To my shock and dismay, to win a race, you have to finish. Damn. And not spin. Damn. And not blow up. Triple damn.
As you'll recall, I embarked on this racing adventure after helping found a low-buck racing series for SR20-powered Nissans. In addition to getting on the track for cheap, the main goal was to use my ghetto-fabbed cheap bucket to tear ass on another SE-R--one meticulously designed and fabricated by our own nerdmaster general Mike Kojima, funded by the deep pockets of retirement queen Grandma Paule (judge of the Grandma test for USCC) and driven by her good-for-nothing crack-gigolo nephew Tom.
The race season lurched toward that goal initially, with our race meeting sending Grandma's team home in shame. I was set to repeat the pummeling at another encounter at Willow Springs until a piece of asphalt, thrown up by another car, found its way through the cooling holes in my bumper and core support and gnawed through several radiator fins. In just a couple of laps, we went from ass-kicking to detonation, a blown head gasket, and the end of my race.
Before losing compression on all cylinders, however, we had seen impressive power gains, documented in the following pages with just bolt-ons and cams. With this engine terminated, seeing the glass as half-full, I figured "well, at least I'll be able to take on a project way over my head, spend lots of money I don't have, and probably fuck up the first engine I ever build." And with that, we dove in.
The SE-R Cup class is regulated by a 15:1 power-to-weight ratio. Make more power, and you have to add more weight to maintain this ratio. With this class structure, a big torque number, not horsepower, is the Holy Grail. Why not then build a high-compression engine to make more horsepower and torque and then tune down from there?
So we did that. And it didn't make all that much more power. And then we blew it up (overheated it, actually) in the kind of red-misting mechanical rage that only tremendous investment with disappointing results can provoke. We also spent two full days on the dyno playing with various cam configurations so you won't have to. You'll be amazed how much of a difference one degree of cam timing can make.
Prior to this race series, we dealt with the car's transmission, the great flaw in an otherwise exceptional powertrain package. Front-drive SR20 transmissions aren't horribly robust, with frequent third gear failures and fifth gear pop out. Moreover, the stock viscous limited-slip differential dies after just a couple of run-ins with R-compound tires and high loads seen on track. At the same time we installed a mechanical limited-slip differential, we had the gears cryo-treated and the case reinforced where it's known to crack. In an attempt to lengthen the life of the tranny, we poured in Redline Shockproof tranny fluid, the color of Pepto-Bismol and the consistency of molasses. It has worked great, so far.
My 1991 SE-R is powered by a roller-rocker SR20DE, which is found in '00 and newer Nissans. This engine claims identical output to the previous SR20s, but boasts a slightly better torque curve. The biggest changes are roller-rocker cam followers and other steps taken to remove frictional losses, and massively decreased rotational mass. For more detail, check out page 117 of the May '02 issue.
But this new SR20 doesn't just drop in and replace an earlier one. Nissan saw fit to completely change the engine management, so making the new engine run with the old wiring harness and ECU requires a fair amount of rewiring, like the sensors and injector leads. This isn't a swap for the cautious or, frankly, the intelligent.
Jim Wolf Technology produces the only performance cams for the roller-rocker SR20. Designed as street cams usable with the stock valvetrain, the impressive power gains we saw point to the extent to which the roller-rocker engine was geared towards low emissions levels. We saw 11 hp and 5 lb-ft of torque over the stock cams with no cam tuning and using a JWT-tuned ECU program. Keeping the early 1991 ECU allowed this tuning, as the code in the OBDII ECUs, which came paired with this engine, has not yet been unraveled.
Whereas JWT spins standard SR20s to 7700 rpm with its standard ECU upgrade, the valve springs in the roller-rocker head are quite a bit softer, so the fuel cut had to be set at 6700 rpm.
A thousand less rpm is not ideal when you have the same gearing as everyone else, but we figured the torque difference would offset this. Many second-gear rolling starts, however, have proven that theory wrong. The solution is heavy-duty valve springs and a higher redline.
Thanks in no small part to our Mad Max-style cage, our car is a little porky. We ran the first race with a stock midpipe and a cobbled-together muffler section, which didn't do much for power. The solution is cheap, simple, and loud. Inspired by the 3-inch JG Edelbrock B-pipe, also known as a dump pipe, which made huge horsepower and torque gains on our Project Civic, we headed to Hotshot, which fabricated the car's previous header system.
Andris Lavins hacked an old B-pipe assembly, welding the two secondary pipes into a single 3-inch section. We guessed on the length, about 3 feet, and gained 6 hp and 4 lb-ft of torque, made more horsepower and torque throughout the rev range and saved 25 pounds over the old system. However, earplugs are now a must.
With 145 wheel hp and 2,200 pounds, we're safely with the 15:1 ratio, and even could stand to lose a few pounds.
The Parts Bin idea
It just so happens that if you overbore the SR20 1 mm, 87mm non-turbo Z32 300ZX pistons will give you an 11:1 compression ratio. JWT stocks these low-cost pistons, and if you have JWT bore your block, it uses a torque plate that makes for a perfectly round bore, which is then plateau-honed for faster ring break-in. JWT also drills and taps the oil passages in the bottom of the block to fit the oil squirters from the SR20DET to aid in piston cooling.
All this planning was dashed when 0.006 of an inch was removed from the block, and .008 from the head to reverse the damage caused by the overheating episode. This raised the compression ratio approximately .3 points, for a total of 11.3:1.
The roller SR20's crank has half the counterweights of the model it replaces, reducing rotational mass by 8 pounds. Sizing the bearings is essential. Unlike many older American engines for which just one stock bearing size is offered, Nissan offers five. By decoding the journals using numbers stamped on the crank and block along with your Nissan service manual, which you, of course, read cover to cover before attempting any of this, you can see what bearings were fitted to the engine when new. This obviously doesn't account for wear.
We fortunately had access to precision micrometers so we could measure the crank and block journals to the 0.0001 of an inch. Plasti-gauge is always an option, but we found it to be far too imprecise for the tolerances required. In theory, you should get the precise bearing match for each journal and order them from your dealer. In reality, we had a selection of Nissan bearings on hand and a race at the end of the week. Fortunately, most of the journals were close to par with what we had on hand, and on the tight side of factory specs, which creates high oil pressure, ideal for road racing where the extended g-loading can produce undesirable low oilpressure situations.
Although the factory crank comes forged, balanced and micro-polished from the factory, as we balanced the rest of the engine internals, we figured it was worth having the crank more precisely balanced, repolished, and sized slightly for our new bearings.
After numerous calls to reputable engine builders, one name kept coming up: Castillo's Crankshaft Specialist. Tucked away in an unassuming industrial complex in La Mirada, Calif., the Castillo brothers specialize in hard-to-do, one-off cranks for companies like Dutweiler and for projects like the stroked-crank MR2 Spyder James Chen brought to the USCC last year.
Because my monkey-ass was doing the build-up, some snafus were inevitable. It seems my monkey twin was working at the shop that did the valve job. Instead of removing the alignment dowels from the head before resurfacing, they machined them flush with the head. I noticed this only after assembling the head. Being the week before a race, I did what any competition-crazed, sleep-deprived primate would do while building his first engine. I ripped some dowels from another engine, used a drill to carefully remove the old dowels, and proceeded with assembly.
Of course, those dowels are hollow, and a head bolt passes through each one. That means we were drilling in a head bolt hole, which goes through to the top of the head, where the cams, rocker arms, valve springs and lots of sticky assembly lube lives. To drill out the dowel, the head has to be upside-down, of course. Gravity pulls all those sharp little dowel fragments down into the top of the head.
Luckily, monkeys have lots of flea-picking practice, so the long night of meticulously chasing every little shard of metal brought back memories of my childhood in the jungle.
SR20DEs are not as good as the Honda B-series at running ludicrous compression on 91-octane pump gas, so making this engine happy requires careful tuning and a little bit of octane enhancement from our old toxic friend Toluene. Toluene is "look at me sideways and I'll give you cancer" nasty, and is commonly available through paint stores, where it's sold as a solvent. Toluene also happens to be 114 octane, and is a common ingredient in octane boosters, as well as one of the more important octane enhancers in Japanese fuels. Remember those crazy 1.5-liter, 1,500-hp F1 engines from the '80s? They drank mostly Toluene, which also has particularly good resistance to detonation.
Internet wisdom dictates that when mixed with 91-octane fuel, a 30-percent ratio of Toluene will yield a 97-plus octane mix. Oh, and this doesn't get the thumbs up from the EPA, so the cops who test your fuel when they pull you over can write you a ticket or something. Yeah, that'll happen.
After drinking a break-in diet of straight 30-weight oil for 20 minutes at 3000 rpm, we changed to 20W-50. The break-in period consisted of trolling the newly born but pissed-off sounding engine through an industrial park, much to the delight of everyone in a three-mile radius. That 3-inch dump pipe has its downsides.
Once the rings were seated, we ripped off a baseline run of 151 hp and 146 lb-ft of torque. SR20s with 11:1 compression often make 160 to 170 hp, so we were a bit mystified by our results. Then I remembered that I built the ingine.
Even with the 97-octane diet, the car was still running a 9.5:1 ECU program, so we headed to JWT in El Cajon, Calif., for a remapping with a stock exhaust system installed.
Our SE-R was mounted to its Dynapack dyno and Clark Steppler hooked up a Horiba A/F analyzer and a Nissan Consult and went to work. After installing a base high-compression map, Clark played with fuel and timing until he had a nice, fat program that erred on the side of safety.
With the mandated 15:1 horsepower-to-weight ratio, torque is the name of the game. In the time between receiving our S3 roller cams from JWT and rebuilding our blown-up engine, JWT came out with a set of trick valve springs and titanium retainers that would raise our rev limit.
Designed for the stock 6500 rpm rev limit, the roller S3s nose over fairly hard after 6200 rpm, not ideal for road racing, where the engine lives mostly between 5000 rpm and redline.
By playing with cam timing, we hoped for increased performance in the upper rpm. We procured a set of anodized, adjustable billet SR20DE cam gears from Stillen to do just that.
Cam tuning on the SR20 is always a pain, requiring removal of the valve cover, but these gears allow easy adjustment with Allen keys once the cover is removed. We've seen a cam gear with only four adjustment bolts slip, so we were happy to see six bolts with anti-slip washers on the Stillen units. Once you have the ideal cam adjustment figured out, another option is to buy preadjusted eight-hole cam gears from someone like JWT, which completely eliminates the risk of slippage and is a good idea on racecars where everything tends to go wrong, always.
With many runs to make during business hours, we had to bolt a quieter exhaust on to the car, which would also better approximate what people will run on the street. We settled on what we could beg, borrow and steal: a stock cat, midpipe, and rear section tubing with an Edelbrock RPM-series muffler.
With the Clark Steppler magic tune, our Dynojet read 148 hp and 144 lb-ft of torque. Not exactly what we hoped for.
Rushing for our first race, we had advanced the intake cam 1 degree and the exhaust 2 degrees, hoping to get midrange and kill the top-end power a bit. We were successful.
Putting the cams straight up, we got 151 hp and 141 lb-ft of torque. Retarding just the exhaust cam yielded both inferior horsepower and torque numbers, so we put the exhaust cam back to zero and moved on to optimize the intake cam. By advancing the intake cam and opening the intake valves sooner, we saw better torque numbers with minor peak horsepower penalties, but with big roll-off on the top end.
The best horsepower numbers were often matched to the best torque figures as well.
The engine was happy with 2 degrees advance on the intake. While we were happy with the horsepower and torque numbers, power was still nosing off heavily at 6200 rpm. We returned our attention to the exhaust cam. With the intake cam advanced, the engine was highly sensitive to having the exhaust cam retarded, thereby increasing overlap. We got our top end, but with huge penalties in horsepower and torque before 6200 rpm. For a street engine, having an extra 13 lb-ft of torque at 3500 rpm is worth a softer top end.
Spending as much time as we do on the racetrack in the last 1000 rpm before redline meant having to make that tradeoff.
We moved the exhaust back to .5 degrees retarded, hoping to fall somewhere in the middle of good midrange and top end. Exhaustive testing complete, we bolted on our race dump pipe for the big number. Over two points more compression, a slight increase in displacement, and much dyno time yielded us a whopping 3 more hp and 3 lb-ft of torque more than our stock cam timing, stock compression engine.
This could be a function of a crescent wrench I left in a cylinder bore or coagulated knuckle blood on bearing surfaces. Your mileage may vary. The moral? A box-stock roller SR20 is one kick-ass engine, can be yours for about $400, and with just a few bits from the wizards at JWT, is a torquey wonder for your street or racecar. Or you can try to build an engine.
So halfway through the season, we have blown engine number two, and harbor an enthusiasm deficit. Like a phoenix, however, the #27 will rise from its head gasket ashes--with a stock compression engine, upgraded cooling system, weaknesses addressed, all while keepin' it cheap, keepin' it real.
JWT's roller cams produced impressive power gains, but were designed for a powerband limit
On a stock engine, the JWT roller cams are impressive, delivering 11 hp and 5 lb-ft of tor
Be careful with your engine parts. The block somehow got a nice ding on the head gasket se
The head was slightly warped in the overheating incident, and needed to be slightly surfac
So you need backpressure to make torque, eh? Isn't it about time that myth died? Lopping o
Good lube is a must, always; we took no chances and used Torco engine assembly lube, coati
JWT machined the block for these stock Nissan oil squirters, which aid in cooling the pist
We cleaned up the bearing on the small end of the rod using a ball hone soaked in oil on
Tap the wrist pins into place using an extra pin on the bottom for alignment.
Heating the pistons causes them to expand slightly to more easily accept the floating wri
The sides of the connecting rods must be shaved .010-inch per side for proper clearance i
For a street engine, this might be a fun setup. Advancing the intake cam gave huge low an
The stock piston/rod assembly is on the left; the extra heft of the Z piston is easily vi
Rings must be checked for proper gap before installation on the pistons. The rings, as rec
Tuning for the high-rpm performance our engine needed killed our earlier torque gains. On
We prepped the pistons by removing sharp edges on the piston surface caused by the factory
With the 3-inch dump pipe back on, over two points more of compression, a slight increase
We double-checked all our clearances as installed with our Castillo-serviced crank.
Don't forget the easily overlooked step of oiling the head bolts before torquing them to s
After assembling and filling the head with metal shavings, be sure to spend some quality t
The Stillen adjustable cam gears have six bolts to minimize the chance of slipping. If you
Castillo's Crankshaft Specialist
P.O. Box 909
6261 Katella Ave., Ste. 2B
Jim Wolf Technology
So. California Region NASA
Puente Hills Nissan