You can see just how much collector length was added to this header setup. On the other side, the collector was about 3 inches longer. The effect this had on output was out of all proportion to the simplicity of the mod, as Figure 1 shows.
Bearing in mind that exhaust systems are about diameter and length, let's set some ground rules. Slightly too small is better than slightly too big, just as was the case with ports in the head. Also, especially in the case of the headers, slightly too long is better than too short. Having said that, what most of us race does not strictly adhere to that second statement. Dealing with the header length first, we find that in the case of a two-plane cranked V-8 and a four-into-one header design (which is about all we race), the header lengths are not super critical from about 28 inches to about 40 inches. If the pipes are between approximately 30 and 38 inches, it will be hard to improve much by juggling these lengths. As a result, valuable time is lost while trying to build a set of headers with totally equal lengths unless pipe routing allows such consistency with big, high-flow curves.
As forgiving as the exhaust system's primary lengths may be, the same cannot be said of the secondary or collector length. Depending on the cam and rpm range, the collector should typically be within 12-20 inches to either open air or a large and abrupt change in system cross-sectional area. If you ignore the collector length you could be throwing away as much as 20 lb-ft and 20 hp, even on a relatively mundane engine. The test results in Figure 1 (page 96) show typical gains or losses from secondary length adjustments.
Shown here are some rear wheel figures as measured on a Dynojet chassis dyno. The red curves are for the extended, tuned-length collectors, and the blue represents collectors with near zero length. As can be seen, the differences for such a simple, low-cost modification are substantial.
9. Mechanical Losses
There are plenty of places within an engine where friction can seriously reduce torque, and consequently, power. The piston/ring pack is the number-one culprit here. If you are building an engine for a class that legislates stock-type pistons, be they cast or forged, you need to be aware that the clearance such pistons are designed to use is close enough to limit piston noise to a minimum. When running, many street-type pistons expand to such an extent that they take up almost all the clearance, and their frictional losses are high. Losses from piston friction can be cut by increasing the clearance. Optimal clearance will vary from piston to piston because of differences in material and design. However, you can be sure that the stock clearances used, which typically range from 0.001 to 0.0027 inch, are way too tight. A good starting point with a 4-inch bore engine using a cast piston is 0.004, and 0.005 for an engine with a forged piston. Don't assume that because a minor increase in clearance is good, more must be better. The more the clearance is increased, the more difficult it is to get typical stock wide rings to seal.
If you want to reduce piston and ring friction and the time it takes to break-in the engine, try this for a torque-enhancing move. Use a Scotchbrite pan-scouring pad to polish the rings until they feel smooth and slippery to the touch. Next, use a pad on the bores in a vertical motion, and use Gunk engine cleaner as a lube. Continue the vertical polishing until the bores feel really slippery. This action takes out a minimal amount of metal and removes or smooths the microscopic tears left behind by the honing process. In effect, this process is similar to a fine plateau honing job.
The next culprit for big frictional losses is a slightly bent or twisted rod. You can usually detect such a rod at teardown, as the wear pattern on the piston skirt will have a slightly diagonal direction to it. Any such rod should be discarded.
If I am building a small-block Ford, I often get my blocks from DSS, a well-known Ford specialist. Every dimensionally critical surface is machined on CNC equipment, and the bore finish is top quality.
As for bearing clearances in a semi-stock wet-sump motor, using top-limit factory clearances seems to be the best thing to do. Although this makes for a little more oil flailing around in the crankcase, it seems to allow production machining errors and stock component distortions.
10. Good Lubes
There is a generous number of advanced oils available to the racer. I have used Red Line, Royal Purple, and Joe Gibbs Racing Oil from CV Products, as well as a number of brands that you probably have never heard of, all to good effect.
All told, I can recommend a multitude of really good oils. With these oils, you could be pouring in about 4-5 lb-ft of extra torque, but sometimes you just cannot get them when you most need them (11 o'clock the night before the race). Not all of them are available at the local parts store. So why not just pour in some quality off-the-shelf oil made for street use? Well, here comes the crunch. Almost all the low-buck race classes call for the use of flat-tappet cams. The contact patch of a flat-tappet cam is the most critical factor considered in the formulation of a lube that will deliver tappet life. Most of the additives used here are zinc derivatives, and current environmental legislation has mandated their removal from oil used for street applications. This leaves the flat-tappet cam, especially one with increased lifter accelerations, at severe risk. But there is a fix-Oil Extreme. At an educated guess, I would say this is at least twice as good a high-pressure, friction-reducing additive as the zinc additives it replaces.