Toyota 18RG Engine

Many miles were accumulated on the transplanted engine.

The four-cylinder, double overhead camshaft 18RG engine is of interest because of the valve gear, derived from work on the Toyota 2000GT sports car. Only a handful of these cars were built. In January 1978, the 18RG was offered in European versions of the Celica Liftback and continued in production for about three years.

The 2000GT was also the genesis of the XS650--Yamaha’s first four-stroke motorcycle engine. Writing in the May 1997 issue Classic Bike magazine, Yamaha’s Ludy Beumer recounted that company engineers borrowed the basic bore and stroke (75 x 75-mm) dimensions from the automotive engine, but shortened the stroke 1 mm to give 654 cc from two cylinders. The same pistons were used and adapted to air cooling by casting them in Lo-Ex. Valve geometry was identical to the 2000GT, except that a single, rather than double, overhead camshaft was specified for the motorcycle. Valve keepers carried the same part number. Production commenced in late 1969 and continued until 1980, for a total of nearly 300,000 units.

This particular engine, an 18RGU, was found in a Japanese junkyard with about 30,000 miles on it. Crankshaft bearings and gaskets were replaced with a combination of locally supplied 18RC parts and 18RG items obtained through Toyota in England. The dual Weber DCOE look-alike carburetors had self-destructed and were replaced with a Corolla manifold and single Weber. The intake restricts power, but one can hang in fifth gear at 25 mph, floating like a DC-3 on landing approach.

Engine Identification

All engines in the family displace 1990 cc with an 89-mm bore and an 80-mm stroke. The suffix "U" indicates smog gear, which was far more complex than used on American engines of the same period, and "E" stands for electronic fuel injection. If possible, avoid the 18RGU, with its throwaway carburetors and arcane smog hardware.

  • 18RG - dohc, 9.7:1 compression ratio, Toyota cylinder head, no smog hardware and 140 hp. Very rare and desirable.

  • 18RGR – dohc, 9.2:1 compression, Toyota head, air injection, and EGR. Rated at 130 hp.

  • 18RGRU – dohc, Yamaha head, 9.2:1 compression, complex smog gear. 125 hp.

  • 18RGU – dohc, Yamaha head, 9.2:1 compression, non-rebuildable carburetors, smog hardware and a throwaway electronic ignition. This 120-hp engine appears to be most frequently encountered.

  • 18RGEU - dohc, Yamaha head, 8.7:1 compression and electronic fuel injection. Said to develop 120 hp.


Rings, rods, crankshaft, main and rod bearings, front and rear oil seals, lower timing sprockets, small timing chain, and most gaskets are 18R/RC items, and still available from U.S. parts houses. Pistons, upper timing chain, chain tensioners and guides, and head gasket are 18RG items. You can order these parts from Europe via the Internet or from Toyosport, a California outfit that can be reached through Toyota Racing Development (213 532-1232).


Your local Toyota dealer may be able to order the "18RC Engine Repair Manual" PN 98196E, which covers the mechanical aspects of the 18R, 18RC and 18RG. Very little is said about the emission controls for these engines, which were designed to met more rigorous standards than were then current in the U.S. That's too bad, since components like the computer-controlled PCV valve merit discussion.

Out-of-print Haynes Celica manuals also deal with the engine.

Problem Areas

The 18RG establishes valve clearance with "buttons" of various thickness, on the model of Offenhauser and other high performance engines. A good automotive machinist can set valve clearances by grinding the seats and valve ends, but this work requires skill and experience. On the basis of the work they did for me, I can recommend Dover Cylinder Heads (2929 Calhoun Ave., Chattanooga, TN 37407,1-800-782-0541).

Intake and exhaust cams, marked "1" and "2," should be set with the slots in each cam at the 12 o’clock position and dead parallel to each other. At one time, Toyota Racing Developoment supplied the special tool used to establish this relationship, but eyeballing is accurate enough to get the engine running. I have not been able to discover the OEM valve overlap and duration specifications, which would make a significant contribution to power.

Ways to deal with the disintegrating carburetors used on the 18RGU are discussed below. There was also a problem with the distributor -- it was designed for magnetic coil pickup and the amplifier did not come with the engine. I substituted a 20R point set and condenser, using a 20R mounting plate that fit without modification. Distributor cap, rotor and point set also came from a 20R. An aftermarket coil used with power coming from an MSD-6 ignition module mounted inside the cab. This setup performed well and compensated for mixture discontinuities associated with the cobbled-up manifold.

The Corolla manifold was water-heated and plumbed in series with the heater inlet. (Connecting the manifold in parallel failed to produce enough flow.) The coolant first passed through the heater valve, then to the manifold and heater. Closing the valve cut off manifold heat for summer driving.

Besides being difficult to work with, the Celica rod-and-ball throttle linkage lost motion at every joint. I purchased a cable from Toyota Racing Development and, many miles later, replaced the inner element with plastic-coated, .063 in. diameter woven wireline purchased from farm equipment dealer. A brass bushing was machined to take up the slop at the throttle bellcrank.

The K&N lookalike air filter behaved like a blind pipe when tested on a homemade flowbench. After pricing a real K&N, I fabricated a filter from .080 in. aluminum plate and a paper element intended for large Briggs & Stratton engines. Because of space limitations, it was necessary to include a second filter, borrowed (I think) from Tecumseh, for the PCV system. Flowbench tests showed that the new filters had almost no restriction, and engine power increased. The setup was, however, noisy.

The original mechanical pump failed within a few miles of startup. An 18R mechanical pump could have been used, but I opted for a "no-name" diaphragm-type electric pump powered from the starter solenoid and through a two-terminal oil pressure sender. This pump could be disassembled – which was fortunate because the valves tended to stick – and it did have a means of output pressure adjustment. Five years or so later, a sealed rotary pump was substituted and run with a pressure regulator to reduce output pressure to three psi.