Air Delivery

Topic:

Subsection:

Other:

Air in motion is what powers the internal mechanism that meters the fuel inside the carburetor. At the heart of that mechanism is the venturi. The venturi is very simply a small section of pipe with an "hour-glass" profile inside it. Intuitively, it appears that the narrowing profile of the venturi would serve to impede the flow of air or fluid through it, when in fact if it's designed correctly, it is not an impediment at all. A properly designed venturi will flow air at the same rate as a plain piece of straight pipe with the same diameter and length. But the amazing counterintuitive efficiency of the venturi pales in comparison to what goes on deep within it...

The nature of the venturi is such that as air flows through it, there is a pressure drop created at the narrowest diameter inside it. On the Nikki, like most carburetors, there is also a smaller "booster" venturi set down into the primary venturi. It amplifies the pressure drop, helping to draw fuel in at low RPMs, when the air flow is low.

Inside the booster venturi, located at the narrowest point where the pressure drop is the strongest, is the fuel outlet pipe. The venturi assembly is situated so that the fuel outlet pipe is above the fuel level in the float bowl, otherwise fuel would simply spill out into the venturi. The fuel pipe comes up from a circuit at the bottom of the float bowl, where the fuel jet limits the rate of fuel that can flow up the circuit and out the pipe.

When air flows through the venturi, the pressure at the end of the fuel pipe becomes less than the atmospheric pressure acting on the fuel in the float bowl, causing fuel to spill out of the fuel pipe and be carried into the air stream into the engine, ideally in a fine mist. The pressure differential, or vacuum, acting on the main circuit is often referred to as the "fuel signal". As the flow of air through the venturi increases, the pressure at the fuel pipe decreases, creating a more powerful signal. The pressure drop is not linearly proportional to the flow through the venturi. The pressure decreases at a higher rate than the air flow increases, which means that the ratio of fuel to air increases, also. But the carburetor has a way to "bleed off" some of that powerful signal to the fuel inlet pipe in order to maintain a usable air fuel ratio; the "Emulsion System". Even a venturi of mediocre efficiency will be able to draw enough fuel at high velocity to perform well in the high RPMs as long as the venturi can flow enough air. Because the emulsion system can divert excessive vacuum present at the fuel pipe, the more efficient and powerful the venturi is, the better the carburetor will be at low air velocity, and thus in the low end of the power band.

I've often said jokingly that a cut up beer can with a straw stuffed in it's side would draw out enough fuel at the top end for the 12a. The fact is, most any carburetor that can flow more than the maximum theoretical needs of the 12a is capable of delivering near perfect mixture at WOT. But the larger the carburetor is, the more the low end will suffer. This holds true for an appropriately sized carburetor with inefficiently designed venturis, as well.

The venturi utilizes something called the Bernoulli principal, named for Dutch mathematician Daniel Bernoulli (1700 - 1782). It states that as the speed of a moving fluid increases, it's pressure decreases. Though this is a bit of physics that has been applied to sewage removal and community plumbing in city dwellings for over 5000 years, Mr. Bernoulli is credited with coming up with the mathematical equations currently used in modern science.

The venturi is named for Italian physicist Giovanni Venturi (1746 - 1822), who dedicated the bulk of his work to fluid dynamics, so much so that the venturi with which he is accredited pales in comparison to the sum of his other discoveries.

There is actually an "ideal" venturi. It's not a theoretical myth, either. The ideal venturi would, of course, have a flow efficiency of 100%, and as large a pressure differential as possible at peak air flow through it. The venturi would have an inlet edge angle of 16 - 20 degrees and an outlet edge angle of 7 - 11 degrees. The equal inlet and outlet diameter would have a cross - sectional area 4 times that of the smallest inside diameter. (In other words, the smallest inside diameter would be � the largest diameter.) At 100% efficiency, the flow rate would be 137.7 cfm per square inch. The greatest pressure differential inside the venturi, measured at a few hundredths of an inch below the constriction, would be 25 inches of water, or about 0.9 psi.

From this ideal venturi, we can calculate the theoretical venturi configurations to meet the demands of any given engine, as well as estimate the overall flow efficiency of any given carburetor. And that's just what we're about to do...

Again, I'll use the 12a rotary stock port as the engine of choice for this exercise.

We've already dremonstrated in the Rotary Engine section the theoretical maximum needs of this engine to be 306 cfm @ 8400 RPM at a VE of 90%. For argument's sake, we could also take the theoretical needs of the 12a @ 10,000 RPM at a VE of 100%, which comes out to 405 cfm, to appease those still unconvinced of the relatively scant needs of the 12a.

Keep in mind that this exercise is going to give us a theoretical venturi configuration necessary to meet the demands of the engine, and as such it's to be taken for granted that the venturis are indeed "ideal" in efficiency, taking into consideration no air flow impediments what so ever. (... in other words, those pesky extraneous components such as booster venturis, shafts and throttle vlaves.)

Maximum Engine Demand cfm	= total area of venturi inlet/outlet
137.7 cfm/sq. Inch

How we divide up the combined area of the venturis is purely subjective. We can choose to apply the information to a two barrel or four barrel carburetor, though we would have to make an educated guess on the spread of the primary and secondary bores on the latter.

For a 2 barrel carb, that's only 1.11 square inches of area for each venturi, and remember, since this is for an "ideal" venturi of 100% efficiency, that measurement is taken at the inlet diameter, not at the most constrictive point.

The diameter of each barrel of a two barrel carburetor with perfectly efficient "ideal" venturis to meet the needs of a 12a with a maximum RPM of 8400 and a VE of 90% is 1.19 inches, or about 30 mm.

The same equation applied for the theoretical demands of a 12a with a maximum RPM of 10,000, and a VE of 100% results in a two barrel carburetor with venturis 1.37 inches in diameter, or about 35 mm.

Now everyone knows from personal experience that this theoretical measurement is very far from reality. Most venturis are far from "ideal" simply due to manufacturing constraints, and then of course, there are the many necessary "extraneous" components I joked about earlier that impede air flow. These variables may seem to make the above equation seem rather useless for picking out a carburetor by it's cfm rating, but that's not what it's intended to be used for. Instead, we can apply this same "ideal venturi" theory math to carburetors with known flow rates and venturi cross sectional areas, and calculate their efficiency. By comparing carburetors with the same maximum flow rating, we can mathmatically determine which choice will yield the higher signal strength to the main circuit, which translates into better low end performance. -Remember, high end performance is mostly determined by the maximum cfm the carburetor can flow because as velocity increases, the internal pressure drop decreases multiplicatively. Even if the carburetor has inefficient venturis, as long as it can slightly surpass the flow needs of the engine, it's nearly gauranteed to be capable of delivering decent high end performance. In nearly every performance application aside of drag racing, of two identical set-ups, the one that also has good low end performance is the one that's going to out perform the other overall.

I'll do this for the stock Nikki, the Sterling Nikki, and the Racing beat Holley 465.

First we measure the venturis to find the combined cross-sectional area. Then we'll plug this number into the following equation derived from our exercise above to find the flow efficiency of the carburetor:

rated flow of carburetor in cfm	= flow efficiency (%)
venturi area in sq. inches x 137.7 cfm per sq. inch

The combined venturi area is multiplied by137.7 cfm per sq. inch to give the maximum flow efficiency through the carburetor. The venturi area is measured at the inlet/outlet diameter because the venturi which we are comparing the real world carburetor venturis to flows at 100% efficiency even though the smallest internal diameter is half the inlet/outlet (largest) diameter.

Carburetor	Venturi Diameters (inches)				Combined Venturi area (inches sq.)	Flow at 100% efficiency (cfm)	Rated or Tested Flow (cfm)	Flow Efficiency
Carburetor	Pri #1	Pri #2	Sec #1	Sec #2	Combined Venturi area (inches sq.)	Flow at 100% efficiency (cfm)	Rated or Tested Flow (cfm)	Flow Efficiency

Stock Nikki	1.09	1.09	1.35	1.35	4.73	651	313 cfm	48%
Sterling Nikki	1.09	1.09	1.35	1.35	4.73	651	465 cfm	71%
RB Holley 465	1.46	1.46	1.46	1.46	6.70	923	465 cfm	50%

The inside diameter of the primary venturis in the stock Nikki measure 20mm, or .79 inches. I increase these to 22mm, or .87 inches, and I also turn the venturis on a lathe to give better inlet and outlet angles. But you can see from doing the math that they are still far from the "ideal" venturi in that the inside diameter is much larger than the half the outside diameter ( ...ideally the inside diameter would be 13.8mm or .55 inches ). As this is all theoretical math, real world experience shows us that even with less than ideal venturi design, at fairly low air velocity the pressure differential created inside the mass produced carburetor venturi is still plenty strong enough to draw adequate fuel into the carburetor for a useable air / fuel mixture. Even so, when the venturi is optimized as much as possible, it does improve performance as well as fuel efficiency. Further flow testing is yet to be done to measure the pressure differential on all three carburetors for comparison.

Also, take note of the fact that the Racing Beat prepped Holley 465 has the same sized inlet and outlet diameters for both the primary and the secondary venturis. The inside diameter, however, is smaller on the Holley primaries. Though this can be considered an illustration of the above paragraph in that even "less than ideal" venturis can pull adequate fuel at a given air velocity threshold, it stands to reason that at least one of the pairs of the Holley venturis is much farther from ideal than the other. Since the venturis in the Holley are cast into place, I cannot remove them to easily measure the inlet and outlet angles. I suppose it wouldn't be too difficult to triangulate by measuring the constriction inside and also how far down the venturi it is. I'll update the chart when I do.

We've discussed the efficiency of venturi design, but we haven't taken into account the rest of the components with regards to how they disrupt and impede air flow. It only stands to reason that anything in the path of the incoming or outgoing air flow through the venturi is going to threaten efficiency, and there are certainly plenty of such components. Probably the worst offenders are the throttle plates at anything other than wide open throttle. But at WOT, the throttle shafts disrupt flow the greatest, followed, at least in the stock Nikki, by the booster venturi support arms. Also present in the direct path of incoming air are the OMP fittings, accelerator pump nozzles, and of course the choke, all of which impede the flow only through the primary venturis. Everything that you can see in the air path will effect air flow, and not in a positive way, either. Even changing the screws that hold the secondary throttle plates on from the stock machine screws to more delicate button-head cap screws had a measurable impact on flow efficiency in our testing.

Carl Perez did all of the flow testing on the development of the Sterling Nikki. I mention on other pages the great lengths that Carl and I went to in our efforts to increase flow through the Sterling Nikki. We found out rather early on that optimizing venturi profile was not even half the battle. We started out with a fairly optimized venturi that was conservatively cut as little as possible so as not to compromise fuel signal strength at low velocity. We then slimmed down everything that was impeding flow by clipping the extra booster supports and putting an airfoil profile on the remaining ones, cutting back the OMP lines that hung over the edges of the primary venturis, and milling both throttle shafts as thin as possible while still retaining a reasonable amount of strength. The result was a carburetor that flowed approximately 425 cfm with great low end power. As I'm convinced that it's plenty enough for even a heavily ported 12a, I could indulge myself here by plugging what I consider to be my best carburetor achievement into the above equation, and I get a flow efficiency of only 65%. But alas, Carl and I agreed that impressive numbers is what sells carburetors, and if we couldn't catch people's attention with a cfm rating that was competitive with the least expensive after market performance carburetor choice for the 12a, the Racing Beat prepped Holley 465, our little cottage industry would falter, and all our efforts would be for nothing.

So I compromised the venturi design to make the Sterling flow the same as the Holley 465. Luckily, after many different cuts and tests, I was able to maintain the great low end of the Sterling 425. Through those testing stages, some early carburetors were shipped out with one of three different cuts, all yielding the same decent low end pull, but with different maximum cfm ratings. There was the Sterling 412, 425, 450... until we hit that magic marketable number, 465. That's the same cut I currently use, and have used for the past several years.

A carburetor can only flow as much as the rest of the system will allow. The most restrictive part of a complete engine system is referred to as the "bottleneck", and for stock and modified Nikki owners, the bottleneck is most certainly the stock manifold. The flow rate of the stock manifold is downright terrible compared to the flow capabilities of the carburetor. Some later models, with the anti-after burn valve installed flow as little as 288 cfm. There is great room for improvement in the stock manifold, even with mild porting just to remove the obvious flow impediments. The Racing Beat cast Holley manifold, on the other hand, is an absolutely beautiful free flowing design, and there's no doubt just by looking at it that it will deliver unimpeded high end performance. But the Holley carburetor is unfortunately not very optimally designed for low end performance, where the rotary engine really needs help, while the Sterling Nikki, flowing the same maximum CFM, is very well designed for low end performance but the manifold keeps it from reaching it's high end performance potential. Both carburetors are capable delivering the same high end performance, as they both exceed the maximum Mazda rotary induction needs. With a better manifold, the Sterling Nikki would plant the Racing beat Holley, hands down. Until the time when a manifold is designed and produced, it is only a "contender". - I am working on that, by the way. This is a stage I plan on focusing my efforts on this coming season (2009).