When this Series of stories began, its stated objective was to meld Cup engine practices with what might be applied to weekly racers, certainly in terms of affordability and doability. And, as we button up the subject, it seems worthwhile to spend a little time discussing some of the better ways to evaluate a given parts combination, notably on an engine dynamometer. In so doing, let's examine how the use of such test equipment has transitioned to today's methods and technology, and discuss some down-to-earth techniques for sensible data evaluation.
Most all of us in this industry/sport recall the early purpose of engine dynamometers. Essentially, by their very nature, they were used to simply measure brake torque; data that was then converted mathematically on a time-rate basis into brake horsepower. Some of the first engine dynos to find their way into motorsports were iterations of dynos that either came from the OEM or were built by persons who had experience in that environment. Early dynos at Holman-Moody's and Smokey's come immediately to mind.
But regardless of their origin or capabilities, the fundamental goal of an engine dyno was to measure brake horsepower. At this point, it hadn't yet become evident to dyno users that the use of in-car pieces (headers, ignition systems, cooling systems, etc.) on the test stand would be a logical step toward the need for linking engine testing with on-track performance. The importance of this fact was brought solidly home to me in the mid-'70s when sorting out some cylinder-to-cylinder mixture distribution fixes in a now bygone "Smokey Ram" intake manifold. Distribution fixes Smokey had determined on the dyno barely resembled what were required on the track.
Long before Land & Sea's DYNOmite...
Long before Land & Sea's DYNOmite Dynamometer, Smokey Yunick "drove" a dyno of his own. Photo by Smokey Yunick
In addition, the rather immobile state in which an engine on a test stand operated didn't duplicate the dynamics (including air movement, spark, and fuel calibration requirements) found on the track. So it seemed something that might be common to both environments would be a useful next step. Fortunately, we landed on the commonality of "combustion efficiency" between dyno and track as one possibility.
I will not take credit for what you will now read. It came from a now-departed good friend who was deeply involved testing all manner of engines at the Champion dyno shop in Long Beach, California, circa 1971. On one occasion, he happened to be doing some "weekend testing" at the Edelbrock facility and I was there helping him. What immediately got my attention was the fact he was only recording brake horsepower (along with fuel flow) for his calculations of brake specific fuel consumption (BSFC). He had no apparent concern for how much power was being made, beyond use of this information.
If memory serves, he ran a series of tests through which only spark timing and fuel enrichment were changed. When he'd completed this exercise, he was finished. So, flushed with curiosity, I asked him to elaborate on why BSFC was so important to him. Paraphrasing his response, it went something like this: "What we're trying to do is an efficient job of converting fuel into heat (power). For any given package of hard parts, unless you start swapping or modifying components, the only meaningful adjustments you can make are to spark and fuel. Once you've adjusted these to the point of minimum BSFC, without an attending loss in power, there's nothing else you can do. Combustion efficiency has been maximized when BSFC, under the conditions I just described, is minimized. And when you get to that stage, you can then observe how much power you're making."
In all honesty, during the next thirty years of engine testing, I have found this approach to be the singularly most valuable technique I've ever found for evaluating engines on a test stand, aside from certain specific design or performance objectives and some supporting testing techniques.
Bobby Clark looks over the...
Bobby Clark looks over the shoulder of engine builder George Pils as they evaluate the results of a dyno run. Photo by Rob Fisher
Over time, I discovered some additional benefits to "BSFC testing" that may be of value to weekly engine builders and testers. In particular, even though engine dyno testing equipment and data acquisition are far more sophisticated than in previous years, there are some fundamental relationships that can be helpful when dyno testing.
First of all, you will find that an engine's BSFC and volumetric efficiency (VE) curves should mirror each other (see figure 1). In a previous Series segment, it was mentioned that VE curves and torque curves (except for the influence of pumping losses) can be essentially overlaid. In addition, unless an engine is experiencing a significant combustion efficiency problem (enrichment dilution from excessive exhaust residue, improper air/fuel mixture ratios, isolated wet-out in the combustion space, or related abnormalities), minimum BSFC should occur at or near peak torque. In fact, it is at this point in an engine's speed range that you can run repeated "spot checks" for fuel calibration adjustments, until further reduction of BSFC nets a power loss. It's then that you've determined minimum BSFC for a given set of engine components and conditions and, correspondingly, the best attending combustion efficiency which in simple terms is the conversion of fuel into power.
Next, this process will also set a bench-mark BSFC that you can use as a template for values below and above the engine's torque peak(s). However, at engine speeds below this rpm point, there isn't sufficient piston speed to generate higher volumetric efficiency, and above peak torque there's not enough time to maintain the VE achieved at peak. As a result, you can expect BSFC numbers to be numerically higher in these two ranges, even though your ultimate goal is to minimize these values.
Figure 1 - As pointed out...
Figure 1 - As pointed out in the text, minimum BSFC. and maximum brake torque should occur at or near the same crankshaft rpm. For a given set of engine components and state of tune, this value of BSFC (once optimized) can be used as a "target" value for best combustion efficiency. Photo by Jim Mcfarland
Finally, remember that BSFC can be used as a measure of combustion efficiency, although the data can be flavored by other conditions that will cause an increase in the actual values. For example, suppose a particular condition exists that's preventing adequate evacuation of combustion residue from the combustion space (e.g., improper valve timing, inefficient exhaust system, etc.) In this case, relatively inert exhaust gases dilute fresh air/fuel charges, upsetting proper enrichment and tending to reduce overall combustion temperature...and combustion efficiency. Consequently, BSFC values will increase.
In another instance, you may note inordinately high BSFC values in engine speed ranges beyond peak torque. If, again for whatever reason, mechanical separation of air and fuel may be taking place (either along the inlet path or in the combustion space), and BSFC values will increase. They will also rise when mixtures are overly rich.
The trick is to compare exhaust gas temperatures with BSFC values (if EGT data is available) and correlate the data accordingly. As an example, undesirably high BSFC values, combined with lower-than-normal EGTs can often be linked with excessive enrichment, even beyond what you can read on spark plugs. Just remember that high BSFC values in the upper rpm ranges can be a blend of the two previously-mentioned conditions, so you'll need to keep an eye of EGTs to pinpoint your analysis.
On The Exhaust Side - Periodically in this Series, we've touched on ways you can influence torque curve "shape." This practice makes more sense if you'll consider some of the major factors that affect where in a given engine speed range torque is produced. What's convenient about this approach is that you'll quickly discover some physical/dimensional relationships between induction and exhaust systems that are common to both. Here's one example of how this can work.
Figure 2 - Changes in collector...
Figure 2 - Changes in collector length tend to affect the "tuning range" (below peak brake torque) in which torque changes can be expected. Generally, altering collector length tends to impact torque output below peak torque. Note that little if any attending changes will occur above peak torque. Photo by Jim McFarland
Torque output below peak is materially affected by header collector dimensions. Adding volume by way of length tends to build torque below the peak. Experience has shown that dyno testing a range of collector lengths, noting the gains, and relative rpm can become an at-the-track tuning tool, given certain track length and surface conditions. Plus, this is a very convenient, quick, and comparatively inexpensive modification. Keep in mind that you can also tune torque out of this range by shortening or removing collectors completely (see figure 2).
In a similar fashion, you can manipulate torque above peak torque rpm by lengthening or shorting primary pipes. Such adjustments can easily be made at the junction between primary pipes and collector, even if the "collector" does nothing more than merge these pipes. Of course, as previously stated in this Series, such adjustments tend to "rock" the torque curve around peak torque rpm (see figure 3).
Shifting the torque peak(s) rpm higher or lower in the rpm range is a function of pipe diameter (section area inside the pipe), largely because a change in inside diameter affects mean flow velocity at any given rpm. This consequence can become a tool to targeting the rpm at which you want a torque boost, simply because you can manipulate where the critical mean flow velocity occurs (see figure 4).
On The Intake Side - Here, the flexibility for making flow passage dimensional changes is far less than for the exhaust side. However, there are some guidelines you may want to consider.
Figure 3 - Typically, you...
Figure 3 - Typically, you will note two fundamental conditions (relative to torque output) when altering primary pipe length. As a rule, shortening pipe length increases torque in the upper rpm range (generally as a penalty below torque peak) and increasing length builds torque in the lower engine speed ranges while reducing it at higher rpm. In either case, watch for attending changes in BSFC. that may require attention. Photo by Jim Mcfarland
First of all, as previously indicated, regardless to which "tuning theories" you subscribe, significantly similar dynamics exist in both intake and exhaust systems. Even the "wave motion" theory maintained by many tuning proponents finds similarities in how both systems (intake and exhaust) can be treated. Despite differences in piston position when optimum pre-combustion conditions favor the intake or exhaust cycles (roughly mid-stroke for the intake cycle and bottom dead center for the exhaust), flow dynamics are surprisingly similar.
Included in the similarities are passage length, section area, and the fact that the roughly 240 feet/second mean flow velocity effects, for all practical purposes, are the same. What this does is aid decisions in the selection, evaluation, and modification of these systems during dyno testing. Even absent the transient load (resistance of a racecar during acceleration) and the dynamic effects not precisely duplicable on an engine dyno, you can still perform some valuable tests, certainly if the data evaluation procedure is valid. We'll address that issue in a moment.
In terms of "tuning" intake manifolds on a dyno, recall our previous comment about the Smokey Ram. Even in manifolds not of the "box" type that don't allow random fuel migration in a comparatively open volume prior to runner entry, manifold flow dynamics related to the part's design can influence cylinder-to-cylinder air/fuel charge mixture distribution. Today's wet-flow benches have provided ample proof of this. And even if you've done a good job of balancing cylinder-to-cylinder air flow on a bench, the "quality" of such flow can materially affect the distribution of fuel in that air during engine operation.
I've seen instances where sorting out distribution problems on an engine test stand can be a good start, and the use of exhaust temperature probes and skillful plug reading will certainly help. Given the response time of today's HC/CO emissions meters, sampling exhaust byproducts immediately downstream of the header flange is better yet.
Figure 4 - Somewhat exaggerated,...
Figure 4 - Somewhat exaggerated, and while you may not observe quantitative values consistent with this graphical representation, the concept will remain valid in terms of how you can expect a "net" torque to be produced (summation of red and blue torque curves show by black curve). Essentially, results from a "two-size" or multiple peak header set will tend to broaden the engine's overall torque output, as you might want when off-the-corner and mid-straightaway acceleration is particularly critical. Photo by Jim McFarland
In fact, this method encompasses the ability to help quantify combustion efficiency (exhaust-diluted combustion, mechanically-separated air/fuel charges prior to and during combustion, etc.) by measuring the amount of unburned fuel (HC) and air/fuel ratios (CO) in cylinder-to-cylinder exhaust passages. Whether you can afford or have access to such equipment, I've seen some pretty useful data produced by the use of emissions sampling equipment from automotive repair shops. At least some information is better than none.
We touched on some of these in early sections. Here are some more you can consider. A few might seem rather rudimentary, but you'd be surprised how often they are overlooked.
First of all, test all the parts you plan to run on the track. While both temperature and related factors encountered on the track may not be duplicated in the test cell, at least you'll get a sense for where in the speed range power is made. That's critical to trackside tuning and gearing choices. Plus, unless you're using the same parts, you'll be tuning to some other combination instead. For example, even though the "dyno" headers you might be using are dimensionally the same as the racecar system, bends, radii, and other little idiosyncrasies can sneak up on you and produce off-target results, by comparison.
One area where dynos may be somewhat lacking in helping you "tune" components is the intake manifold. Consider the following piece of information. Let's say you've conducted back-to-back dyno tests using two essentially different manifolds. Further, let's also say they contributed to corrected power curves that are essentially the same, quantitatively. However, you noted that even though you minimized the BSFC curves (once plotted), one is trending numerically lower than the other. So, which manifold should perform better on the track? What this comes down to might be called "transient" horsepower or torque. The latter may be preferable.
Essentially, you'd like an engine to accelerate quickly (under load) throughout its operational rpm range. And, obviously, the quicker it can do this the greater the chance for a racecar to accelerate, all else being equal. Over time, it has been demonstrated that the engine operating with the lower BSFC will also tend to produce the quickest acceleration. Interestingly, you can apply this same logic when evaluating other major engine components, including headers and cylinder heads.
How do you compare this evaluation technique with so-called "acceleration" tests performed on the dyno by which a time-based and controlled unloading of the power absorber is applied? Not certain you can, even though this method does provide useful brake power data (for test-to-test comparison purposes) at minimum wear and tear on an expensive race engine. Unless an engine is accelerating a mass (load) and experiencing the changes in dynamics this involves, you've not basically brought the track to the dyno. However, programming throttle positioning as a function of load, if properly done, can create an operational environment that simulates specific track conditions. Higher-end engine dyno facilities do this on a routine basis.
There's clearly a limit to what can be accomplished in the pages of a magazine. Explanations often create unanswered questions. Editorial attempts to cover all aspects of a given topic or sub-topics can suffer the same consequence. However, even recognizing these limitations, this Series was intended to minimize those handicaps and attempt stimulating both thought and reason about and for the subjects discussed.
Make certain you spend a few minutes with the sidebars attached to this month's wrapup. Both are stuffed with helpful tidbits, particularly the one from Charles. And, finally, he and I hope you'll allow this Series of stories to be a stimulus for telling CIRCLE TRACK'S Editor (firstname.lastname@example.org) what else you may want to explore. You might be surprised at the response and pleased with the results. Meanwhile, thanks for your interest. It's been a pretty good run and hope you've enjoyed it as much as we have.
Engine builder George Pils...
Engine builder George Pils tunes the carburetor on our Dirt Late Model project car's engine in an effort to evaluate our intake system. Photo by Rob Fisher
There was a time when a race engine's parts were selected based on the performance potential of each component. In fact, specialty parts manufacturers often designed toward the optimized potential of their parts with little regard for their impact on other components. Today, for engines from Sprint Cup to Saturday nighters, the functional integration of basic components is essential to reliable, appropriate, and repeatable power. Most of the responsibility for making component choices falls on the engine builder or racer assembling their own engines. So with that in mind, we offer this sidebar as part of the concluding steps in this Series of stories.
Shaping The Torque Curve We previously noted that an engine's volumetric efficiency (VE) characteristics are closely tied to the shape of its torque curve. Aside from the influence of pumping losses and based on its achieved level of combustion efficiency, VE and torque curves are fundamentally similar.
Let's restate that. If you want an engine to produce torque in its lower- and mid-rpm, it's necessary to select parts that enhance breathing characteristics in this span of engine speed. In fact, that approach applies to whatever rpm range of operation you've selected. And while an increase in mechanical compression ratio will also provide torque boosts in the selected range of rpm, increased cylinder pressure at higher rpm can be hindered if the engine fails to have sufficient air. So, VE is key to making torque where you want it produced.
In earlier CT tech material, we briefly discussed the concept of how airflow velocity relates to volumetric efficiency to the extent that you can design, modify, or build intake and exhaust systems on the basis of how they affect air rate. We further suggest that flow rates associated with torque vs. rpm (notably torque boosts at selected rpm) could be based on flow passage section area. Here's the linkage stream: Flow rate vs. torque boost vs. rpm vs. piston displacement is tied to flow passage section area. And while other engine parameters have an influence on the characteristics of a given torque curve, the design of intake and exhaust systems can have significant impact on VE performance.
Much has been written to describe the dynamics of intake and exhaust systems. To elaborate on the complexities of how contemporary design and analyses of these systems can be accomplished at the Sprint Cup level serves no immediate benefit to the Saturday night racer. However, those elements that can be reduced to simpler and more hands-on suggestions are helpful. It's in this context that the above information and what follows has been compiled.
Functionally Integrating Intake And Exhaust Systems These are powerfully influential components in a race engine, certainly in terms of where in the rpm range torque is produced. Regarding how they influence engine output, you can view them separately or in conjunction with each other. But with either approach, it's important to recognize and understand that both have significant impact on VE (torque) performance. To optimize their effectiveness, you need to decide on the range of rpm where the engine will be most often operated; minimum to maximum (we alluded to this in a previous Series segment.) Making such a determination isn't always a simple process, but the outcome can materially affect your parts selection.
Before you begin tuning on...
Before you begin tuning on a dyno, you have to check everything to avoid problems like this. Note the fuel in the float. Photo by Rob Fisher
As a rule of thumb, the wider the range of operational rpm, the farther apart (in rpm) you'll need to "tune" the intake and exhaust systems. Conversely, as in the case of tracks where the rpm spread is narrower, you can bring the intake/exhaust tuning points closer together. So how do you do that? Let's return to the flow velocity vs. flow passage section concept.
Take a look at the illustration (figure 4) describing how flow passage section area relates to an engine's peak torque rpm (volumetric efficiency). Recognizing that we're presenting this material in a very simplistic format, compared to more complex computational methods, it turns out the concept still holds true. You may recall a previous point that a mean flow velocity of about 240 ft/sec occurs at peak torque. Based on that information, note how peak rpm shifts upward as a function of increased flow passage section area. At least in terms of this concept, you can evaluate intake manifold runners and exhaust header primary pipes in the same fashion, for purposes of selecting, designing, or evaluating existing components.
You will also find that intake and exhaust systems can be tuned to different rpm points, within the range of anticipated rpm. This can be useful in a number of ways, based on final drive gearing, length of track, and track conditions. In addition, you may find it useful to select or configure intake manifolds with different passage section areas. This will allow you to broaden (flatten) the contributions made to a total torque curve, thus enabling a wider range of effective torque; e.g., coming off corners and continuing past mid-straight-aways.
On a personal note, I've previously worked with NASCAR teams who not only used this concept to their benefit, but combined different intake and exhaust passage size with the appropriate intake and exhaust lobe designs and timing-e.g., short valve events associated with the longer (lower rpm) intake and exhaust flow passages, and longer events for the shorter (higher rpm) intake and exhaust passages.
The approach amounted to treating the engine as a multiple set of single-cylinder engines by using different intake and exhaust lobes to match the tuning points for the intake and exhaust systems (relative to where in the operational rpm range torque boosts were desired). When combined with the proper gearing to match engine speed and track conditions, this method turned out to be a targeted way to create track-specific engine and gearing packages. Properly done, the results were sometimes spectacular and frequently beneficial.
Overall, as pointed out in previous segments of this story Series, it's both possible and helpful to integrate an engine's torque performance with gearing and track conditions (including length, banking, surface, etc.) By so doing, you will move toward optimizing on-track performance by linking engine performance potential with track requirements and opportunities.