RX-Rated: Mazda’s Early Rotary Cars, Part 2


The big news for 1975 was the Mazda Cosmo AP, introduced in October after debuting at the Frankfurt Motor Show. A big four-seat coupe reviving the name of Toyo Kogyo’s first rotary production car, the Cosmo AP had some commonality with the Luce, sharing the same wheelbase and track, but in the best American tradition, the coupe was longer, lower, and wider. The new Cosmo’s styling had a distinct Detroit flavor as well, resembling both GM’s “Colonnade” intermediates and the 1974 AMC Matador Coupe.

The new Mazda Cosmo shared the Luce’s transmissions and engine lineup — the familiar OHC four in 1,796 cc (110 cu. in.) and later 1,970 cc (120 cu. in.) forms for the cheaper models, the 12A and the 13B for rotary cars — but suspension was more like the Capella’s, trading the Luce’s rear leaf springs for coils, trailing arms, and a Panhard rod. All Cosmos had four-wheel disc brakes and 14-inch (356mm) wheels. Japanese-market models also offered the choice of a JATCO automatic or a five-speed gearbox with a fluid coupling rather than a plate clutch, an arrangement shared with the JDM Mazda Luce, but not offered on U.S. export cars, which had a conventional five-speed instead. Upper-level Cosmos were lavishly equipped, in keeping with their lofty prices; the RE Limited started at ¥1,795,000, equivalent to more than $6,000.

1976 Mazda 121 front 3q copyright 2011 Alain Durand (used with permission)
Export Mazda Cosmos of this generation were usually badged Mazda 121 when fitted with piston engines and Mazda RX-5 with rotary power, but surprisingly, Mazda retained the Cosmo name for the American market. Non-U.S. Cosmos were 176.2 inches (4,475 mm) long — 179 inches (4,545 mm) with bumper overriders — on the same 98.8-inch (2,510mm) wheelbase as the Mazda Luce, weighing around 2,650 lb (1,200 kg) in 13B Limited form. Federalized Cosmos were 182 inches (4,623 mm) long and weighed about 2,825 lb (1,280 kg). They used the same 110 hp (82 kW) 13B engine as the U.S. RX-4. (Photo: “The second generation CD Cosmo appeared in 1975 and lasted until 1981. It was known as the Cosmo AP in Japan, and sold internationally as the Mazda RX-5, though in some export markets its piston-powered counterpart was called the Mazda 121.” © 2011 Alain Durand; used with permission)

If Toyo Kogyo had hoped for a piece of the lucrative American personal luxury car market, they were to be sorely disappointed. The federalized Mazda Cosmo arrived in early 1976 to lackluster reviews and general buyer disinterest. Extra weight made the Cosmo slower than the cheaper RX-4 coupe and the Cosmo’s Americanized looks did it no favors with U.S. import buyers; customers who wanted a Detroit-style luxury coupe could buy a Chevrolet Monte Carlo for less money. One New York dealer tried to drum up some interest by running a Cosmo in the 1976 24 Hours of Daytona, but Mazda’s U.S. organization was still struggling to interest buyers in any new Mazda, much less a $6,000 luxury coupe. We don’t have a breakdown of U.S. sales by model, but Mazda only sold around 35,000 cars in the U.S. in 1976 and we suspect very few were Cosmos.

The Mazda Cosmo met a similar fate in other export markets, but it turned out to be a big hit in Japan, which took the lion’s share of production: over 50,000 in the first year alone. In July 1977, the Cosmo even spawned a still more American-looking variation, the Cosmo L Landau, with a notchback roof and padded vinyl top.

1976 Mazda 121 rear 3q copyright 2011 Alain Durand (used with permission)
Although it didn’t seem to go over well with American buyers or critics, the Mazda Cosmo’s ‘glass pillar’ hardtop roof became its signature, retained on the subsequent HB Cosmo. While the chrome frames are fixed, the glass is retractable, an unusual touch. The interior is archetypal seventies luxury, with plush velour, woodgrain dashboard, and a genuine wood steering wheel. (Photo: “The CD Cosmo/RX-5 series was a flop internationally as Mazda tried too hard to ‘americanize’ the car.” © 2011 Alain Durand; used with permission)

The Cosmo was popular enough at home to make it profitable, but mediocre export sales made it less successful than anticipated and it was not quite enough to stave off the company’s mounting losses.


The end of the OPEC embargo had not brought much of a recovery for Mazda. Total sales for 1975 were down an additional 22%; the company declared a loss of ¥17.3 billion (roughly $60 million) for the fiscal year. By the end of the calendar year, Toyo Kogyo was close to bankruptcy.

That was not a happy prospect for the Hiroshima area, where Toyo Kogyo was responsible, directly or indirectly, for about 25% of the industrial jobs, or for the Sumitomo Bank, which was the company’s second-largest shareholder and creditor, with a total financial exposure of around ¥60 billion (more than $200 million). The bank tried to convince Toyota to purchase an equity stake in Toyo Kogyo in hopes of bolstering the company’s capital position, but when that didn’t pan out, Sumitomo Bank agreed to step in directly to keep Toyo Kogyo afloat.

1977 Mazda RX-3SP front 3q
The final U.S.-market Mazda RX-3 was offered only as a coupe. Introduced in March 1977, the RX-3SP was 40 lb (18 kg) lighter and $150 cheaper than the previous RX-3 coupe, bringing its base price below $4,000. Since 1976, the federalized 12A engine had used a two-barrel carburetor rather than the previous four-throat, but porting changes had improved output to 95 hp (71 kW) and 102 lb-ft (138 N-m) of torque. The RX-3SP was no longer available in California, although buyers in other states now benefited from a five-year, 75,000-mile (120,750-km) engine warranty. (author photo)

Kouhei Matsuda, who had succeeded his late father as president in 1971, has ended up shouldering much of the blame for the company’s near-collapse. Unlike his father, who had been a persuasive leader of considerable vision, Matsuda ruled through intimidation as much as respect, and many employees and executives were afraid of him. Toyo Kogyo had always been essentially an autocracy; there were many executives and executive committees, but most actual decision-making power resided with the president. Since no one was eager to give Matsuda bad news, much less argue with him, he was slow to recognize problems and slower still to respond even when business turned sour.

Even in the face of the company’s financial crisis, Matsuda resisted outside attempts to change the company’s direction. In January 1976, he reluctantly allowed Sumitomo Bank to appoint bank executive and troubleshooter Tsutomu Murai as Toyo Kogyo’s new executive vice president. Murai and managing director Hiroshi Mineoka, another Sumitomo appointment, enlisted the support of other Toyo Kogyo executives to develop and implement new procedures and strategies throughout the company. Murai also made some important appointments of his own, including making Hiroshima plant manager Yoshiki Yamasaki the new head of production.

1977 Mazda RX-3SP side
The Mazda RX-3SP’s rather gaudy spoilers, stripes, and decals were nominally optional, a $345 appearance package. It remained available in the U.S. through the 1978 model year, although Savanna production had ended in 1977. This car’s wheels are once again not stock. (author photo)

Perhaps the most unusual strategy Toyo Kogyo adopted during this period was the dispatched worker program, which put volunteers from throughout the company through a two-week sales training course and then shipped them off to dealerships around the country for temporary stints as car salesmen, making up for the recent losses to Mazda’s sales force. At its peak, the program involved almost 3,000 employees.

In December 1977, Matsuda agreed to relinquish the presidency to Yamasaki, ending the Matsuda family’s five-decade dominion. By the time of Matsuda’s departure, Toyo Kogyo had managed to reduce costs, increase sales and revenues, and improve relationships with its suppliers. Since the company union had agreed to accept delays or reductions in bonus payments as an alternative to salary cuts or layoffs, the cost in jobs and wages was relatively modest. The company reported a net profit of ¥15 billion (perhaps $65 million) for fiscal 1978 and did even better the next two years, due in part to Ford’s belated acquisition of an equity stake: 12.9% in 1979, rising to 25% in 1980. Only seven years earlier, Henry Ford II had declared Toyo Kogyo worthless except for the rotary. Now, Ford was more interested in Mazda’s piston-engined compact pickups than in the rotary engines.

1977 Mazda Luce Legato front 3q press - copyright Mazda
Mazda’s eighth new rotary of the decade was the Mazda Luce Legato, introduced at the Tokyo Auto Show in October 1977 as a replacement for both the Luce/RX-4 and the Roadpacer. The Legato was even bigger than the Cosmo — 182 inches (4,625 mm) long and 66.5 inches (1,690 mm) wide on a 102.8-inch (2,610mm) wheelbase — and was offered as a four-door sedan, a station wagon, or a four-door pillared hardtop. Like the Cosmo, the Legato was available with a choice of four-cylinder piston engines or the 12A and 13B rotaries. Piston-engined versions were sold in some other markets as the Mazda 929L, but the rotary cars were not exported. (Photo circa 1977, courtesy and copyright Mazda)

While the company’s late-seventies recovery had a lot to do with the new piston-engined Mazda Familia (a.k.a. Mazda GLC or Mazda 323, depending on market), introduced in January 1977, the financial crisis and reorganization did not mean the end of the rotary engine. Kenichi Yamamoto — by 1977 a director of the company as well as manager of rotary development — persuaded Sumitomo Bank of the rotary engine’s continued viability, promising even greater reductions in emissions and fuel consumption. With the bank’s support, the rotary would survive into the 1980s, albeit on a more limited basis. Yamamoto’s career also continued to advance: He became senior managing director in 1983, president in December 1984 (shortly after Toyo Kogyo officially became the Mazda Motor Corporation), and chairman from 1987 to 1991.


By the late seventies, Toyo Kogyo’s ambitions for an all-rotary line were fading. While the company mostly avoided canceling products outright — the main exceptions were the Mazda Roadpacer AP and the REPU, both dropped by 1978 — it gradually pruned its export offerings. As existing rotary models reached the end of their lives, they were not necessarily replaced. The new Familia was not offered with a rotary engine, nor was the next Capella, introduced in Japan in October 1978. The luxury-oriented Mazda Luce and Cosmo would continue to offer rotary power, but would no longer be exported in any numbers. Both models were dropped from the U.S. lineup at the end of the 1978 model year, vanishing from Australia and other export markets by mid-1979.

1984 Mazda HB Cosmo RE Turbo front 3q copyright 2007 天然ガス Tennen-Gas (CC BY-SA 3.0 Unported)
Introduced in 1981, the third-generation Mazda HB Cosmo was offered in two- or four-door hardtop form or as a four-door sedan, all sharing a lot of structural commonality with the Luce/929. At launch, the Cosmo was available only with a piston engine, but a new six-port 12A engine was added in 1982, followed by a 12A turbo and the latest version of the 13B. Blackout pillars and retractable headlights mark this 12A turbo coupe as a facelifted Series 2 model, introduced in 1984. (Photo: “Mazda HB Cosmo 001” © 2007 天然ガスTennen-Gas; resized 2011 by Aaron Severson and used under a Creative Commons Attribution-ShareAlike 3.0 Unported license)

The Mazda Savanna was another matter. Even with the energy crisis, the Savanna/RX-3 had been a solid success, selling 285,887 units worldwide, and its racing exploits had given the coupe a strong performance image in Japan. Rather than sacrifice that recognition, Toyo Kogyo applied the Savanna name to the company’s new rotary-powered two-seat sports car, the Mazda Savanna RX-7, which debuted in Japan in March 1978. In other markets, the new car would be called simply Mazda RX-7.

1985 Mazda RX-7 GS front 3q
The first-generation (SA/FB) Mazda Savanna RX-7 was sold from 1978 through 1985. This is a 1985 U.S.-spec RX-7 GS, powered by a carbureted 12A engine with 101 hp (75 kW). The pricier GSL-SE had a fuel-injected 13B with 135 hp (101 kW) as well as four-wheel disc brakes, a sunroof, cruise control, and air conditioning. (author photo)

Although the Savanna RX-7 would become Mazda’s best-selling rotary model for the rest of its 26-year lifespan, the company continued to offer rotary engines in the Japanese-market Mazda Luce and Cosmo into the 1990s. The 1982 Cosmo RE Turbo would be the world’s first turbocharged rotary production car, while the 1990 JC Cosmo, sold through Mazda’s Eunos sales channel, would be the first production car to offer a three-rotor engine.

Mazda would continue to develop rotary engine technology into the 21st century, adding catalytic converters, fuel injection, resonance supercharging, twin-scroll and sequential turbocharging, and even a hydrogen-fueled version. Mazda rotary engines also saw considerable competition use, culminating in the remarkable four-rotor 787B that won Le Mans in 1991.

1990 Eunos Cosmo front 3q press - copyright Mazda
The JC Cosmo, sold through Mazda’s Eunos sales channel from 1990 to 1996, offered either a turbocharged 13B engine or the twin-turbo 1,962 cc (120 cu. in.) 20B-REW three-rotor engine, conservatively rated at 280 PS (276 hp, 206 kW) and 297 N-m (402 N-m) of torque. Top-spec ECCS models were laden with luxury features, including a 12-disc CD changer, an in-dash television, GPS navigation, and touchscreen audio and climate controls. Despite its mandatory four-speed automatic and 3,600 lb (1,640 kg) curb weight, the Cosmo ECCS was capable of 0-62 mph (0-100 km/h) in 6.2 seconds, although top speed was electronically limited to 124 mph (200 km/h). Total production was 8,875 units, only a few of which ever left Japan. (Photo circa 1990; copyright and courtesy Mazda)

The company’s most recent rotary model, the Mazda RX-8, made a promising debut in 2003 with first-year production of 60,100 units, but declining sales and difficulty meeting Euro V emissions standards led to its withdrawal from European markets in 2009. It was dropped from the U.S. market at the end of the 2011 model year. Mazda announced in October 2011 that Japanese production would conclude in June 2012 with 1,000 copies of a final Spirit R limited edition.

In 2007, Mazda released information about its next-generation rotary engine, known as the 16X. Still a two-rotor engine, the Mazda 16X has a narrower rotor housing than the 13B or current RENESIS engine, but a greater eccentricity and epitrochoidal radius (comparable to reducing the bore and increasing the stroke of a piston engine) for a total geometric displacement of 1,598 cc (98 cu. in.). Despite the increased displacement, Mazda claimed that a lower surface-to-volume ratio and the adoption of direct injection would give the 16X fuel consumption comparable to the company’s four-cylinder 2000MZR engine, making the 16X at least 30% more efficient than the RENESIS. As of this writing, a shortage of development capital has forced Mazda to table the 16X project, but Mitsuo Hitomi, general manager of powertrain development, has maintained that the company has no intention of abandoning the rotary.


Add a Comment
  1. The US RX-2 was updated for model year 1974 with the new single distributor engine. The car received the new taillight assembly and highly styled dashboard, but retained old style front end bodywork to accommodate the new large 5 mph bumpers. It was visually very distinct from the 1973 and earlier cars. It is a rare car because the RX-3 was cheaper and newer, and the new RX-4 had the luxury upscale side of the rotary market well covered.

    The first generation RX-3 12A engine 4-door sedan is a very rare car. The twin distributor 12A RX-3 was mostly sold in the US, and very few sedans were sold (which is why it was dropped for 1974). RX-3 coupes and wagons were the big sellers.

    The RX-3 may not have been a substantially better race car than the RX-2, but when Mazda got serious about sedan racing in about 1973 and 1974, the RX-3 was early in its model run, and the RX-2 was later in its cycle. It would make sense to “market” through racing success the car that was likely to be offered for a longer period of time. The RX-3 was also marketed as a “sporty” car and the RX-2 was more of a luxury car, though the mechanical specs were similar.

    Late in the RX-3’s production life, the US only RX3-SP was introduced, actually an 808 with the rotary engine and certain specific trim changes and driveline upgrades. The optional stripe and trim kit was manufactured in California and installed at the port. This iteration of the RX-3 encouraged grassroots racing of the rotaries and helped keep the rotary alive as a “sporty” engine until the arrival of the RX-7.

    1. Thanks for the clarification on the ’74 RX-2. Because of its rarity, detailed information on it was rather scarce.

      The Savanna wasn’t really that much younger than the Capella/RX-2 — the difference was less than 18 months — so I don’t know how much that really had to do with it. The Savanna had the potential for a better power-to-weight ratio in racing trim, and more advantageous dimensions. Aside from being substantially shorter (a plus in some types of event), the Savanna was actually 15mm wider than the Capella, and its tread width was 15mm/10mm greater, which makes for less body roll, even with stock suspension. Of course, they did pitch the Savanna at a different audience than the Capella, particularly in Japan.

      The RX-3SP is mentioned in the text. It’s not exactly an 808; the 808 was the export designation for the Grand Familia, which was the piston-engined version of the Savanna. Although Mazda marketed them as separate entities (rather than as a single line with a choice of piston or rotary power), my understanding was that the differences between the Savanna/RX-3 and comparable Grand Familia/808 weren’t any greater than the differences between the Capella/616 and RX-2 — larger fuel tank, some minor changes to accommodate the rotary’s exhaust system and emissions controls, etc. Six of one, half dozen of the other.

      Since I don’t think the RX-3SP arrived until after Mazda had already announced that they would be launching a new sports car, I assume you’re right that it was basically an appetizer for the RX-7. I don’t have production figures for the U.S. model, unfortunately, but Mazda listed total Savanna/RX-3 production as 9,825 units for the 1976 calendar year (encompassing part of the ’76 and ’77 model year) and 1,606 in 1977 (encompassing the latter part of ’77 and the ’78 model year), so it can’t have been a lot — although it’s clear there were at least enough to homologate it for IMSA and SCCA competition!

      1. The RX-3 was always distinct from the 808 in its interior fittings and dashboard. The SP’s interior is an exact duplicate of that year’s 808. There is actually no “rotary” badging except for the “Rotary Engine” badge on the trunklid (and the “RX3SP” on the right rear fender). The RX-3 always had chrome window driprails and rocker panel trim. Later 808s did not and neither did the SP. The SP taillight assemblies were parts binned from the Australian 808. Beyond the VIN, a fake SP is easy to tell if it is made from an earlier RX-3, but hard to tell from a late 808, if the bolt-on parts are changed. The 808 derivation of the SP is not supported by documentation, but instead by careful observation of how the various year RX-3s and 808s were put together.

        Keep in mind too, that by mid 1975, Rx-3s were essentially special order vehicles, manufactured in small batches. 808s were still being turned out in the tens of thousands per year. Which leads me to production numbers.

        Worldwide RX-3 production was recorded by calendar year. But US production was by model year, starting the previous August or September. Also, only US cars were included in the S124 serial number sequence (12A engine) until late 1973. Other markets got the 10A engine and a separate S102 serial number sequence until late 1973. Overall roughly 80,000 non-US market 10A engined (S102A) coupes and sedans were built, 75,000 US market 12A engined wagons (S124W), and about 85,000 mostly US but worldwide 12A engined coupes and sedans (S124A) were built.

        Here is what makes the understanding of production numbers difficult. Most RX-3 US production for a given model year is actually produced in the last part of the prior calendar year. So the calendar 1976 production is largely 1977 model year cars for the US. The other thing to keep in mind is that from late 1973 through mid 1976, the 12A engined production sequence (S124A) is not just the US any more, but represents worldwide production.

        So US sales for the early twin distributor car through late 1973 were about 60,000 coupes & sedans (S124A) and 55,000 wagons (S124W). Worldwide, about 20,000 of the second generation coupes and sedans were sold in 1974 and 1975. The 1976 “Savanna” model (special taillights and the introduction of the ‘air dam’ front lower valance) totals about 2,000 cars worldwide. The SP (US only) totalled about 3,000 for 1977 and 800 for 1978. Most of the 1977s were built in late 1976 and the 1978s were built in the fall of 1977. All continued the S124A serial number sequence, out to just over car number 85,000 (VIN S124A-185000). So those last 25,000 cars or so were sold worldwide, not just in the US. That is why the later cars are so hard to find in the US or anywhere else.

        Just to confuse things further, a small batch of “V-100” cars were built for Japan. They are essentially right hand drive RX3SPs, right down to the big bumpers, but with alternate badges. But rather than using the S124A VINs in sequence, they pick up the numbers in the 460000 range. I do
        not recall the prefix used, but the last part of the number appears to be in sequence with the 808s of the time. Perhaps the US cars stayed in the RX-3 sequence to satisfy the US authorities. The V-100 VIN sequence also suggests the late rotaries were a batch pulled from the 808 production.

        1. Given the slowdown in sales, it would make sense that later Savannas and RX-3s would have greater commonality with the piston-engined cars. During that period, Toyo Kogyo was also trying to increase inter-model commonality in general, as part of the effort to reduce unit costs. Different dashboards are expensive, especially for what’s become a slow-selling specialty model.

          If I may ask, where do your production numbers come from? I wasn’t able to get U.S. production breakdowns from Mazda, and the only figures I have there are model year totals that aren’t divided by model. Mazda’s figures only show total worldwide calendar year production, and the numbers I have are only for rotary models. (I asked about piston-engined cars, but wasn’t able to get that information.) I realize there’s a big discrepancy between model year and calendar year production (which among other things makes it hard to tell how many might belong to each model year, or each generation in years where they overlap).

          Thanks again for the info!

          1. The RX-3 is the only older rotary in which the US model year sales can be roughly discerned. Every other model uses one VIN series worldwide, with no country by country breakdown. Because for most of its production life, the RX-3 used separate VIN sequences for the US (due to the larger engine size), the US sales can be estimated, and also by model year.

            I use Mazda’s old dealer manuals that include VIN sequences as part of the model/year ID process at the dealer level. I guesstimate the numbers from memory, but now I have the books in front of me.

            Total cars about 240,000. Early non-US 10A engines, about 80,000. Early 12A US 55,999 coupes/sedans plus 50,103 wagons. So the total early cars about 186,000 of the 240,000 total. Second generation (now worldwide) roughly 27,000 coupes/sedans plus 22,000 wagons. This figure includes the “Savanna” cars.

            Finally, roughly 4,000 US RX3SPs (more than people think there were) and maybe? 1,000 Japanese V-100s.

            Roughly 2/3 of the second generation 27,000 are probably US, but a maximum of 1,845 of the US cars are the model year 1976 “Savanna” style (but still badged as RX-3, not Savanna). That one is the rarest US car, and very hard to find here (The US version did not get the alternate dashboard package on that model). It is the one style more abundant abroad rather than in the US, but still not common anywhere.

          2. Okay, that makes sense. Serial number sequences aren’t always 100% accurate (sometimes numbers are skipped for various reasons, or used out of sequence, etc.), but it’s certainly a reasonable basis for estimation.

            The figure Mazda provided for total worldwide Savanna/RX-3 production is 285,887, starting in calendar year 1971 and ending in 1977. Unfortunately, they didn’t have any sort of model breakdown.

  2. The maintenance and repair of the rotary engine was carefully planned at the dealer level as the cars were introduced. It was assumed that the dealers would disassemble the engines to replace the “wear parts” and gaskets at specified intervals. Special disassembly tools and procedures were put into place.

    In fact, the engines had a multitude of issues, and a variety of unplanned catastrophic failures (partly due to improper or poor owner maintenance habits). The orderly disassembly of engines was cancelled and an engine exchange program was introduced. The old core engines would be rebuilt with new parts by Mazda and cycled back into the dealer pipeline.

    The US engine license with Curtiss-Wright did not allow for engines to be purchased separate from a whole car, so the engine exchange was the way things had to be done. One could not go to Mazda and purchase a whole engine, not even a competition engine (which could be purchased outright in Japan). Only parts could be bought and built up into an engine, or a rebuilt engine could be swapped into a car for the old one.

    On the engine exchange program, the car’s VIN had to be recorded by the dealer and turned in to Mazda with the core engine, in order to get a new one (again, the C-W license restrictions). While Mazda may or may not have made money on the engine exchanges, the dealers certainly made money, and competed for the business. Some enterprising dealers went to the junkyards and pulled whole engines, while carefully recording the junked vehicles’ VINs, so they could stockpile new engines ahead of time and offer quicker turnaround to the customers. Years later, 10A and early 12A engines could sometimes be found in dealers’ back rooms, still sealed in the boxes, and were let go for a song to clear the space of unwanted inventory.

    1. Interesting — thanks for the info!

  3. The Cosmo RX-5’s profile was echoed by the later 1977-1979 Ford Thunderbird, though the RX-5’s “opera” window actually rolls down, unlike the T-bird’s fixed glass.

  4. I was waiting with bated breath for the second part of this series.

    Have you seen the (in)famous Mazda "boinger" commercial?

    Neat bit of rotary-related ’70s camp.

    1. That is pretty funny — I knew they did commercials along those lines, but I hadn’t seen that one.

  5. The early RX-3s had a strange quirk that could be pretty frightening if you weren’t prepared for it. Suddenly let up on the gas after a long hard blast and the car would backfire apparently from the thermal reactor – it sounded like a 12-gauge shotgun going off under the car and one would expect to see pieces of the exhaust system left behind on the road. That wasn’t the case, but it was still pretty unnerving.

    1. A lot of the period reviews noted something similar. It appears that it was a side effect of the rich mixtures used to increase the reactor’s effectiveness, and that when you closed the throttle, it momentarily jumped from rich to [i]very[/i] rich, causing the backfiring when unburned fuel hit the hot thermal reactor.

  6. I had an RX3-SP and ruled the streets for awhile in my College Neighborhood of Northridge, California back around 1978. Loved how the Mazda magazine car ads claimed the ‘SP’ did not stand for ‘slow poke’. The specs claimed my RX3-SP could ‘stick’ a Boss 302 V8, and it really did. Before my ride got ‘totaled’ in a car wreck with a Lady from Texas, I was slowly converting my Mazda for weekend Amateur Racing. Had the car ‘dropped’ a bit and added high performance Koni Shocks. The engine was stock, with the big Rotary engine seal replaced, under some kind of Mazda ‘Re-call’ thing. She could easily hit 123mph on the 101 fwy when the fuzz weren’t around.

    1. I had one and am wondering if the speedometer reached 180 or 160?

  7. Just Picked up a Nice 1978 RX3SP from Puerto Rico and had it shipped to Florida, my first one was back in 1977 in EL Paso Texas, while serving int he US ARMY,I have had all the models from the RX2 to the 1993 Mazda RX7, Mazda Rotary Pickup but none of them come close to what you really feel for the RX3SP, its a nice toy to have and to pamper and take care of. THe sound of a Rotary Engine Bride Ported and overlap is music to any Rotary Fan.
    Would like to post picture but I guess on this site it cannot be done. But Thanks for the memories.

  8. Going to high school in 1980’s Oregon, rotary powered Mazdas were a common sight. Including mine, there were at least three rotary pickups in the school parking lot on any given day. My yellow 1975 rotary truck (along with a new Cosmo) was purchased by my uncle in ’77 or ’78 from a warehouse of unsold inventory in Japan and shipped back by the navy.

    It was an epicly fun little truck, but had some very odd quirks. They had a significantly stronger suspension system than other late ’70s small pickups and had no trouble reaching freeway speeds with very heavy loads. The small hatch in the side of the bed behind the passenger door in your picture is where the battery lives. That would corrode quite badly over time. The exhaust system was also one of the heaviest I’ve ever seen. In addition to the standard exhaust pipe with mufflers and such, there was a small straight pipe directly from the thermal reactor. I was told that was a pressure relief pipe to keep the muffler from exploding when the engine backfired.

  9. Thanks for the write up, I don’t think I’ve ever seen or heard of the FWD rotary Luce, or even a piston 1500cc 1500, only seen sedans.
    4dr sedans no matter what the make are more common than coupes/hard tops.
    I bought Mazdas for all my first cars, a 1974 RX3 sedan I bought in 1992, then a lot of 808s for spares and run abouts while working on the RX3 (changed the 12A and 3spd auto to a 13B extend port and 5spd manual out of an RX5), I think I owned about 4 808s, saving the best panels from them when I got them taken away for scrap.
    Of course in 1993 I also bought a 1973 RX4 hard top for another project car.

    Still have them, but stripped down in parts as unfinished projects these days unfortunately, but couldn’t bare to get rid of them.

    Lots of old RXs are still around in Australia and Puerto Rica. If you search places like youtube for ‘mazda rotary cruise’ you’ll see and hear a lot still running on the roads.

  10. Largest Australian Holden was the Statesman, Premier was a shorter wheelbase. FAB site.

    1. Oops, you’re quite right. Thanks for the correction — I’ve amended the text.

  11. Looking for a picture of the Mazda rotary 4 door sedan I owned in the 1980s bought used in Upstate New York. The interior was really plush. I loved this car but it began to rust. Cannot remember the model name. I do not see it listed in this article. Thanks. Paul

    1. Well, if it was a U.S.-market car and a four-door sedan, it was either a) a 1972–1974 RX-3 (a.k.a. Savanna), b) a 1971–1974 RX-2 (a.k.a. Capella), or c) a 1974–1978 RX-4 (a.k.a. Mazda Luce), which are described in the article (although naturally I may not have a picture of the specific version you had).

      Based on the timeframe and plushness, my guess would be the RX-4. The RX-4 (known as Luce in Japan and Mazda 929 in piston-engined export form) was the biggest and plushest rotary sedan Mazda sold in this country.

      If the car was RHD and not originally a U.S. model, it might also be a later Mazda Luce or the four-door hardtop version of the Cosmo (which is pictured in this article), which were never officially imported here, although a few have shown up as gray-market imports and there may be some in Canada. I’m assuming if you had a RHD car you would have mentioned it, though — generally, people who have RHD Japanese-market imports have a pretty good idea of what their car is! — so my guess would be the RX-4.

  12. My father had a 1978 RX-4 rotary. I used to ask for the keys to “go to a friends house”. In reality I was headed for hwy 100, the nearby cruise boulevard. Noticing that the rotary had a strong pull from around 3000 rpm and up I found a way to run with the muscle cars back then. The dark green 4 door was not quick off the line with an automatic transmission. I definitely had to avoid a drag race, but those were not as common as the 35-40 mph floor it till 60 mph or so bursts to see who could pull ahead by a fender length or more. That is where I realized that if the transmission was left in first gear the little rotary was in it’s power band. Many drivers were surprised by that little import and the strange screaming sound it made as it pulled ahead of them and then back firing when lifting off the gas. Running in first gear up and down the cruise strip on Fri. and Sat. nights eventually took it’s toll. After two summers of this kind of abuse I was able to leave a smoke screen under certain conditions. Interestingly my father though I was putting gas in the car as he kept business mileage records. Maybe fuel efficiency was better at higher rpms. It sure liked to rev. Power band like a two stroke dirt bike, not much initially but hang on when it starts to pull.

  13. I too had a 1978 Rx3-SP. $4500 new. 5speed stick. Got it up to 125mph a few times. Beat a few 350cu in V-8 vettes, and could hold the lead up to about 100mph. It was a rocket off the line, and light weight (in lbs) car. Vettes were heavy V8 iron engine dogs, even with Fiberglas bodies.
    Rarely ever lost a race in 5years in the suburbs of Chicago, to any car.

    Car rarely needed service. But you had to replace the plugs every 6mos. Due to oil by design was injected into rotary “cylinders” for lube.
    I cleaned them religiously 3mos, and replaced them every 6mos, with OEM Mazda plugs, which cost $3 each….2x normal plugs at the time.
    Rotary engines were awesome. I really thought they were the future, and not camshaft cylinders. Oh well, bring on the Electric cars!

    1. As I understand it, the plugs in a rotary engine inevitably lead a fairly hard life because they’re buried in the “combustion chamber” area of the housing. In a reciprocating engine, the tip of the plug is exposed to uncompressed intake air during each intake stroke, but that’s not the case with a rotary, where the mixture is already compressed (and hot) before it gets to the plug. This is why the housings of (as far as I know) all production rotaries are aluminum: the aluminum is more conductive to heat than iron, and without that, the housing ends up cracking due to thermal stress. (This happened to various test engines using cast iron housings, which in certain other ways would have been preferable.)

  14. hello how many us 1973 rx3 sedans were made cause I have one

    1. I’m afraid I’ve been unable to find sales breakdowns by model or body style — sorry! I wish I knew.

    2. A U.S. model 12A engine sedan is more rare, and a rest-of-world 10A engine sedan is more common. The U.S. version was the only one with the 12A engine in 1973 (twin distributors), and while the rest of the world bought quite a few RX3 sedans, the U.S. buyers preferred coupes and wagons. As all true RX3s have become quite rare these days, rest assured that the car is not common any more, in any case. Basically all parts of the car are fragile (paint, interiors, sheet metal, trim, engines), and do not hold up well to abuse, poor maintenance habits, or exposure to the outside elements.

      I believe 1973 was the last year for RX3 sedans in the U.S., but I could be wrong about that.

      1. On the last point, as best I can determine that is correct, but I’m not positive either.

  15. From 1974-1976, I worked for Tilton Mazda in Sioux Falls, SD. At that time I bought a new 1975 RX-3. I found the window sticker about 3 months ago. WOW! Prices were: dealer prep $50.00, Freight $85.00. The AM radio set me back $69.00. MSRP for a Flare Yellow 2-Dr was $3697.00. Total was $3901.00. It listed the Engine No.as 12AS57958. The sticker was printed 5/19/75 and I bought it in July 1975. As I recall I paid $3500.00. This car was a blast to drive. It came with white vinyl interior, did they offer black? That engine was so smooth. I loved the buzzer to let you know when to back off.

    1. Many of the RX-3s I’ve seen have had black vinyl interiors, so I’m guessing the answer to your question is yes, although white might be slightly less painful on a hot summer afternoon!

      1. RX-3s in the U.S, were offered with black or white interiors through model year 1975, but always with black dashboards and carpets on the white interiors. Tan was added as an interior choice for model year 1976 (the “Savanna” trimmed RX-3). Black was the only interior choice for the model year 1977 and 1978 RX3-SP in the U.S.

  16. Back in around 1978 I was in Taupo, New Zealand. One day I heard a banshee wail, clearly a Rotary in the distance. It took quite some time to come into view. Eventually there was a Mazda Ute, resplendant in the company blue and white livery, with a pretty radical exhaust. I did not know that rotary utes were ever issued, this being s right hand drive one was probsbly concerted from a B1600. This dealer’s workshop vehicle had a VERY overloaded B1600 on the other end of a tow rope. With little torque and heaps of revs in low gear the whole cavalcade crawled into view with a cacophany that all Taupo heard. It was a ROTARY!

  17. When the RX2 was released here in Oz, guys I knew would go to the dealer for a test drive, and see if the chaperone who rode with them knew what the over-rev buzzer meant! I also used to go to sedan car races when Alan Moffat’s RX7 was running – boy did it have a sharp barking exhaust note! Especially when being scrutineered under a tin roof!


  18. A fellow student bought a new R100 when they first went on sale. He marveled at the power and how smoothly it was delivered. One day he happened to return to his closed garage after a normal drive and discovered a glow under the car. I think it was caused by the muffler being red-hot.

    1. It was probably the thermal reactor, which was a common issue in the ’70s. (Early catalytic converters had the same issue; some ’70s cars had “catalyst” heat warning lights for that reason!) The thermal reactor was essentially an afterburner, used to reduce hydrocarbon emissions at the tailpipe. This naturally generated a lot of waste heat as well as increasing fuel consumption because the exhaust leaving the manifold had to be fairly HC-rich for the catalyst to work effectively. Fortunately, more precise mixture control and three-way catalysts largely eliminated the problem by the mid-eighties!

  19. Thanks to all contributors for such informative and interesting reading. I have loved rotary Mazdas for many years. This love has recently been revived with my purchase of a lovely RX3 coupe in phoenix blue. A 1971 build with cherry taillights, and according to the seller (an old Mazda dealer in Belgium), the very first LHD world market RX3 ever produced. The chassis number is S102A130000. I even have the original factory windscreen printout tag if anyone can translate japanese. Unfortunately the 10a died some time ago so originality is somewhat lost with a 12a but maybe I will stumble over an 0866 10a one day.

  20. Hi there, any info on the South African assembled Cappella rs? Very hard to find info on!
    Am looking at one trying to confirm originality. It has the single dizzy 12a, but has a spare wheel well which is odd. It also has the s series chassis rather than c.

    Any help appreciated.

    1. I’m afraid this is the first I’ve heard of it — I didn’t run across it in my original research. If someone has more insight on it, I’d be curious to know as well!

  21. Apparently the Mazda Chantez Kei Car was planned to use a 360cc 3A single-rotor Wankel engine until other Kei Car manufacturers protested (even if their concerns in retrospect would be unfounded).

    Would have been interesting seeing such a car reach production and Mazda in turn remaining in the Kei Car segment instead of abandoning it for over a decade, the closest comparison would have to be the NSU Spider and even then Mazda Chantez is both smaller and lighter.

    1. Well, the concern a 3A kei or keitora would have presented, aside from fuel economy, was that a 360cc rotary engine was not on the same footing as a 360cc reciprocating engine, which is troublesome in a market segment defined by strict displacement limits. Japan ended up settling on a 1.5:1 formula, so the 3A would have been deemed 540cc, out of line for the strict requirements of the class.

      The kei market has gone through some significant doldrums over the years. Part of the reason a lot of manufacturers backed away from it for a while was that as the Japanese consumer economy took off, kei cars and sub-1,000cc cars like the original Publica turned out to be a tougher sell than either their makers or MITI assumed. There was too much stigma attached. A Toyota Corolla or Datsun Sunny wasn’t that expensive to run, by comparison, and didn’t scream “poor person’s car.”

      1. The 3A would have still been eligible for the 550cc Kei Car era from 1976-1990 had Mazda decided to remain in the segment and slightly delay the introduction of the engine or develop an earlier 240cc unit.

        Is it known what the output was for the 3A single-rotor engine or any other figures exist? Cannot make out the info for the engine at the online Mazda Museum.

        It would have been interesting to see how the 3A single-rotor unit would have evolved during the 550cc and 660cc Kei Car eras prior to the max 64 hp limit (or whether it would have any influence on increasing the limit to around 75-80 hp), along with how Mazda would respond to other Kei car manufacturers using turbos and superchargers.

        Envision rotary Kei versions of the Autozam AZ-1 and Mazda-rized version of the Suzuki Cappuccino.

        1. I unfortunately can’t find a photo of the museum display engine that makes the sign text readable (the museum’s Japanese website doesn’t appear to have pages for specific exhibits), and I can’t find any indication of whether Mazda quoted an output for it. It would of course have been constrained by the contemporary kei class limits.

          1. Understand.

            Have to wonder though how long hypothetically speaking Mazda would have been able to continue producing Wankel-engined Kei Cars, especially given the inevitable trend towards front-wheel drive (with the larger Familia becoming FWD from 1980).

            That is unless Mazda cannily decided to develop earlier in-house Wankel-engined Kei Car-sized sportscars akin to the larger MX-5 years before rivals in order to retain the rear-wheel-drive layout

          2. Well, rotary engines are certainly not incompatible with FWD — the R130 Luce Rotary Coupé was not what I’d call an especially modern FWD layout, but it does illustrate the point. The compactness of the rotary engine is helpful in that regard. If anything, the packaging challenge would be that if you really take advantage of that compactness in designing the rest of the powertrain and the space allotted for it, you could end up with a car that can’t easily take a conventional I-3 or I-4 engine.

          3. It is probably why the rotary would be reserved for Mazda’s kei sportscars despite the likely 440cc displace (equating to 660cc via the 1.5 formula), though it is possible they would look for ways to make up for the displacement shortfall (and other limitations) compared to conventional engines such as turbos, superchargers or hybrid.

            While Mazda re-entered the Kei Car segment with a rebadged Suzuki, did they explore developing any 3-cylinder engine projects either an all new design or based on the Mazda B engine? Am assuming they would have eventually developed some other alternative Kei Car 3/4-cylinder engine had Mazda remained in the segment.

  22. You have overlooked the first Bertone designed generation of the Luce – the 1963 Luce 1000. So all your Luce generations are off by one.
    Otherwise excellent write up.

    1. The 1963 Mazda Luce was designed by Bertone, but it was a prototype, not a production car. (Its evolution is discussed at length in the 2015 Curbside Classic article “Automotive History: Mazda, Bertone And The Alfa That Wasn’t” — in the interests of full disclosure, the author did consult with me (albeit in a minor way) about it prior to its publication, as is noted at the end of the text.) Mazda describes the 1972 Luce as the second generation and the 1977 iteration as the third generation, so they don’t regard the prototype as a distinct generation.

      1. Thanks for clarifying.

  23. A detail on the 1976 RX-3 engine specs. The RX-2 and RX-3 12A engines had always used a four-barrel
    Nikki carburetor. The carburetor primaries were always inboard toward the center of the engine, looking left to right from the front of the car, and fed the center (looking fore to aft at the side of the engine) side housing intake ports, giving very short intake runners for the primaries, and longer intake runners for the secondaries. Intake runner length turns out to be critical for rotaries; moreso than for the typical piston engine. The later Le Mans Mazdas actually used variable intake runner lengths that would adjust during the course of the RPM curve, to maximize power over the RPM range.

    Back to the RX-3. For 1976, the RX-3 went from a Nikki carburetor to a Hitachi carburetor, a slightly smaller spec from the Hitachi carburetor used every year on the 13B engined vehicles, the RX-4, Cosmo, and REPU. The manifold runners were also reversed, with the secondaries feeding the center ports of the engine, and the primaries lengthened, to feed the outboard ports. This served to change up the runner lengths, by lengthening the primary intake runners and shortening the secondary runners. The 13B engines of 1976 shared this change, which required distinctive intake manifold castings. This was a one-year-only spec, that reverted to the old way of doing things for the 1977 year cars. If I recall correctly, the unique 12A Hitachi carburetor was also a four-barrel, not a two barrel. A four-barrel carburetor allowed the secondary venturies and ports to remain unused at lower engine RPMs, which would save fuel, theoretically at least, over a two-barrel arrangement. In any case, updating or backdating 1976 year intake parts to other year cars is a challenge.

    1976 also saw the introduction of “tall and skinny” intake runners, which served to improve fuel mileage, in both the intake manifold and the engine side housings. For 1977, the 12A RX-3 reverted back to the old Nikki carburetor and “short and fat” runners in the intake manifold feeding the center engine ports from the primaries (reintroduction of basically the 1974-75 manifold). So there was a real mismatch between the “short and fat” intake runners in the manifold, and “tall and skinny” intake runners in the engine, at the face where the manifold bolts to the side of the engine. Not ideal. This mismatch carried over into the early run of the RX-7.

    In the face of rapidly declining sales of cars with rotary engines, and renewed emphasis on piston engined cars (the new 323/GLC on the way, along with the next gen 626), it suggests a number of things going on at Mazda in the mid-to-late ‘70s. Certainly the company correctly understood that the rotary would appeal in a light sports car (the planned RX-7). But it also suggests that capital for ongoing rotary engine improvements was available in the mid ‘70s, funding elaborate but trivial-in-the-real-world manifold and carburetor modifications, but absent later in the decade, when “parts binning” from available engine parts on the shelf seemed to be the way things were done. The other question is why Mazda stayed with the rotary option in the various car lines, despite few and declining rotary engined car sales, and corporate image issues emanating from rotary difficulties of various sorts. Observing from the sidelines, perhaps Mazda was simply buying time to get the RX-7 out there, but maybe Mazda was also so bound up in the rotary engine, that the company, as a corporate culture matter, simply couldn’t let go of the rotary. Perhaps it took a substitute corporate automotive icon, the MX-5/Miata, to be established before Mazda could let the RX-7 go, and not have a rotary engined car always available in the product line every year.

    1. I wouldn’t assume that the real world impact of the various engine changes was trivial from a corporate standpoint. That period was one in which progressively more stringent emissions regulations were being phased in, in Japan as well as in the U.S. (the big years for JDM emissions standards were 1976 and 1978), so a lot of powertrain changes were the product of automakers’ struggle to balance emissions compliance with driveability, performance, and fuel economy. Because of the regulatory phase-in, many automakers ended up fielding some weird one-off interim variations, some lasting only a year or so. Making sense of these after the fact isn’t always easy, but the likeliest explanation in many cases is that the engineers were working on some more considered solution that just wasn’t going to be ready in time. The interim solutions may have been kludges, but they worked that worked well enough to pass regulatory muster and could be rolled out quickly enough to buy the manufacturer another year or two to work on the next, more stringent set of emissions standards.

      It’s also important to keep in mind the lag time between when things are developed and when they actually come to market. As the article discusses at some length, Toyo Kogyo did face a significant financial crisis that really became apparent throughout 1975. By that time, a lot of the work on mid-seventies products had already been done, for better or worse, but the cost-cutting that resulted from the crisis and internal reorganization probably had a lot to do with the mix-and-match approach you describe on subsequent models.

      As for why they kept at it with the rotary, the strong links between the rotary and Mazda’s corporate image was undoubtedly part of it (even decades later, Mazda was very reluctant to give up on the 16X project for that reason), but I don’t think that was the only reason. In the latter half of the seventies, Toyo Kogyo belatedly recognized that it didn’t make sense to stick the rotary engine in product they made, but they obviously concluded that it was still worthwhile for high-end models as an alternative to a six-cylinder piston engine, which Mazda didn’t then have.

      From a sales standpoint, one could certainly argue that the later iterations of the rotary Luce and Cosmo were a waste of time. However, since Mazda did not have a six-cylinder piston engine until the latter part of the eighties, abandoning the rotary engine or limiting it only to the Savanna RX-7 would have forced an unpleasant choice between rushing to develop an all-new six, which would also have been costly, or simply abandoning any presence in the high-end prestige segments, which would have been a pretty significant blow to their brand image. Mazda did not, so far as I’ve ever seen, ever have a particularly big slice of those segments in Japan, but a small player is still a player, and it spared company executives and shareholders the embarrassment of having to drive a Capella when their rivals drove Crowns or Soarers. Also, as the success of the Celica XX and Soarer demonstrated, the personal luxury segment had at least the potential to be quite profitable, and throwing away that potential probably didn’t seem prudent. (The eventual JC Cosmo was a spectacular effort to beat the Soarer at its own game, and arguably a successful one at least from a technological standpoint, although the collapse of the market for personal luxury coupes made it at best a Pyrrhic victory.)

      In the nineties, after the MX-5 debuted, the situation had changed dramatically: The coupe market was imploding, Mazda now had V-6 engines (including its very sweet small-displacement versions of the K-series V-6, which we saw only in the MX-3 GS, but were more widely used in JDM models and in some other export markets), and exchange rate issues were picking off the stragglers in the GT market, which made the RX-7 seem like it was being kept alive more out of corporate pride than good sense. (It pains me to say that, because I am extremely fond of the FD RX-7, but it’s hard to argue; it was finicky, high-strung, and way too expensive for the demographic to which it most strongly appealed, and its annual sales figures were miniscule.) However, that wasn’t true or even foreseeable 15 years earlier, so I don’t think Toyo Kogyo was unreasonable in deciding to play out the string.

  24. I’m having my 13b being rebulit now and have a 12a misson hooked up to it, question is the misson is a short vision of the RX 7 non turbo 5 speed misson. It is going back into my Datsun 1200 1972, dose anyone know of a short vision (lenght) 12a misson, not sure what it came out of but I know it’s old, maybe out of a 10a????? I picked up two rx7 1980 misson out of a junk yard and I was shocked to find out the mision in my car was almost 5″ shorter.

Leave a Reply

Your email address will not be published. Required fields are marked *

Comments may be moderated. Submitting a comment signifies your acceptance of our Comment Policy — please read it first! You must be at least 18 to comment. PLEASE DON'T SUBMIT COPYRIGHTED CONTENT YOU AREN'T AUTHORIZED TO USE!