The only readily available pieces that I could find were a 4 into 2 into 1 header (tri-y), a cone type K & N filter and adapter, and an upgraded e-prom. However, since starting my original research, a few more pieces have come onto the market such as alloy pulley sets and a reworked throttle body. Still these are really not enough the stir the soul, and will realistically only give about 12-15 h.p. over stock with all the components in place. Most of these bolt-ons only increase the peak horsepower and affect torque and low-rpm power minimally.
I decided to tackle this project in a logical manner. The engine in question needed to be rebuilt anyway, as it had been autocrossed, time-trialed and basically thrashed on since it was purchased new, and had accumulated over 120,000 miles on it. Its' future use included additional time-trialing, hill climbs, and a future in PRO-Rally. But unlike most competition cars it would also be retained as daily transportation, so driveability and reliability were paramount, and relying on racing fuel on a daily basis was out of the question.
One of the first steps was determining if the engine had any inherent weaknesses. Other than an imbalance in the intake-to-exhaust flow proportion of the cylinder head (more on this later), the only major shortcoming was a tendency for connecting rods to seek the atmosphere. Numerous friends of mine in the machine shop and repair businesses had commented on seeing GA16's with holes in the block and broken connecting rods. Further investigations revealed the connecting rod bolts as the probable culprit. These 8mm bolts have a torque spec of only 17 to 21 lb. ft. of torque yet are the most highly stressed bolts in the engine. This most likely explains the GA16's low factory-set rev limit (compared to its big brother the SR2O). Knowing that more horsepower was in the cards, as well as eventually more rpm's, I needed to upgrade the standard connecting rods and fasteners without breaking the bank.
The original connecting rods were removed and then magnaflux-inspected for cracks. I proceeded to polish the beams of the connecting rods and slightly lighten the ends as well, then I sent them out to be shotpeened with S-230 shot at an intensity of .012-.015A. This is a slightly higher intensity mil-spec shotpeening operation than is usually used for automotive parts, but it costs the same. Most automotive parts are shotpeened at an intensity of .006-.012A.
A call to ARP confirmed than they had never produced a rod bolt for this engine, but I sent a sample and they came up with a near match. The shank diameter was perfect and the overall length was .080" longer but it would work fine. The head of the bolt did not fit the Nissan rod but was easily machined to fit. The torque spec of the ARP bolt is 30 lb. ft., so it is at least 50 percent stronger in clamping force. I'm sure that the actual metal used is of a higher quality as well. The connecting rods were sent to a machinist and the big-ends were resized with the higher strength bolts torqued in place. Finally, I balanced the rods end-to-end to within a gram of each other.
To cap off the connecting rods, I had Venolia make a custom set of higher compression, oversized pistons. Some people may wonder why I had Venolia make the pistons and not J.E. or Wiseco, who make tight tolerance pistons suitable for street or race. To be honest with you, I had ordered over 100 pistons in the past from J.E. for other engines, and they never really quite got them right. They also proved to be difficult to deal with, since their attitude was basically "It's only for a 4-cylinder." I tried to contact Wiseco, but they never returned one of my seven phone calls to them. Enough said. I had a good working relationship with Venolia already, and although they had never made a 0A16 piston before, they said "Sure, just send an original piston, and a combustion chamber pattern." I did that and gave them the new compression ratio I wanted, and 4 weeks later had a set in my hands. They specify a piston clearance of .005"- .007", but I've usually installed them with .0045" clearance for the street, so that was what the machinist bored the block to.
Incidentally, although the original compression ratio was rated at 9.5:1, the true calculated ratio turned out to be 9.59:1. Most automotive companies are a little optimistic, but Nissan appears to be a little pessimistic in its calculations. I settled on an increase to 10.5:1 compression and a bore increase of .040", as the cylinder walls didn't look thick enough to accept a larger bore increase. Displacement increased slightly from 1597 cc's to 1639 cc's.
For good measure, I had Airborne Coatings apply their thermal barrier coating to the piston tops and their dry film lubricant to the skirts. Moly-coating top piston rings, Total Seal gapless secondary rings and stainless steel oil scrapers were installed along with the original Nissan piston pins.
In an effort to improve throttle response and engine acceleration, the crankshaft was sent out to have .750" removed from the counterweights. It was also turned undersize .010" and rebalanced. In all, 3.25 lbs. came off the crankshaft. It may not seem like a lot, but it's roughly 11.5 percent of the crankshaft's original weight that was removed. Before the shortblock was assembled, the block was internally deburred and painted with Glyptal, and the top was decked to ensure trueness. The standard oil pan was installed with only one modification. A 12-guage piece of steel was welded onto the bottom to act as a skid-plate for the expected roughness of PRO-Rally usage. At the back of the crankshaft, the standard flywheel was fitted along with a 30 percent stiffer pressure plate and a sprung-center clutch disc with eight semi-metallic pucks on each face. I may have the standard flywheel lightened at a later date, or even may install a 10,000 RPM racing clutch on an automatic flexplate for an ultra-light assembly. But street driveability would be jeopardized, so it's still a toss-up as to what the owner wants to do.
With the shortblock done, work was concentrated on the cylinder head. I flow-tested the standard cylinder head to establish a baseline. The intake port of the standard head flows 102 cfm and the exhaust port flows 67 cfm. (measured @ 10" of water and .350" valve lift). This displays a typical Nissan shortcoming in the exhaust flow, as the exhaust-to-intake proportion is only 66 percent, while I like to see somewhere in the region of 75 to 80 percent for optimal power. Since oversized valves were not in the budget, and maintaining proper airflow velocity is crucial to street use, the cylinder head could not be radically ported. Instead of just enlarging the ports, they were only mildly enlarged and reshaped, with emphasis on the exhaust port. The end result was 115.5 cfm on the intake port and 83.5 cfm on the exhaust port. The proportion was better, at 72 percent, but not as good as hoped for.
Future development will probably include oversized valves, but for now I saw 13 percent more intake flow and 25 percent more exhaust flow than standard, while maintaining excellent velocities. In fact, this head flows nearly identically to a stock SR20 head but with much smaller ports. The head received a three angle valve job and a light resurfacing, then was assembled with the valve springs shimmed up slightly to increase tension.
With the cylinder head finished, the camshafts could be reground to match the cylinder head's airflow characteristics. The stock intake camshaft measured out to .342" valve lift and 208 degrees of duration @ .050" cam lift. The exhaust camshaft measured out at .312" valve lift and 196 degrees of duration @ .050" cam lift. It was bad enough that the head flowed badly out the exhaust port, but Nissan saw fit to make the exhaust cam smaller too! Keeping in thought the intended use of the car, the cams were mildly reground to identical intake and exhaust profiles.
The new valve lift is .354" and the duration is 212 degrees @ .050" cam lift. In addition, the intake cam was "offset ground" two degrees to tighten up the static lobe separation a bit. One thing to keep in mind here is that Nissan's variable intake cam timing on the GA16DE constantly changes the lobe separation angle dependent on throttle loading and rpm, so typical estimates about the proper camshaft for a given engine combination are confused by this technology. Still, I thought it was better to start small and work up, as nobody could tell me from experience what the engine "liked" from its camshafts. My own time schedule prevented me from doing the actual engine assembly. The shortblock assembly and the time-intensive shimming of the lash caps was handled by Richard of DeCormier Nissan in Manchester CT. Thanks to him for helping to pull this project off.
The now-completed engine was installed and the accessories began to be put in place. The stock intake manifold was installed but with SR2ODE injectors in place. The car's owner had years ago had a custom four-into-one header made for the car (before there were any aftermarket ones), and this was installed as well. A new four-into-one header is in the works at this time, because the one currently in use has primary tubes that are about 18 inches long, which would concentrate the power peak at a somewhat unrealistic 9000 rpm or so. At least the tubes are only 1 1/2" in diameter though. The one being built will use 34" long primary tubes of 1 1/2" diameter tubing and a 2 1/4" collector. If it tests out well, it will probably go to market. I'll have an update as soon as it's up and running. The exhaust that I had originally built for the car was retained as well. It has a 2" diameter mandrel bent downpipe leading to a single 2 1/2" Simons muffler that I imported from Sweden. This produced more power and very noticeably more torque than the previous Supertrap and 2 1/4" exhaust that the owner had on the car and sounds better as well. Small displacement engines such as this are very sensitive to exhaust pipe diameter. Bigger is not always better. The Stillen adapter and cone filter previously employed was reused, and Magnecor plug wires were installed. The ignition timing was set at 15 degrees B.T.D.C., which is up 5 degrees from stock.
The good news is, the car started up immediately and seemed to run well. After some break-in miles it was put through its paces at the Mount Washington hillclimb this year. It had a lot more grunt than the stock engine for sure, but seat of the pants experience tells me it was flattening out, and Tim, the driver, said it was worse at the higher elevations. A plug inspection showed a rich air/fuel mixture than would get worse at high rpm's when the pressure regulator and the higher altitudes (less oxygen) came into play. Still, the surprising thing was the torque and the smoothness of the engine, even though it has a lightened crankshaft and some hotter cams. It idles like a stocker at 900 rpm's, and the throttle response is excellent. In retrospect, I wish I had chosen more radical cam profiles still, because now it feels downright tame. That will come sometime down the road, as I'm working on locating some blank cam cores.
Of course, part of the performance felt at this point can be attributed to the 4.47 gear ratio installed in place of the stock 3.89 ring and pinion. And the custom limited slip installed helps to put everything down to the ground.
Tim tried installing the original GA16 injectors, but surprisingly this made it run worse. So the SR20 injectors went back in and I had a little brainstorm just then. Why not install the SR20 throttle body? We had a spare GA16 intake hanging around, so I removed the upper half and bolted an SR20 throttle body onto it (same bolt pattern). I scribed the larger throttle bore into the intake, removed the throttle body, then ported the GA16's intake to match. Then, I modified the throttle arm to accept the GA16's throttle cable, and removed the SR20's throttle position sensor and fitted the GA16's in the reversed position so it would fit the GA16's harness. The lever arm that actuates the T.P.S. was also reversed. A vacuum fitting had to be added as well by drilling a hole and epoxying a brass tube in place and then it was done. The SR20 connector duct between the throttle body and the airflow meter was used with a rubber sleeve used on the GA16's airflow meter to match the I.D. of the connector duct.
This worked better than expected. The throttle response was improved even further, and the sound it made held promise for more power. Sure enough, on the road it was dramatically better and didn't flatten out like the previous combination. The plugs even looked better. Just jabbing the throttle right off idle in first gear is enough to induce slight wheelspin. We seemed to be going in the right direction, so an SE-R airflow sensor was then modified to fit the GA16's "hot wire" assembly and electronics. It was a lot of work, but it was unfortunately also a complete failure. The engine would barely run and produced clouds of black smoke as it chugged along. Oh well, you never know until you try. More recently, a Racer Wholesale pulse-width and air-fuel meter was installed. As expected, full throttle and higher rpm's pegged the meter at full rich, with injector duty cycle at 80 percent. Either we'll have to go back to the GA16's injectors and bump the fuel pressure, or reduce the fuel pressure to use the SR20's injectors. Or reduce the pulse width of the SR20 injectors. That can only mean going to an aftermarket injection control unit. I was looking at the Simple Digital System from Canada. If it was upgraded with the distributorless ignition, it would open the door for endless future modifications. Maybe next year.
Robert Legere
Legere Motorsports Development
last modified November 14, 1999