetically, from the Porsche 944 — Mazda's engineers have never pretended Porsche wasn't the target — and for that it has spent thirty-five years being called derivative by people who never drove one. They should have driven one. Because underneath that borrowed suit was something Porsche never built and never could bring itself to: a spinning triangle where the engine should be.

That triangle is the whole story. The 13B — twin rotors, 1.3 liters of displacement that behaves like nothing with pistons — doesn't make power so much as release it, in a smooth unbroken pull toward a redline that arrives sooner than your instincts expect and sounds, at the top, like tearing silk. The turbocharged version, the one Japan mostly kept for itself, made around 200 horsepower, which the spec sheet will tell you is unremarkable and the spec sheet will be wrong, because the FC never felt like its numbers. It felt like a car sanded down to only the essential parts and then balanced on a knife's edge — light, honest, a little nervous in the rain, willing to rotate the instant you asked and occasionally when you didn't. It asked something of you. That was the point. It assumed you wanted to be involved, and it was right.

And here is the cruelty of it. The rotary is gone now, killed by the same emissions math it could never quite solve, and every year the survivors get rarer, because a rotary rewards devotion and punishes neglect with the patience of a landmine — miss a rebuild and the apex seals will remind you what you owe. So the FC has quietly become the thing nobody expected a Porsche-shaped middle child to be: proof that the interesting answer and the sensible answer are almost never the same car. It was never the fastest, never the prettiest, never the one on the poster. It was just, stubbornly, unlike anything else on the road — and thirty-five years later, that turns out to have been the rarest thing it could possibly be.

01First, the one that earned the name

You cannot defend the FC without a nod to the car that made it possible. The first-generation RX-7 — the SA/FB, the wedge — arrived in 1978 and did something a small Japanese company had no business doing: it went to the great endurance races and beat the establishment at its own game. On its IMSA debut at the 1979 Daytona 24 Hours, RX-7s finished first and second in the GTU class and fifth and sixth overall. In 1981 a Tom Walkinshaw Racing car won the Spa 24 Hours outright — the first Japanese car ever to win a 24-hour race — beating the Porsches that were supposed to own the podium. It won the British Touring Car Championship. In Australia, Allan Moffat took the 1983 Australian Touring Car Championship in one and stood on the Bathurst podium three times.

That is the inheritance the FC was handed: not a styling exercise, but a nameplate that had already humiliated better-funded, bigger-engined machinery using a 1.3-liter engine most of the paddock regarded as a novelty. The FB was the legend because it was first, and because being first and winning is a thing you only get to do once. The FC's job was harder. It had to grow up without getting boring.

02"It should have been called the RX-8"

When Mazda revealed the second-generation car for 1986, it framed the concept as "New Adult Sport" — a rotary sports car that kept the delicacy but added refinement, comfort, and a genuine attempt to challenge Porsche on its own turf. The old live rear axle was gone, replaced by a sophisticated independent rear suspension. Recirculating-ball steering gave way to rack and pinion. Four-wheel disc brakes became standard. It won Motor Trend's Import Car of the Year for 1986; the Turbo II landed on Car and Driver's 10Best. This was not a soft car. It was a more complete one.

And here is the detail that reads, in hindsight, like a prophecy nobody caught. At the 1986 press launch, Motor Trend reported that the assembled journalists — driving a car so thoroughly new it shared almost nothing but its name with the wedge — decided, half-joking, that it deserved a new name too. Their suggestion:

"New enough that all the journalists at the Mazda press preview decided it should have been renamed RX-8. At the very least."

Hold that thought. It will matter enormously by the end of this piece, because seventeen years later Mazda would build a car called exactly that — and it would turn out to be less the FD's successor than the FC's.

And the grown-up brief was not timidity. It was an answer to a question Mazda went and asked. Chief engineer Akio Uchiyama sent a small team to live in the United States for three months to learn what Americans actually thought of the RX-7, and by the spring of 1982 the company was running dealer clinics with three explicit proposals on the table: a Realistic Sports Car at nine to twelve thousand dollars, a Civilized Sports Car at around eleven, and a Technologically Advanced Sports Car with a rear transaxle and adjustable suspension at thirteen to fifteen. The Civilized Sports Car won, conclusively, and everything the FC is followed from that afternoon. The car is soft the way a decision is soft. Mazda asked, and America answered.

There is a second frame worth having, because it explains the FC's whole position in the family. Yoshihiko Matsuo, who drew the original Fairlady Z, split the category in two: the Lightweight Sports Car and the Super Sports Car. The FB sat near the lightweight end of that line at around 2,420 pounds. P747 was aimed deliberately at the other end — bigger, faster, heavier, better equipped. Which left a hole where the lightweight car used to be, and Mazda president Kenichi Yamamoto knew exactly what a hole like that was worth. The car that eventually filled it was the MX-5. The FC did not merely get overshadowed by its siblings. It stepped aside and created one of them.

A word on the Porsche 944 comparison, since it follows the FC everywhere: it is true, and it is overstated. Mazda openly named Porsche as the performance target and developed the independent rear end specifically to sharpen cornering. But the resemblance is a family one, not a copy — the proportions rhyme, the details do not, and the driving experience is unmistakably Mazda: lighter in its responses, smoother in its power, and dominated by a soundtrack no Stuttgart four-cylinder has ever made. Borrowing a silhouette is not the same as borrowing a soul.

It is also worth knowing how Mazda arrived at the shape, because the story is better than the accusation. Five designers were sent to Mazda's California studio in early 1982 under Yasuji Oda. Ten renderings were cut to two: one that looked a great deal like the Supra that would arrive in 1986, and one that became the RX-7. Resin models of both were shipped to America that autumn and set out anonymously alongside a first-generation RX-7, a 280ZX, a Supra, a Corvette — and Porsche's new 944.

The Porsche won the clinic. The eventual RX-7 placed a clear second.

So the FC did not copy the 944. It lost to the 944 in a blind test, went away for four years, and came back and beat it on a road course — a result you will find further down this page. Along the way it rejected a Z31-style headlight arrangement, kept its width under 1,700 millimeters to stay out of a punishing Japanese tax bracket (which is why the bulging wheelarches on the clay models never made production), and settled at a drag coefficient of 0.31, or 0.29 with the full aero kit.

Oda himself put the design brief in terms this article cannot improve on. The first-generation body, he said, had something rough-hewn about it — it felt like a child that had not grown up yet, and yet it was completely distinctive. The second generation was drawn to mature and refine exactly those features. The people who built the middle child knew precisely what they were doing, and they used the same word for it that we are using here.

03The machinery the technofreaks drooled over

The 13B does not have conventional pistons. Two triangular rotors turn inside oval housings, opening and closing the intake, compression, combustion, and exhaust phases as they spin. That gives the car its defining traits: the engine is physically tiny, so Mazda could sit its mass low and far back in the chassis; it revs fast and, when healthy, feels almost eerily smooth; and instead of the thump of big pistons it builds a rising mechanical hum that sharpens toward the redline. The turbo adds useful midrange, but the car still wants revs. Its appeal was never shove — it was the combination of smooth boost, quick response, a light nose, and a gearshift period testers kept calling one of the best in the business.

But the engine, strangely, is not where the FC's engineers showed off most. That honor goes to the rear suspension, and to a piece of mechanical choreography Mazda called DTSS — the Dynamic Tracking Suspension System — which is quite possibly the finest passive suspension of its era. Passive, meaning it carries no sensors and no computer deciding what to do; it simply reacts, mechanically, to the forces the car generates. Through an arrangement of links, bushings, levers, and bearings that period reviewers described as "mechanical monkey motion," the DTSS toes the rear wheels slightly out at gentle cornering loads — which makes the car feel agile and eager to turn — and then, as lateral load climbs past roughly 0.4 to 0.5 g, toes them in, which makes the car planted and stable exactly when stability matters most. Agility and security had always been opposing demands. The DTSS delivered both, with no electronics at all.

What that description leaves out is that DTSS almost did not happen, and that its survival came down to one engineer refusing to accept a vote.

Uchiyama handed the chassis to Jiro Maebayashi, who pulled in a man from another department named Takao Kijima — a name worth remembering, because he would go on to head development of both the RX-7 and the MX-5. Double wishbones were the obvious answer and would not package, so a semi-trailing arm layout was chosen as the compromise and became the basis for everything that followed. Two mules were built by the spring of 1982, one carrying the new triaxial floating hub, the other on Chapman struts for comparison. Mazda's own racing drivers — Yoshimi Katayama, who had helped sort the first-generation car, along with Yojiro Terada and Takashi Yorino — were brought in to drive both.

By the end of that year the consensus was to abandon the semi-trailing arm entirely.

Kijima went back to the computer instead. He gave the floating hub different bushing values, split each trailing arm into two parts so the car gained anti-squat and anti-dive behavior while taking up less room, and added camber-compensating links to answer a separate complaint. The first revised mule ran in May 1983; seven were eventually built and one was shipped to the Nürburgring. A final prototype in early 1984 added forged aluminum A-arms and a rear subframe after the earlier cars proved unbearably noisy. That is the car that reached production. The idea underneath it — near-limitless toe control without electronics or hydraulics — had come down from the MX-02 concept's four-wheel steering, which was simply too expensive to build for money.

So the finest passive suspension of its era was a compromise its own development team voted to kill, rescued by one man with a computer and a conviction. Thirty-five years later it is still the best thing about the car.

Nor did the cleverness stop there. Up front, to kill the harshness that anti-dive geometry usually inflicts under braking, Mazda fitted the wide-based lower A-frame with bushings that were soft fore-and-aft but rigid side-to-side — supple over bumps, precise in cornering. The steering was electronically assisted, with a computer metering the boost from sensor data to keep the effort constant from parking lot to fast sweeper. An optional Auto Adjusting Suspension read speed, braking, and steering angle, calculated lateral g, and switched the dampers between three firmness levels in microseconds — jumping to full-hard the instant it sensed anything destabilizing. In 1986, in a car this size, from a company this small, that was genuinely leading-edge hardware. The FC's reputation as the "soft" RX-7 was always a slander written by people reading the curb weight instead of the engineering drawings.

041989: the year it grew teeth

1989 was the FC's turning point. In April, Mazda rolled out the substantially updated version enthusiasts call the Series 5 — and it was far more than a facelift. Output climbed from the earlier 185 PS to 205 PS; the compression ratio rose; the twin-scroll turbo was revised; dampers and suspension mounts were improved. Visually, the quickest tell is the trio of round taillights that replaced the earlier units — if you see three circular elements per side, you're looking at the desirable later car. Because the change happened mid-year, a 1989 registration can be either the earlier S4 or the facelifted S5; the taillights are the tell.

Underneath, the specifics matter more than the horsepower headline. The twin-scroll turbo gained an electronically controlled wastegate, which is where the S5's improved low-speed response actually comes from. Compression went from 8.5 to 9.0 to one, and premium fuel stopped being a suggestion and became a requirement. The anti-roll bars were remounted on ball joints to take noise out of the structure. The automatic gained electronic control with manual hold on all three lower gears, the shift linkage was reworked and the lever itself redesigned, and the cabin got a new fascia with proper circular gauges and seats lifted from the limited-edition cars. The GT grade was dropped outright; the GT-X inherited an aluminum hood. American cars gained fourteen horsepower on the naturally aspirated engine — a higher 9.7:1 compression, a new ECU, a redesigned airflow meter, a lighter flywheel and lighter rotors — for 160 horsepower at 7,000 rpm.

And here is the loop closing. Among the ideas rejected during the original P747 design work in 1982 was a set of circular taillights, proposed as a nod to the earlier rotary cars and turned down. Seven years later they arrived: dual circular lamps under a single smoked lens. The detail you now use to identify the good one is a 1982 idea that finally got built.

Then Mazda did something no company does for a car it has given up on. The facelifted range went on sale on the first of April 1989 with prices that had gone down, and a launch film in which Paul Frère and Nino Vaccarella drove the new car around the route of the Targa Florio. Two Le Mans winners, a road that had stopped hosting a race a decade earlier, and a mid-cycle refresh of a Japanese sports coupe. Somebody in Hiroshima still believed in this thing.

There is a darker note under all of it, too. In October 1988, five months before the best version of the FC reached showrooms, Felix Wankel died. He had announced plans only five years earlier for a new engine using two differently shaped rotors, and it never came. For a great many people in Europe, the idea died with him — which makes the S5 something like the last argument made on the rotary's behalf while it still had a future.

One point of pedantry that matters to buyers: in Japan the S5 turbo trims were the GT-R (the plain, cheaper starting point), the popular GT-X, the plush GT Limited, and a GT Limited Special Edition above that — all sharing the 205 PS engine, differing in equipment, and priced from roughly ¥2.10 million to ¥3.12 million. A Savanna GT-R or GT-X is not the same thing as a "Turbo II." Turbo II was primarily the export badge. They're mechanically kin, but the trims, equipment, and names were different — a distinction worth knowing before you pay JDM money for what someone's calling by an American name.

05The whole family, by the numbers

The FC was never one car; it was a range, and the range is where you can see Mazda's ambition. Here is the US-market spread, from the everyday coupe to the stripped-out enthusiast special to the two versions of the Turbo II that bookended the generation. (Period figures; sources and test methods vary, so read these as the shape of things rather than the gospel.)

*Approximate period new prices. The GTUs was the enthusiast's pick: a decontented, lighter non-turbo with the good brakes, sport seats, a chin spoiler, bigger wheels, and a shorter final drive — even air conditioning was optional.

But the US table is only half the family, and the better half never left Japan. From the moment Takaharu Kobayakawa took over the RX-7 program in January 1986 — one of Mazda's original 47 Samurai, and a man who made no secret of preferring lightweight sports cars to luxury grand tourers — he began quietly undoing the Civilized Sports Car brief from the inside. The instrument he used was a series of limited editions called Anfini, badged with an infinity symbol, and they are the FC's hidden halo.

His own explanation of their purpose was blunt: they existed to strengthen the brand's image and to work as development vehicles so the RX-7 series could keep evolving, and the chassis work in particular carried a great deal of new technology. In other words, these were not trim packages. They were rolling R&D, sold to the public in batches of a few hundred.

Read that table as a sentence and it says something remarkable: the FC's own chief engineer spent five years making it lighter, sharper and less civilized, four hundred pounds of paperwork at a time, right up until the FD arrived to do it properly. The Anfini cars are the FD's ideas arriving early, wearing the middle child's body.

They were also barely advertised. Kobayakawa has said the budget was so small he had to rely on press coverage to tell anyone the cars existed. Car Graphic, having driven one, called it wasteful that so few were being built, certain that demand would outrun supply. It did. When Road & Track finally got one shipped to America to try, it recorded 0.91 g on the skidpan and observed that where the standard Turbo takes on a noticeable understeering set when pushed to its limit, this one simply stays neutral — as neutral, the magazine said, as Switzerland.

The same instinct produced one more car worth knowing. In June 1991 a Mazda 787B won Le Mans outright, and two months later Mazda built a thousand Savanna RX-7 Winning Limited models to mark it — GT-R underneath, Anfini above, with sixteen-inch BBS wheels, the full aero package, a strut tower brace and a Momo wheel, in Shade Green or Blaze Red. A commemorative edition of a car nobody was buying, celebrating a victory by an engine everybody was about to abandon.

And the American equivalent of that instinct was the GTU. The 1988 car did without power steering and most of the electrical gadgets, weighed some 66 kilograms less than the Turbo, and was described by Road & Track as one of the truly great driving machines of its day — a superb balance of everything that makes a car go without giving up the things that make it pleasant to sit in. Mazda then moved the GTU badge onto the base car for 1989 and created the stripped GTU-S above it: aluminum hood, four-pot front calipers, limited-slip, a short 4.30 final drive, sixteen-inch alloys and an AM/FM radio. About a hundred sold before it was quietly dropped from the price list. If you are hunting the purest FC that ever wore American plates, that is the one, and almost nobody knows it exists.

06What the FC did on a track

The nameplate's racing heyday belongs partly to the wedge, but the FC did far more than coast on the FB's reputation. In America, the RX-7 became the most successful car in the history of IMSA — more wins than any other model, full stop. It took the GTU drivers' championship seven years running and gave Mazda eight consecutive GTU manufacturers' titles from 1980 through 1987, which took the most-championships record off Porsche. And from 1982 it won its class at the 24 Hours of Daytona ten years in a row. The competition weapon those later years produced is one of the great cult racers: a spaceframe, four-rotor RX-7 GTO wearing FC bodywork and making some 640 horsepower, which won four races in 1991 on the way to the drivers' and manufacturers' titles and screamed at a pitch that still empties grandstands in your memory. That car also took the nameplate's 100th IMSA victory, at San Antonio on 2 September 1990 with Pete Halsmer driving, 12 years after the first one at Daytona. Its four rotors came out of the same program that would win Le Mans outright the following June. On the grassroots side, the FC became an SCCA staple — cheap, tough, easy to work on, and quick enough to embarrass more expensive metal in improved-touring and showroom-stock racing for decades.

It traveled, too. In Australia — where Moffat's first-gen cars had already made rotaries folk heroes — the FC went on to win the Bathurst 12 Hour three years straight, from 1992 to 1994, in production-car trim, with a fourth win in 1995 once the event moved to Eastern Creek. (The great might-have-been is the turbocharged Group A touring car Mazda planned and then abandoned when the rules and the budgets changed; on paper it would have been a monster.) In Japan it ran the domestic touring championships; in Europe the nameplate carried the endurance pedigree the FB had built at Spa and Le Mans. It was, in short, a genuine racer wearing a road car's clothes — which is exactly the reputation the FD would later inherit and bank.

07The second life: touge, drift, and the White Comet

Here is where a forgotten car quietly became immortal. The FC's defining pop-culture role is not a movie cameo — it is the strategist's weapon in Initial D, the manga-turned-anime that taught a generation what a Japanese mountain pass sounds like at 2 a.m. The hero's AE86 gets the glory and the yellow FD gets the second-most posters, but the car that runs the whole operation belongs to Ryosuke Takahashi: a Crystal White Savanna RX-7 ∞ III — an FC3S — known as the White Comet of Akagi.

Notice the badge on that name. The ∞ III is not a trim level invented for a cartoon — it is the Anfini III from the table above, one of six hundred cars, with 215 horsepower, a Torsen differential, BBS wheels and Pirelli P Zeros. The writers did not hand their strategist a nice FC. They handed him the rarest and most focused FC Mazda ever built, and then wrote him as the man who wins by thinking. (One caveat for the pedants: the factory color options for that specific run are worth confirming before you argue about paint codes on a forum.)

And the casting is perfect, because it is exactly the argument this piece is making, dramatized. Ryosuke is the brain — the analyst, the tactician, the one who wins by reading the road like a chessboard rather than bludgeoning it. His younger brother Keisuke drives the FD: louder, angrier, all aggression. The writers gave the thinking driver the FC and the hothead the FD, and nobody who understands these cars found that surprising for a second. The FC was always the driver's car for people who wanted to be involved rather than merely fast.

The show handed the strategist an FC and the hothead an FD. That casting is the entire case for the middle child, told in cartoon form.

Off-screen, the FC became one of the pillars of drift and tuner culture — rear-drive, near-perfect balance, a light nose, and an engine bay so cavernous relative to the tiny rotary that owners have spent thirty years stuffing bigger hearts into it, right up to triple-rotor 20B swaps. As FD prices climbed into supercar territory, the FC inherited the mantle the FD vacated: the last genuinely affordable, endlessly modifiable, truly unusual way into rotary ownership. Cults form around cars like that — cars that ask for devotion and give back character no piston engine can. The FC has one of the most loyal.

08The one on the wall

Then came the FD, and it is only fair to bow to it. The third-generation RX-7 is one of the most beautiful cars Japan has ever produced, and one of the few whose styling genuinely does not age — it doesn't look like 1992, it looks like a shape that arrived fully formed and simply waits for the rest of the world to catch up. Sequential twin turbos, a chassis of near-telepathic delicacy, a curb weight kept fanatically low: it is routinely and correctly called one of the greatest driver's cars ever built. The FD is the extremist of the family, the car that took everything the rotary promised and refused to compromise any of it. No argument here. It earned the wall.

But the FD's brilliance is also the source of a category error that has followed the RX-7 for thirty years — the assumption that the FD is the template, and every other RX-7 is measured by how close it comes to that hard, uncompromising ideal. By that ruler the FC looks soft. By any other ruler, the FC is simply a different and equally valid idea of what a rotary sports car is for.

There is one more FC that never existed, and it belongs here. While the FD was being drawn, Mazda was testing a three-rotor version of the second-generation car. It made 220 horsepower and 195 lb-ft with what the engineers described as exceptional flexibility, and it weighed only about 65 pounds more than the naturally aspirated 13B. It never entered production. The three-rotor engine eventually reached the road in 1990 in the Eunos Cosmo, breathing through sequential twin turbochargers — the arrangement that would define the FD. For a moment in the late eighties, the middle child was being fitted with the future, and Mazda decided to give the future to somebody else.

09The heresy: the RX-8 is the FC's true heir

Which brings us back to that 1986 joke. The journalists who said the FC should have been called the RX-8 were laughing at how new it felt. But they were also, without knowing it, describing a lineage — because the car that eventually did wear that name turned out to be the FC's spiritual child, not the FD's.

Look at what the RX-8 actually was, underneath the four doors everyone fixates on. A naturally aspirated rotary — the 232-horsepower Renesis — in a front-mid, rear-drive chassis with a 50:50 balance and, by most accounts, handling sharp enough to trouble the FD. It reached 60 in the mid-six-second range. It was comfortable. It was practical. It was fast enough to satisfy and civilized enough to live with every day. It was, in Mazda's own framing, a rotary sports car built to be usable, refined, and welcoming — a car that assumed you had a life to fit around it.

Now read that description again and tell me it isn't the FC's. "New Adult Sport." The grown-up rotary. The one that flows rather than attacks, that is quieter on the freeway than its styling suggests, that trades the last few percent of hardcore edge for the ability to be a real car on a real Tuesday. The FD was never going to have a successor in spirit, because the FD was a one-off act of obsession — you cannot follow "no compromises" with anything but more of the same, and Mazda couldn't afford to. The RX-8's brief was the FC's brief, thirty years on: take the miraculous engine, wrap it in something a normal person can love and use, and make the case that a rotary belongs in your driveway, not just on your wall.

Chronology says the RX-8 followed the FD. Character says it followed the FC. The two of them are the bookends of the same sane, humane idea — the rotary as a car you drive, not a car you worship — with the FD's beautiful obsession sitting in between them like the exception that it was. The journalists in 1986 were more right than they knew. The FC really should have been called the RX-8. It just took Mazda seventeen years to build the rest of the sentence.

Kenichi Yamamoto, asked in 1988 whether every future RX-7 would be rotary-powered, answered without hedging: Mazda and the rotary were inseparable, it was the image of the company, abandoning it would be like losing the company's identity — and the RX-7 could not and would not exist without it. He was right about all of it, including the part he did not intend. When the rotary finally went, the RX-7 went with it, and so did the RX-8. What is left is the small population of cars that carried the idea while it was alive.

10The Hagerty view: what one costs, and what to look for

The market has finally started to notice, though it remains, for now, the bargain of the rotary world. Hagerty flatly calls the second-generation RX-7 the least expensive rotary sports car you can buy, and the sale data backs it up: the average FC changes hands in the mid-teens, with clean S5 Turbo II five-speeds and the handsome NA convertibles carrying the premium, exceptional cars occasionally cracking into the thirties or beyond — the record is a $63,000 Turbo II — and rough, project-grade examples still turning up for the price of a decent used commuter. Against an FD that now routinely trades from the forties into six figures, the FC is the last RX-7 a normal enthusiast can actually afford.

The scarcity is not a market quirk, either. It is the record of a car America stopped buying while it was still getting better. US sales fell by about a third over 1988. In 1990 the RX-7 accounted for 9,743 of the 226,000 Mazdas sold in the States. In 1991 it was 6,986, and there was no 1992 model year at all — the leftovers were simply sold off until the stock ran dry and the FD arrived. The best FC was the slowest-selling FC. Every clean one you see now survived a decade in which nobody wanted it.