NOTE: Due to the heavy telephone and e-mail traffic of our spring season, our technicians will not be fielding any general technical questions related to the data in this document. We appreciate your cooperation. Group K
The Market -The future emission mandates for pwc after 2000 will likely mean the use of fuel injection and direct injection on many future pwc models. While these systems are emissions friendly, they cannot be easily retrofitted onto older design engines. Even if these systems could be fitted, the fuel delivery "mapping" programs cannot be altered or adjusted. This means that many owners of older pwc models (and modified pwc's used for competition of all sorts) will be seeking carburetors of improved designs (that are adjustable) for their machines.
There is certainly no shortage of "high performance" pwc carburetors on the market. Unfortunately there is a stark shortage of ways to compare the technical abilities of all these carbs, and even less comparative information with respect to those abilities. In light of the considerable costs of these various racing carbs, consumers (and engine builders) would have every right to desire "a lot" of information about them. This document will attempt to offer some of this absent information. Along with this information, we will attempt to outline some standardized evaluation procedures for individuals wishing to make relevant comparisons of their own.
There are too many brands of aftermarket pwc carburetors to include all brands in this document. We will attempt to include the most prevalent brands of the various design styles. Group K is not a distributor for any aftermarket carburetor brand, so we have no vested interest in favoring any brand. It is not our intent to promote any one brand, but rather promote the desirable abilities that might make one brand more desirable than the others are for a particular application. This document will not attempt to declare a "winner" the laws of physics does that.
Basics - As any engine builder will tell you, an engine is basically a big air pump. "The more air in and out...the more power it makes." Thats simple enough...why not just install huge carbs to let gobs of air into the engine? The reason...we have to get fuel in along with that air. There in lies the BIG problem. Air is light, and fuel is heavy. This means that we can only add air as long as we can bring the correct amount of heavy fuel along with it. This makes the job of good carburetion infinitely more complex. Getting the carburetor to always draw the correct amount of heavy fuel, along with the lightweight air, is the design goal of every carburetor. Getting a carb to do this job is actually not too tough. Evaluating how well the carb is doing that job...thats a lot tougher. Among other things, the following text will carry definitions of terms, and descriptions of the variables, that help engineers evaluate how well a carburetor is doing its job.
Signal - During the intake cycle of a pwc engine, the reeds (or rotary valve) will suddenly expose the strong low pressure (aka vacuum) in the crankcase, to the atmospheric-pressure column of air in the inlet tract. In essence, the crankcase is creating a negative pressure "signal" that literally sucks the air through the carb throat in to the crankcase. This moment of negative pressure is referred to as the "engine signal".
At the same time that this negative pressure is pulling air into the crankcase, the air traveling across the fuel circuit openings in the carb also create a negative "signal" that pulls fuel out of the carburetor. This is "carburetor signal", which basically means the amount of signal that the carb circuits can "see". The design of the carb throat, and fuel atomizer hardware, play the primary roles in how much "signal" the carburetor circuits see. The degree of carb signal that the jet circuits "see" predicts how much of the heavy fuel can be pulled along with the passing air. If a carburetor does nothing at all, it "must" have good carb signal. For simplicity, engineers refer to both the crankcase negative pressure signal, and the carburetor observed signal, as just "signal". Even though they are different things, its assumed that the meanings apply to the particular item your discussing at the time.
The Flow Bench - Most folks figure that everything you need to know about a carburetors function, can be told by a flow bench. This is mostly true, depending on the added apparatus you use with the flow bench. Here is a very basic explanation of flow bench function and vocabulary.
A flow bench is basically a flat table with a hole in the top face, with a variable pressure vacuum blower directly underneath. Any object placed over the table opening (like a carburetor) offers resistance against the air vacuum from under the table. A "probe" connected to an upright (water-filled) manometer shows the level of resistance (or lack thereof) that the carb offers against the vacuum. In a conventional battery of flow bench tests, where different carbs are being tested, the variable air pressure valve is re-adjusted for each carb until the manometer reaches a standardized specified water level. This water level, in the upright manometer tube, is commonly referred to as "inches of test pressure". The resistance of each carb being tested is reflected by the water level in another separate angled glass tube. This tube, also mounted to the bench, is called a "flow scale". The Flow scale is angled so that its easier to observe subtle changes in flow during a battery of tests. The Flow scale indicates the percentage of flow change during a battery of tests. The results of this flow scale reading are calculated (with the total flow reading) to yield a "CFM" number. CFM is cubic feet per minute (of air) being passed through the test carb.
Theoretically, the relative data from any flow bench should match the data from the same test done on any other flow bench...providing that the original test pressure is set at exactly the same value (in inches). Here in lies the problem with flow bench testing and claimed CFM numbers...there is no standardized test pressure for pwc carbs. The (generally accepted) worldwide SAE (Society of Automotive Engineers) test pressure standard for automotive applications is 20 inches of water. So, why doesnt "everyone" simply do their flow bench tests of pwc carbs at a test pressure of 20 inches of water? The reason is simple. A flow bench that has enough power to pull 20 inches of vacuum through a 50mm hole is one hell of a big (and expensive) flow bench. As a result, many folks have opted to purchase a smaller less expensive flow bench, and conduct their tests at pressures between 5 and 15 inches of water (depending on the power of the blower motor on their flow bench). The data gathered at these lower pressures can still very valid information for the purpose of discovery and development testing of a particular unit. However that data just doesnt correlate to any of the standard 20-inch measurements that everyone else has. This scenario of testing below the SAE standard pressure, has become so common that many flow bench users have adopted an "unofficial" standard pressure of 5 inches. Thats because virtually any flow bench can pull 5 inch vacuum level through a 46 - 50mm carb throat.
Full Throttle CFM and Signal Measurement - Obviously, the CFM ability of any given carburetor is important data. As a general rule, the best way to increase CFM is to simply make the carb throat larger. However, as previously mentioned, all that incoming air is worthless if there is no additional fuel being pulled in along with it. Higher airflow (cfm) always comes at the expense of lower air velocity (i.e. energy), and lower carb signal. This is the time where a well-equipped flow bench becomes an invaluable tool. A flow bench, with the correct attachment hardware, can be equipped to take accurate measurements of carburetor signal ability.
For a "signal" test, vacuum tubes are screwed into the threaded main jet and pilot jet holes of a carb that is already mounted on the flow bench. These two vacuum tubes are then connected to individual vertical flow scale tubes (also filled with water). Then the flow bench is turned on and brought up to standard pressure (preferably 20 inches). As the air passes through the throat of the carb, the flow scale tubes, screwed into the jet holes, will indicate the exact amount of negative pressure that is available to pull fuel from the individual jet circuits. Higher carb signal numbers mean that fuel will be more swiftly supplied when the signal comes from the crankcase. Lower carb signal numbers mean that the passing air is less likely to draw the correct amount of fuel, on its way to the crankcase.
This test, performed with the butterfly 100% open, is the most common procedure for measuring and comparing carburetor signal. For this reason, our data in this document will deal specifically with these measurements...and only these measurements.
Partial Throttle Signal Measurement - As stated above, the afore mentioned test is typically performed with a 100% open butterfly. At this setting the CFM, and the demands on the fuel circuits, are at maximum. This effectively portrays a worst case scenario where air delivery ability will be at its best, and fuel delivery ability is most crucial. If the flow bench technician were to slightly reduce the butterfly opening during the course of this test, the CFM numbers would drop significantly, and the signal numbers (i.e. fuel demand) would change significantly.
As the butterfly position alters the CFM and signal numbers, it also alters the respective signals of the main and pilot circuits in relationship to one another. That is, the percentages of total fuel being supplied by the main and pilot circuits is constantly changing. There are openings where the pilot circuit has more signal (and more fuel delivery) than the main circuit, and vise versa. This means that partial throttle measurements are infinitely more complex to accurately document, and even more difficult to compare (from one carb to the next). As a result of all this "gray area" complexity, there is no standardized procedure or protocol for partial throttle signal measurement.
This doesnt mean that no one does partial throttle testing, it merely means that "nobody" agrees on an effective and accurate way to make these comparisons. We will be the first to concede that partial throttle signal is "very" important. We will also concede that not "all" carbs have strong and linear signal curve through the full range of butterfly openings. This truth makes partial throttle signal data even more important.
Given the rats-nest of variables and expert disagreements related to partial throttle signal testing and comparison, we will leave that subject for the racing carb makers to sort out. We think its an important subject, but we also think its a lengthy subject that is best left for another document written by someone else.
A Note About Signal and CFM Data - The effects of flame arrestors will be discussed later in this document. For now, suffice it to say that installing a flame arrestor (during a flow bench test) will improve signal numbers and decrease CFM numbers. On this same tack, installing a "somewhat restrictive" flame arrestor will increase the signal numbers "a lot" and reduce the CFM numbers "a lot". It is a common practice for carb makers to conduct their flow bench "signal development" work with a flame arrestor mounted on to the carb being tested and thats fine. However once all this development is done, some makers will remove the flame arrestor for the purpose of getting "their CFM" numbers. This practice would obviously not offer a valid representation of the carburetors real "signal and CFM" abilities.
But if we want to talk real, all pwcs must be operated with a flame arrestor mounted...and most "standardized" flow bench tests are performed with no arrestor at all. In the interest of doing "real" tests, we considered fitting a flame arrestor to all the carbs we tested. Unfortunately, not all carbs will work best with the same arrestor. After much consideration, we felt that installing flame arrestors for our flow bench tests would introduce too many potential variables. Since the generally accepted standard "20 inch" test is performed with no arrestor or filter in place, we decided to do the same.
Testing the net effects of different arrestors, on different carbs, could certainly uncover a lot of interesting information. However we consider that to be another lengthy subject that easily merits a document of its own. In the meantime, we would ask readers to be wary of data from other sources that may be presenting "arrestor fitted" signal data, along with "non arrestor" CFM data.
About Fuel Flow Meters - Flow benches can be very effective tools for the testing and development of signal. However most technicians agree that flow benches can only do a "reasonable job" of portraying signal through the full range of butterfly openings...not a perfect job. These same technicians agree that the only "real" way to plot the signal strengths (or weaknesses) of a carb is by using an on-board fuel flow meter. Fuel flow meters cannot actually show a numerical value for signal. However they do show the real-world fuel consumption rate, at every throttle opening, that is the result of signal. These digital meters show the exact amount of fuel being consumed at every rpm throughout the range. This fuel flow testing is superior to flow bench testing because it includes the real world effects of exhaust system resonation, crankcase pressures, inlet tract resonation, flame arrestor restriction, etc.
This fuel flow meter testing is what we use at Group K to identify signal "difficulties" of various inlet system components. Such testing allows us to quickly establish the differences between jetting related problems, and signal related problems. Our experience, with respect to signal, has been that of observing its results with our flow meters. During these tests, we have had test boats that exhibited very significant "weak spots" in partial throttle signal that were very difficult to duplicate during flow bench testing. This is an additional reason that we tend to give limited credibility to "non-standardized" partial throttle flow bench signal testing that indicates "no dips in signal strength". However our full throttle on-water flow meter tests have always mirrored the results of "standardized" full throttle flow bench signal testing. This unsolved inconsistency between partial throttle flow bench tests, and our partial throttle on-water tests, was another important reason for us to portray only 100% open throttle signal data in this document.
The only down side to on-water fuel flow testing, is that it must be done on pure glass water. This insures that the loads are consistent, and as high as possible. This stable and heavy loading causes the engine to create maximum signal at all rpms, and portrays the highest fuel demand scenario. As soon as the driveline encounters the interrupted load of rougher water, the engine signal (and thereby fuel flow) will drop at any given rpm.
Unfortunately, this means you have to be testing on one hell of a big piece of smooth water. Running along at 60+ mph while viewing the relationship of digital fuel flow and tachometer numbers is a style of riding that requires great "caution"...and lots of safety margin space. At Group K, we are lucky enough to be near a part of the Colorado River that offers about 30 miles of uninterrupted glass water most every day, hence we conduct "lots" of on-water fuel flow meter testing there.
BSFC and "Going Rich" - Every pwc/motorcycle rider has described the high rpm "end" of the power band as "falling off the pipe", or "falling off the power band". Modern dyno technicians describe this same phenomenon as "going rich". While the riders "seat of the pants" description of the end of the power band is easy to relate to, the dyno-technicians description is much more accurate, and much more telling of possible solutions.
Most of todays modern dynamometers are equipped with the hardware/software to record "Brake Specific Fuel Consumption" data. In laymans terms, BSFC boils down to how many horsepower is being produced per pound of fuel consumed per hour (at a particular rpm). As an engine on a dyno gets beyond the power peak, the hp numbers start to drop sharply while the fuel flow remains very much the same. A rider would feel this loss of power, and say the boat is "falling off the power band". In this same situation, a BSFC equipped dyno "thinks" the power is dropping off because the fuel flow is too great for the power being generated, or that the mixture is simply too rich (hence the term "going rich"). Initially it sounds stupid to say that an engine (well past its power peak) is "going rich". However there are carburetor and exhaust system design techniques that can slightly raise the rpm at which an engine goes rich. On the water, these techniques result in an engine that revs slightly farther before the power band ends. One such technique is to change the design of the carburetor atomizer in such a way that it is "less effected" by the unfavorable resonations connected with the "going rich" phenomenon.
At this point, its important to note that the phenomenon of going rich not just about carburetion. "Going rich" is connected to a collective of pressure waves and resonations of the tuned exhaust, crankcase, and entire inlet system. The lengthy description of how all these related components are involved in the "going rich" phenomenon is easily complex enough to be the subject of another separate document. Rather than engage in all those explanations here, we will try to stay within a part of the subject that relates to the carburetors.
As a particular engine arrangement reaches its horsepower peak, crankcase signal and fuel efficiency are at their peak. However as the engine passes that HP peak (as virtually all engines do on the water), the signal from the crankcase becomes somewhat erratic, and less "observable". The result is that the engine, revving beyond its HP peak, receives an excess of fuel for the power being generated. On dyno BMEP printouts, this shows up as a sharp upturn in fuel-consumed per horsepower-produced, or becoming excessively rich for the power being generated. On the water, this same phenomenon allows a boat to accelerate strong up to peak power, and then "nose in" badly as the point of "over-richness" takes place. Leaning out the high-speed screw might slightly abate the nosing-in at "past HP peak" rpms, however that means you would be dangerously lean at the true HP peak.
The whole concept of "going rich" is important because particular carburetor atomizer designs can be less effected than others by the "unfriendly" signal coming from the crankcase at rpms beyond the horsepower peak. A carburetors ability to meter accurately, while receiving a less than ideal signal from the crankcase, can allow the engine to rev slightly farther before the "going rich" phenomenon causes the boat to "nose in". We have observed that carbs with stronger full throttle signal tend to have this ability to meter more accurately under the conditions that induce "going rich".
Self-Compensating Characteristics - Anyone that has tuned a high performance watercraft knows that pwc engines require measurably more fuel when they get the uninterrupted load of glass water, as opposed to the easier "interrupted" load of riding through rough water. Unfortunately, rough water riding is a series of heavy-load/high fuel demand moments, separated by a series of no-load lower fuel demand moments. If a carburetor does not compensate instantly for the lower fuel demands of the unloaded moments, it will literally "go rich" when it is unloaded. In the next instant, when the engine is under load again, the excess of un-used fuel at hand induces a rich condition that results in a "landing with the brakes on" kind of feeling. During rough water operation, the carburetors are constantly being asked to react instantaneously to these varying fuel demands. Carburetors with strong signal, and atomizer designs that are more resistant to the resonations of going rich, have no problems reacting instantly and accurately to these constantly changing fuel demands. A carb with weak signal strength, and/or an atomizer that is less resistant to resonations of going rich, still offers some correction of mixture, however its range of correction ability will be significantly narrower. That means that the margin of error (for low speed and high-speed screw setting) would be much narrower on a carb that has the weaker signal. Using a carb of weaker signal, a tuner would be forced to seek the safety of a slightly richer setting for loaded operation, and simply accept the over-rich condition that would exist during unloaded operation. Having to make this kind of compromise increases the occurrence of "going rich" moments during high speed rough water operation.
Mechanical Fuel Injection - Within the pwc community, the only aftermarket fuel injection system that has stood the test of time is the one offered by MRD. This mechanical injection system has an advantage over all carburetors in the sense that it does not depend on "carb signal" to conduct its metering work. Hence, MRD owners can use huge venturi throat diameters with little concern of harming "single point" metering. Single point metering refers to the amount of fuel being delivered at one particular load level, and one particular rpm. The MRD fuel delivery curve is strictly rpm based (driven by a fuel pump at the crankshaft), and allows the owner to make various range adjustments via a series of adjuster screws and pop off valves. Many pwc owners (grudge racers in particular) have had very good peak speed results with this fuel management system on many different engine platforms.
Unfortunately, all this simplicity comes with a few important compromises. The first, and foremost, is that the MRD system has absolutely no ability to deliver less fuel to an unloaded (out of the water) engine that is offering a weaker signal. The result of this inability to self-adjust (as all conventional carbs do) is a very noticeable over-rich condition when operated on rough water conditions. This same phenomenon contributes to the MRD's very high rate of total fuel consumption (higher than virtually any conventional design carburetor). On the water, this inability to correct fuel metering for unloaded operation can lead to a "landing with the brakes on" feeling as you run through rough water with high throttle openings. MRD owners, who ride on rough water, are forced to make a compromise of the mixture that might be ideal on glass vs a mixture that does not foul the plugs during unloaded operation. With such a compromise setting, the risk of detonation, or piston scoring, is greatly increased if you suddenly find a big piece of glass water. Furthermore, mechanical fuel injection will not correct for varying air density conditions (as a strong signal carburetor can).
All this said, the MRD system can work well so long as you are willing to regularly fine tune it for any changes in the prevailing water and weather conditions that you will be riding in. While the MRD system has little in the way of self compensating abilities for mid range operation, it "can" be easily tuned to meter the correct amount of mixture for full throttle smooth water operation (this is called "single point" operation). Because of its ability to work very well under single point operation, the MRD systems have been successfully used by many grudge racers who ride only on smooth water conditions (where no compensating abilities are ever required).
From our perspective, this system can work well for individuals that operate on glass water conditions in areas that do not experience significant swings in air temperature or air density. While this might be fine for some owners, most of the high performance pwc owners we deal with are seeking a fuel management system that is more versatile, and more forgiving to adjustment.
About Atomizer Designs - Among the biggest design differences between all the aftermarket carburetors is the apparatus used to introduce the fuel into the passing air stream. These devices are called atomizers because their main function is to introduce the fuel into the air column as an "atomized" spray.
By far the most common atomizer design is the "booster venturi" design (used in most production pwc carbs). Many folks (us included) have referred to the booster venturi as a "bomb-sight" atomizer because it has the appearance of a gunners bombsight.
The booster venturi gets its name from the circular bell-mouthed venturi mounted at the top of a main jet fuel outlet. This venturi helps to boost the negative pressure signal over the main jet supply orifice (hence the term booster venturi). We have been reminded, by the carburetor folks, that nothing inside any carburetor is being "bombed", so they would really prefer that we use the technically correct term, booster venturi...fair enough.
Another popular design is the "atomizer tube" which is a brass bar that replaces the traditional booster venturi. There are variations in the shape and fuel outlet orifices among different aftermarket atomizer tube designs, however the basic concept of all these variations is the same. The atomizer bar represents a much smaller obstruction to airflow than the traditional booster venturi. Unfortunately the atomizer bar design does not have any additional device to induce strong signal directly over the high-speed orifice outlet (as the booster venturi has). One other design that has experienced intermittent popularity among pwc competitors is the "annular discharge" atomizer. These carbs introduce fuel (for the main jet) via a series of orifices around the full diameter of the carb throat. The intent of this design is to offer better performance than other designs because of superior fuel atomization abilities.
A Note About Fuel Atomization - Many of the aftermarket carb makers go to great lengths to promote the exceptional fuel atomization abilities of their carbs. The "real" priority here is the need to deliver well-atomized and well vaporized mixture "into the combustion chamber", not the crankcase. In truth, the mixture being admitted into the crankcases (no matter how well atomized it is) goes through a separation process before making its trip to the combustion chamber. High-speed internal photography of engines in operation has revealed that the spinning crankshaft tends to throw the heavier fuel to the outside of the crankcase, while the lighter weight air remains within the center area of the crank case. The fuel in the crankcase is in a constant state of being atomized by being thrown off the outside of the flywheels (in the same way that water spirally sprays off a road tire). This process, which represents the biggest percentage of fuel atomization and vaporization, is performed in the crankcase and transfer port passageways...not in the carburetor.
The inside of the crankcase of a high output two stroke is best described as a miniature hurricane of biblical intensity. The intense turbulence of this hurricane does a great deal to atomize the fuel being delivered for the next cycle. The walls of the transfer passages can further enhance atomization if their surface finish is rough enough to disperse any "un-atomized" fuel droplets that "make contact" on their high-speed trip to the combustion chamber. This is why Group K porting includes a mild "rough-finish" on the walls of the transfer passages...(instead of polishing). This helps to insure that all fuels induced into the crankcase, by the carb, will experience significant atomization during its millisecond whirlwind journey to the combustion chamber.
For sure, it helps matters slightly if the fuel metering apparatus of a carburetor can assist in atomization, however (on a 2 stroke pwc engine) the impact of that help has more to do with fuel economy/efficiency than sheer power output. The best example of this point, is the MRD mechanical fuel injection system. These systems perform virtually no atomization at all, yet still yield very good single-point full-throttle power. The lack of atomization of the mechanical fuel injection systems does contribute to a very significant wasting of fuel. But the MRDs, nearly non-existent, atomization does not have a measurable cost in the area of lost power.
The Popular Carburetor Designs - There are several different pwc carburetor throat/atomizer designs in todays marketplace. This document will not go into the history or evolution of these designs, but rather their sheer technical function. There are some brands that are very similar in basic design, so we will discuss the most popular of each design style.
In alphabetical order by brand name they are:
AMC - butterfly diameters 44/46/48 mm - Annular discharge carbs made from a Mikuni Super BN. They feature changeable throat inserts, and employ an integral flame arrestor spigot. Used with integral or remote fuel pumps. These carbs are made (and primarily available) in Japan.
BMR Full Spectrum (not to be confused with Buckshot BRM) - butterfly diameters 46/48mm Booster venturi carbs, manufactured from billet, have integral flame arrestor spigots and are designed for use with a remote or integral fuel pump. The BMR employs a booster venturi design atomizer that has several fuel outlet orifices around the venturis inside diameter (ala annular discharge).
Buckshot BRM butterfly diameters 45/47/50mm - Booster venturi carbs, manufactured from billet, have integral flame arrestor spigots and are designed for use with a remote pump. The BRM employs a booster venturi design atomizer that has several fuel outlet orifices around the venturi's inside diameter (ala annular discharge)
Buckshot Magnum - butterfly diameters 44/46/48 mm - Atomizer tube carbs made from a Mikuni Super BN. They employ an integral flame arrestor spigot and are used with integral or remote fuel pumps.
Keihin 42 CDK II butterfly diameters 42 mm - Atomizer tube carb with integral pump and flat top-face that accommodates various bolt-on flame arrestors.
Mikuni Super BN - butterfly diameters 38/44/46 mm - Conventional booster venturi carb with integral pump and flat top-face that accommodates various bolt-on flame arrestor spigots. These carbs are original equipment on many Yamaha, Polaris, and Tigershark pwc's, and the base unit that many aftermarket carb makers start out with.
Mikuni "I" Body - butterfly diameters 38/40/44/46 mm - Booster venturi carb with integral pump and flat top-face that accommodates various bolt-on flame arrestor spigots. These carbs are original equipment on all Sea Doo 720, 785, 951 pwc's, as well as the Yamaha GP800 and XL1200 Limited.
Novi Maxflow - butterfly diameters 42/44/46/48/50 mm - "Sonic" booster venturi carbs made from a Mikuni Super BN. They employ an integral flame arrestor spigot, a full length tapered throat, and are used with integral or remote fuel pumps. Most Sea Doo sets are supplied with a built in vapor separator.
Red Top - butterfly diameters 38/44/46 mm - Annular discharge carbs made from a Mikuni Super BN. They employ an integral flame arrestor spigot and are used with integral or remote fuel pumps.
Comparing Signal Statistics - Having made all these different statements about the importance of signal, the next logical thing to do is post the actual signal data (as measured on a flow bench). Nearly each of the aftermarket carb makers produces (and strongly promotes) the 44/46mm butterfly versions of their line. Given this, we considered it most relevant to list the data for these diameters. The only exceptions were the Buckshot BRM (they promote the 47mm instead of a 46mm), and the Keihin 42mm (the 42 is the largest race carb that Keihin makes).
The following data was gathered on a Super Flow 600 flow-bench set at 20 inches of water for all tests.
Along with this data, we have also denoted the full throttle CFM numbers, as measured at 20 inches of water. They are listed in alphabetical order by carburetor brand name.
Carburetor (signal@100%open) Main Circuit Pilot Circuit CFM @ 20" (100%open)
Buckshot Magnum 44 31.0 23.3 268
Keihin CDK II 42*** 28.8 22.5 195
Mikuni Super BN 44*** 49.0 30.0 230
Mikuni "I" Body 44*** 32.2 22.5 239
Novi Max Flow 44 66.0 25.7 269
Red Top 44 37.9 20.3 269
AMC 46 31.7 21.5 293
BMR Full Spectrum 46 58.0 23.0 299
Buckshot BRM 47 41 - 47.5** 20.5 293
Buckshot Magnum 46 37.2 26.0 286
Mikuni Super BN 46*** 46.0 34.8 253
Mikuni "I" Body 46*** 33.0 23.2 246
Novi Max Flow 46 65.0 28.0 304
Red Top 46 34.8 21.0 296
** The signal reading randomly fluctuated on a 2-4 second cycle within this range, with no stabilized 20" reading lasting over 5 seconds. Signal reading would not stabilize until pressure was reduced to 8".
*** Tests were run with bolt-on arrestor adapter.
The Obvious Question - "How can some racing carbs have so much less signal than others and still work?" This is a very reasonable question. The answer is, a racing carb does not need to have strong signal in order to "work". If your are a top flight tuner that constantly monitors and adjusts mixture, carbs of weaker signal can be made to perform "almost" as good as carbs with considerably stronger signal. It merely requires that you manually perform a lot of the metering compensation that "strong signal" carbs perform on their own. Since strong signal carbs have the ability to correct and compensate over a wide range of extreme air density and load conditions, they are much more "forgiving". Weak signal (within limits) does not mean weak power, it simply means weak fuel delivery abilities and weak self compensating abilities...strong signal means strong metering compensating abilities (weak signal can also result in less than optimum throttle response). Having carbs with very strong signal is sort of like getting carbs that have a built in full time tuner.
The Next Obvious Question - "If I have carbs of weak signal, is there any way I can improve the signal strength?" This is another reasonable question...and the answer is "yes". Unfortunately, the means of "improving signal" are often a "give and take" proposition (where you often give up more than you get).
A Great Example of Improving Weak Signal - During the development of our Sleeper kits for the single carb 720 Sea Doo, we got a swift education in the weakening and strengthening of carb signal...and thereby performance.
The rotary-valve Sea Doo pwc engines tend to have less engine signal than similar displacement reed valve motors do. Along this same line, smaller displacement engines tend to send less signal than larger displacement engines. This made our single carb 720 a motor that doesnt give the carb much signal to work with. In addition to this collection of black marks, this particular 720 motor has a "2 into1" inlet passage arrangement. That means that the signal from the engine is further weakened by the fact that there is double the inlet port passage area of a normal inlet tract. Despite all these negatives, the stock inlet "system" has fairly good signal and great metering stability at all rpms. We were able to verify this good signal strength with our on-board fuel flow meter, which showed a smooth increase in fuel flow that mirrored the increases of butterfly opening.
Among our first tests on the 720 was to modify the carb throat (as we have done on countless dual carb 785s with great success). During the first tests with this modified carb, we opted to leave off the stock air box lid and reduce the number of screens in the stock arrestor element. We immediately found that the overall acceleration of this setup was very good. However every time we applied full throttle, the fuel flow numbers were normal for a moment...then dropped off dramatically (compared to the stock setup). Assuming we had a lean condition, we jetted up. This jet change dramatically increased the fuel flow numbers at 70-80% throttle. But as soon as we drew 100% throttle, the fuel flow numbers again descended, creating a nearly fatal lean condition. Installing the stock carb (with the same modified arrestor setup) did not fix the problem. The stock carb now appeared to have the same fatal full throttle lean condition.
After numerous tests of all kinds, we installed the completely stock flame arrestor element and case on to our modified carburetor. Immediately, the engine carbureted great, and our full throttle fuel flow numbers stabilized (with the smaller main jet we were using earlier). In addition to this, the performance of the modified carb was now clearly better than the stock carb on the same setup.
With hindsight in hand, this 720 taught us a few important lessons.
#1 Installing a free-er breathing flame arrestor is only a benefit for inlet systems that have lots of engine signal and carb signal "to spare".
#2 Bigger jetting has absolutely no impact on a lean condition caused by weak signal. The lean spot is a function of carburetors inability to deliver the fuel at a particular butterfly opening, not a function of the engines inability to draw the fuel from the carb.
#3 Restricting the air inlet, on the flame arrestor side of the carb, increases the intensity of the negative pressure in the inlet tract, and thereby increases the signal at the carb.
Armed with these lessons, we revisited some other "large carb" engine setups that had shown similar difficulties with metering stability at full throttle. As a quick and dirty test, we merely taped off 50% of the air inlet surface of the free breathing flame arrestor elements on these setups. In several cases, our fuel meter immediately showed a stronger and more stable fuel delivery at full throttle, along with improved peak rpms. In short, we could successfully improve the signal on most of these setups by slightly restricting the air supply. As these tests progressed, we learned that this air inlet restriction was only a benefit for inlet systems (and/or carburetors) that had weak signal. As a general rule, this applied to "large carbs" of weak signal. Except for the single carb 720, we never experienced a benefit by restricting the air to setups that used a smaller, more conservative, carb throat diameter.
Good Example #2 - During the 95-96 racing seasons, one of the most popular inlet systems around were the Buckshot Magnum carbs. These carbs were widely used on many different engine platforms with good success. At the same time, Ocean Pro was marketing their "Vortex" flame arrestors for use on race machines. Ocean Pro got endorsements and rave revues from many racing engine builders, for the great performance abilities of these Vortex arrestors. Given all this attention, we wanted to see if the Vortex's would net these same improvements on our engine kits. We tried the Vortex flame arrestors on several of the Sleeper and Hammer kits that we sold to racing customers. All our kits, of that time, utilized 38mm or 40mm carbs (that have very strong signal). In every one of our on-water tests, the Vortex flame arrestors offered considerably less acceleration and less peak rpm than the free breathing K&N type arrestors that we were using at the time. We contacted some of the other engine builders that had claimed such great results from the Vortexs, thinking that we might have been missing some tuning detail. These competent tuners all reasoned the same as we did (at the time)..."If the Vortexs can make a noticeable improvement on Buckshot 46s, they "have" to be able to work on 38s!"
Based on our more recent testing, we believe that the Vortex arrestors were offering restriction to the Buckshot 46s that made for significant improvements in signal. These other reputable tuners were correct that the Vortexs were the hot setup...for carbs with relatively weak signal. We believe that the Vortexs never benefited our engine sets because the smaller carbs we chose to use required no such signal increase. This is why the smaller "strong signal" carbs, of our kits, reacted best to the free breathing flame arrestors that we were already using.
All this information begs the next obvious question " Which is better, a smaller "strong signal" carb with a free breathing arrestor...or a larger "weaker signal" carb whose signal is being improved via a more restrictive arrestor?" The only way (we see) to accurately answer that question would be with an extensive battery of signal/CFM flow bench tests involving different carbs, and arrestors of varying "restrictive" levels. To be sure, a battery of tests like this would be a tremendous amount of work, and would merit another document of its own.
Having said that, we do have our own opinions about this small carb vs. big carb comparison, based on our own on-water tests. From a fuel economy standpoint, the smaller carb is clearly better. From a raw acceleration standpoint, the larger carb is often better. From a peak rpm standpoint, there is disagreement. Our tests lead us to believe that the stronger signal of the smaller carb makes it much less likely to go rich at high rpm, thus allowing better rpm peaks. For this to happen, the throat diameter of the smaller carb must be carefully chosen to assure that it has enough CFM ability to feed the air needed for any particular setup.
Based on our on-water tests, the larger (weaker-signal) carb is much more likely to prematurely go rich at full open throttle. The only way to stave that off, would be to further restrict the incoming air in an effort to increase carb signal at 100% open throttle. In time, the CFM reduction (of the big carb) caused by this air restriction would simply net the same performance and, CFM ability, of the smaller carb with a free breathing arrestor.
Signal Example #3 - Like many other pwc performance shops, we were very excited about the performance prospects of the Sea Doo 951 engines. In fact, the 951 engines reacted exceptionally well to all the normal modifications. Among the most unique features of the 951 motor are the 46mm Mikuni "I" bodied carburetors. These carbs are the first (pwc) carbs that have been developed as much to satisfy emissions mandates as performance needs.
By re-positioning the booster venturi farther from the butterfly (as is done on these 46s), it is possible to dramatically reduce partial throttle richness (and thereby emissions). In truth, this "richness reduction" happens because of a designed-in loss of partial throttle carb signal (at about 70% opening). Since the stock 951 is designed to perform well with the somewhat restrictive stock arrestor in place, these 951 carbs were constructed to have a lot more of this signal reduction than might have been chosen for a free-er breathing aftermarket arrestor. With the stock flame arrestor in place, the signal (and fuel flow) at the 80% throttle setting is very "acceptable". However when a very free breathing aftermarket arrestor is installed, the air restriction that masks this intended weak-signal spot is gone. That means that there is a dangerous lack of carb signal somewhere between 6500rpm and 6800 rpm. Our on-water tests with a fuel flow meter (on a free breathing 951) showed that the machine could be operated all day long at full throttle with no problems. However when the throttle of the free-er breathing 951 was released into this weak-signal area, detonation and engine shut down were almost inevitable. Our fuel flow meter showed about 59 liters/hour @ 6800rpm, 56 ltr/hr @ 6600rpm, and a lethal 38 ltr/hr @ 6700rpm. We tried changing every jet in the carburetor in an effort to eliminate this "lean spot". However, as mentioned before, nothing solved the problem because it was a signal issue, not a metering issue. Installing the stock flame arrestor restored the 6700rpm fuel flow to a much safer 50 ltr/hr (still a little lean, but not lethal).
While the ranges above and below this 6700 rpm were adjustable by jet changes, the 6700 rpm butterfly opening was not. That meant that we literal had to "jet" the 6700 opening with varying levels of flame arrestor restriction, and then re adjust the ranges below and above to suit that restriction.
It bears noting that retaining the stock choke plates (thereby slightly increasing air restriction) in the stock 951 carbs can help to slightly improve signal in this 70% open area. Given this, many 951 owners have successfully used free breathing arrestors with no problems. However if the choke plates are removed, and the engine has any other significant modifications, the risk of an unsolvable and dangerous lean spot are greatly increased.
This is the reason that we recommend the use of aftermarket ( strong signal) carburetors for virtually any 951 that will be modified for higher performance. Additional modifications like increased compression, rev limiters, and cylinder porting will broaden the rpm range, but they also compound the severity of this lean condition (using the stock 46mm I bodied carbs). As always, we tend to recommend carbs that have the best combination of strong signal along with high CFM numbers.
80% Throttle Fuel Delivery - For many pwc's, the heaviest load (and highest fuel demand) the engine will ever experience is at full speed on glass water. However this is not necessarily the case for aggressive handling machines (like closed course race boats).
At peak water speeds, most pwc hulls are at their highest level of plane. That means the hull has the least amount of water contact surface area (and the low resistance that goes with it) under these conditions. In the early years of pwc's, there were no hulls that could predictably hold a turn a peak speeds. However today there is a wide range of fine handling high performance pwc hulls that can hold high speed turns with ease. These hulls are often the boats of choice for closed course racers. These aggressive turning hulls are able to offer predictable high speed turning because (among other things) they plant a lot of hull-surface contact-area against the water in a high-speed turn. All this additional hull surface drag means a much heavier load on the engine. This load can often be greater than the load of driving the boat at peak speed on smooth water. In addition to this, the load of turning left is much harder on the motor than the load of turning right. This happens because of the waters angle of entry in relationship to the impeller blades rotation. In short, a full speed left hand turn puts a "much" heavier load against the motor than a full speed right hand turn.
Under racing conditions, its very seldom that riders negotiate high-speed turns at full open throttle. More often, they will slightly release the throttle to the 70%-80% range in order to "set up" for the turn. After this, racers will often maintain a constantly varying partial-throttle opening in a high-speed turn in order to manage the radius and exit apex of the turn. During this time, the additional water-contact surface-area loads become so great that fuel demands will be very near what is needed for a full throttle straight line run. If all this is happening in a left hand turn, those loads add further to the increased fuel demands.
All this means that most race machines will experience their heaviest engine loads, and nearly highest fuel demands, while the throttle is being held at 70-80% throttle opening...not at full throttle. This reality underlines the fundamental importance of having "very" strong carb signal within the 70-80% throttle operating range. In fact, an ideal racing carb would have a very small reduction in signal between 100% and 75% throttle opening. This reality also makes it clear how specifically "un-ideal" the 951 46mm "I" bodied carb is for a closed course racing application (given its exceptionally weak 70%-opening carb signal).
In addition to these fuel demand issues, the sheer centrifugal force loads of turning at high-speed further increases the difficulty of maintaining sufficient fuel supply. These loads, by themselves, can be so great that a carb with weak signal will not be able to deliver fuel to the atomizer against the centrifugal force. Carburetors with strong signal have a much easier time of providing the needed fuel against these centrifugal forces.
Calibration Tactics - There is plenty of good information, and good advise, available regarding the adjustment and fine tuning of pwc carbs. In most cases, this adjustment and tuning is very straight forward. However there is a growing number of mechanical factors that are having a profound effect on the tuners "approach toward the carb tuning on a particular machine. The following will be a brief description of some of those factors, and the "calibration tactics" that seem to net good results in those situations.
Vibration - As pwc engine displacements (and peak rpms) are increasing, so too does the issues of high frequency resonating engine vibrations. This problem is exasperated by the fact that pwc engines must be rubber mounted to the fiberglass hulls (otherwise the motors would tear the mounts out of the hull). Most pwc owners have observed that the engines in their machines have noticeably different levels of vibration at different engine rpms. Pwc makers do their best to select a balance factor for the crankshaft that offers fairly smooth operation in the rpm ranges where the engine will spend most of its operating time. Even with the addition of harmonic engine balancers on many models, there will always be a part of the rpm range where vibration will be a bit "nasty".
With respect to carburetion, all this extra vibration tends to represent a problem with pop off pressure, or more specifically the needle and seat. This high frequency vibration will literally shake the float needles away from the seats, resulting in very excessively rich operation at low speeds. This problem can be minimized (and even solved) by the installation of smaller size needle seat. Unfortunately these small seats make for a very high pop off pressure that can cause a hesitation at low speeds. A richer low speed jet can be fitted to cover "most" of this low speed hesitation. However the effectiveness of this calibration tactic relies very heavily on the presumption that the carbs being used have very strong signal on the low speed circuits. If the carb signal on the low speed circuit is anything but exemplary, this approach will leave you with an intermittent low speed hesitation that you will never be able to adjust out. Again, it is impossible to fix a weak signal problem with metering changes, and no atomization qualities can equal the positive effects of exceptionally strong carb signal.
A further spin on this same subject is the changes in vibration resonation caused by the installation of lightweight after market exhaust systems. The heavyweight stock exhaust systems, on many modern pwc's, are mounted very solidly to the motor. All this solidly mounted weight helps greatly to dampen and reduce the effects of high frequency engine vibration. When this weight is removed, in favor of a better performing aftermarket exhaust system, there is a marked increase in the engines resonating vibration. Again, the float needles are very susceptible to being "shaken" away from the seats by this changed vibration. Raising the pop off pressure can be a partial "fix" for this problem, but it is seldom the entire solution.
A part of effecting a solution to these vibrations is by installing an aftermarket fluid balancer (available for many popular models). The installation of a fluid balancer is actually a very good upgrade for any high performance pwc engine, but much more needed on the higher revving race engines.
Another very effective approach can be rotating the carburetors 90 to a position that subjects the float needles to less of the forces that are shaking them away from the seats. This is no small piece of work. This change normally requires the installation of a different inlet manifold and different linkage, but if your desperate enough for clean mid-range carburetion, it may be your only option.
About Air In The Fuel Supply - During the 1998 season, we obtained one of the (much-talked about) on-board detonation sensors. A special sensor washer that is fitted under the spark plugs actuates this instrument. These pressure sensitive washers are connected to an adjustable micro-chip. This adjustable chip allows the sensitivity of the unit to be adjusted from 5% - 50%. That is, the LED warning light flashes on when the total percentage of detonation strikes exceeds the microchips setting. It should be understood that many racing engines experience between 5% - 10% detonation under normal full-load operation. However as the percentage of detonation strikes increases, so too does the risk of overheating and engine failure. We used this instrument extensively during the testing of several of our endurance racing machines. After finding the ideal engine settings to allow for "detonation free" operation, we conducted a series of "extreme condition" tests (with the deto-sensor in place). By far the most interesting data gathered during these tests, were the effects of "air in the fuel supply".
The most common way that air gets into the fuel lines is from sheer agitation of the fuel in the gas tank. That is, when the "foamed" fuel in the tank is pulled into a fuel pickup, that air is the forced through the carb metering circuits. During the moments that this takes place, the engine experiences a brief, but serious, lean condition. This lean condition can easily send any high output pwc engine, being run under load, into lethal levels of detonation. We were able to simulate this condition by switching the fuel valve back and forth between the reserve and standard pickup settings during sustained operation at particular rpms. Each time we switched the fuel valve, the deto-sensor showed a solid 4-8 seconds of heavy full-time detonation. Shortening the fuel lines reduced the time span of this detonation slightly, but we could not completely eliminate it.
The implications of this test explains, for us, many situations where we have observed a high output machine to suddenly seize from an unexplainable "lean condition" after operating fine for many operating hours. Of course, this phenomenon is not new information. However actually "seeing" how lengthy and intense the detonation result is...was very newsworthy to us.
From this testing, we gained a whole new respect for the importance of fuel system "air separators" on competition modified pwc's. For now the only bolt-on air-separators on the pwc market are made by Novi Eng. We installed these separators on several of our 1998 endurance race machines...and the results were impressive. With the separator in place, these machines never again experienced the "low-fuel bogs" that virtually every rider has experienced. In addition to this, these machines could easily negotiate full speed turns with a very low fuel level, and never experience the "low-fuel bogs" that can often take place. It bears noting that these fuel-starved segments of "bogging" operation represent the greatest risks of detonation.
Unfortunately, these bolt-on fuel separators do have a couple of disadvantages. The most important of these is that they cannot be easily fitted to "all" machines. These bolt-on separators require the use of a remote fuel pump (the pumps on the carbs must be removed, and replaced with block off plates). This allows the separator to provide a single source of "air-free" fuel for all the carbs. This additional hardware makes the air separator conversion a bit more costly and complex. However that cost is still more affordable than fixing a scored piston.
The other disadvantage of the separator is that there is virtually no notice before you run completely out of gas. When you run out of gas with one of these separators...you are TOTALLY out of gas. However most competitors would rather get the max performance out of every ounce of fuel they have on board, as opposed to bogging around for the last lap.
Besides the bolt-on separators, the only other available separators are "built-in" to all Novi "Maxflow" carburetor sets. Novi constructs a part they call a "fuel rail" for each of their (Sea Doo) carb kits. This fuel-rail has all the separator hardware already built in. In many cases, these fuel rails can be purchased separately for use on other (non-Novi) types of carbs. It bears noting that a remote fuel pump is still required with these Novi fuel-rails.
As any pwc becomes more heavily modified, it will become less tolerant of "air in the fuel". All in all, we consider these air separators (of either kind) to be a mandatory item for any high output modified pwc.
Some Summary Information - Aftermarket pwc racing carburetors are (and always have been) relatively expensive components. For all the money that these carb makers have asked for their wares, few have offered much comparative data on the abilities of their carbs. For the money these parts cost, we believe that customers are entitled to a lot more information than they have been getting. It is our hope that this document marks the beginning of access to more information.
We have no doubt that the contents of this document will spark much controversy and discussion among high performance pwc owners, and some aftermarket carburetor makers...we invite that, and we think thats great. We also suspect that there will be many folks that will want to refute the data in this document...we think thats great too. We would like to submit that anyone who claims this data to be flawed, had better be willing to provide a more effective means of comparison, based on a more accurate and impartial means of testing. If a better means of testing and comparison exists we would gladly defer readers to it.
There are many "side issues" related to carb function that we intentionally did not include in this document. These issues are not unimportant, however they were unimportant with respect to our primary focus subject of full open carb signal. We would ask readers to strongly consider the writings of others, with respect to the many subjects we have not covered here. However we would also caution readers of others that might claim our omission of those subjects is an attempt to "hide something"...we hid nothing.
It has not been our intention to make this document the last word on high performance pwc carburetors...only the first word. Whatever subsequent information comes forward from other sources, there are a few conclusions that we feel are firm.
1 - Precise and responsive fuel metering is a function of carb signal, not airflow.
2 Full throttle carb signal strength is a function of atomizer design.
3 - Big CFM numbers have questionable value, unless they are accompanied with big carb signal numbers.
4 - Carburetors with strong signal are easier to tune than carbs of weaker signal.
5 - Lean spots caused by weak carb signal CANNOT be altered or affected by richer jetting.
6 For high performance pwc engines, good carburetor atomization is nice, but not mandatory.
7 There is no such thing as a "best" flame arrestor for all applications. There can only be a "best" flame arrestor for a particular design carburetor, on a particular engine package.
[Back to Home Page]