1) Because of slight variations we have observed in piston skirt lengths and crown contours, sweeping a group of pistons on a surface plate with an indicator "does not" yield an acceptable level of measurement accuracy. After much examination we have changed the way we measure and compare these pistons. The new measuring procedure requires only a wrist pin and a pair of dial calipers. With the wrist pin half way in the piston, you should measure the distance from the top ring land on the piston (where the "L" ring seats down against), to the bottom of the wrist pin. This procedure allows for the quick and easy measurement of numerous pistons.

2) The variation in this dimension (between the two pistons) should be no more than .004" (.1mm). If you find yourself stuck in a situation where you must (for some reason) use a pair of pistons that vary, the shorter of the two should always be used in the rear cylinder. This will help to facilitate the slightly lower compression that the rear cylinder prefers anyway.

3) We have not seen a pattern of better performance that comes from using "taller" or "shorter" pistons.

Piston Crown Variations - Our earlier entries on this document have outlined the importance of checking and setting squish clearances with regards to correct thickness and cylinder offset. During the recent assembly of a customer's race engine, we uncovered another squish clearance variable that overshadows all our previous information. On this particular engine, we went to great lengths to assure that the parts were prepared in a way that would assure absolutely even clearances throughout. However the final assembled inspection showed one cylinder to have .006" more squish around the entire diameter. On a stock engine, .006" might be an acceptable variation. However on this high rpm race motor...it wasn't acceptable. We disassembled and re-measured all the parts...everything was perfect. The only variable that we had not taken into consideration was a possible variation in the crown height of the pistons. Piston specifications (including crown heights), in all pwcs, are typically held to such close tolerances that checking for crown height variations is seldom (if ever) done. However in this case we found the entirety of the squish discrepancy in the crown heights of the pistons ... a full .006^'.

The next thing we did was collect together every LR piston that we had in house to find out which of these two was the out of spec "Frankenstein" piston. Port opening on the LR engine takes place when the top "L" ring exposes the port windows (unlike other pwc engines whose ports are exposed by the piston crown itself). Given this, we considered the most critical dimension to be the "wrist pin hole to top ring land" dimension (where the "L" ring rests) . The next most important dimension was the "wrist pin hole to piston crown" dimension. After "sweeping" an indicator over a dozen pistons on a surface plate, we found that they were all within .001" of each other on "pin to skirt base", "pin to top ring land", and "skirt base to top ring land". The accuracy of the skirt length dimension was important because it allowed us to set up many pistons next to one another on the surface plate, and "sweep" many pistons at once. After establishing the consistent accuracy of these dimensions, we then swept the piston crowns at their center peak, and at their outer top edge. In this sweep of 12 pistons, we found a stunning .016" variation. This variation is greater than any dimension variation we had ever seen on mass produced pwc pistons.

At this same time we were in touch with a large Sea Doo dealer who was having similar squish measurement difficulties in the assembly of his race engine. With the revelation of our indicator sweep, he pulled all his new pistons off the shelf to sweep them. Like us, he found perfect accuracy on the "lower" dimensions of his pistons. However he found a .035" variation in crown height from the tallest to the shortest. To say the very least, the measurement variations he found shocked the hell out of us, and has caused us to review our entire "port timing/compression" format. To help identify the relative variations, we had to establish a fixed number as our "skirt base to crown peak" ZERO DIMENSION. That dimension is 4.300" (109.22mm). All our further documentation will refer to plus or minus heights from this zero point.

Before we explain any further on this subject, it's important to understand that this variation represents no mechanical "danger" to any LR engine that has had the squish clearance inspected and set within factory spec range. This variation can, however, cause some very significant variations in performance from one boat to the next (we have often seen these kind of "unexplainable" variations on stockers). Each LR motor is very likely squish tested and assembled at the factory with pistons made from the same "batch". This would likely explain why all the stock engines we've inspected, so far, have had two of the identical crown height pistons in it. The only reliability risk would be caused if a +35 piston were installed to replace a failed piston in an engine that came from the factory with a pair of +5's. The actual port timing of both cylinders would still be identical. However the squish clearance of the cylinder with the new +35 piston would be .030" less, and the head volume in that one cylinder would be reduced by 3.9cc (that's alot). At this time we are measuring as many pistons as possible in an effort to establish what the full range of specification is, and finding out what heights make up what percentages of available inventory. To date it appears that most pistons (about 60%) are in the +4 ~ +12 range As we measure more parts, this percentage may change. We will be glad to post percentage numbers offered to us by parties who have measured other large batches of pistons.

For the average LR owner, who checks the squish of both cylinders after piston replacement, there's not much to be concerned about. However, for the folks who plan on building a high performance LR motor...you better get yourself a surface plate and an indicator. The folks who build high performance LR engines realize that port timing and squish clearance are two separate measurements with their own individual importance. However if you perfect a porting format with a set of +5 pistons, those "perfect" port heights will need to be raised .030" in order to have the correct squish clearance with a new set of +35 pistons. Changing those port openings by .30" is sure to have horsepower consequences. To give these numbers some perspective, Group K holds port height tolerance within a .004" range when we port a set of LR cylinders.

WHAT CAN THE AVERAGE GUY DO? - Unless you have an uncommonly good relationship with your Sea Doo parts man, DO NOT ask him to measure all his pistons so you can have a matched pair. When you tell him that the crown heights may vary by .035", he will look at you like your nuts...and/or tell you to get out of his store. We suspect that, in time, dealers in the know will measure and match pistons for the engines they rebuild (if Sea Doo doesn't eliminate the variation first). As for everyone else, we would recommend a pre assembly "no base gasket" squish measurement of both pistons after one (or both) pistons have been replaced. If it happens that you find a height variation between the two squish measurements (.008" or less), we would recommend to "always" use the shorter piston (with the thicker squish measurement) in the rear cylinder. This will offer the rear cylinder slightly less compression, and a greater resistance to scoring as a result of crankshaft torsioning (see our Menu B document "Rear Piston Seizure").

At this time, there is no particularly "good" or "bad" crown height to have. However we will say that "for any given port layout" there eventually WILL be a "good" height to have. We expect that the Limited class racers will soon be all over this one. Since we do not build limiteds, we will not be testing this. Please don't call to ask us about it. As for our Sleeper/Hammer ported cylinders, we intend to engage testing to find out how the various crown heights affect performance. For now we can say, with certainty, that our porting delivers excellent overall performance with +2 ~ +12 crown heights. Given that we are in the peak of our season at this writing, we don't expect to complete this affore mentioned testing until autumn 97.

Impeller Selection and Water Velocity - During the spring of '96 we were fortunate enough to have a brief conversation with Kenny Stuart (owner of Skat Trak) about his thoughts on the XP pumps and impellers. At that time he was headlong into testing components for use on the 1996 Westcoast race team boats (we were finishing the tests of our XP Sleeper kit). He explained that he was experiencing some test results that were creating more questions than answers (a common occurrence during testing).

He explained that most folks (myself included) looked at impellers and nozzle diameters in the same way that we looked at gearing on a motorcycle. That is, high revving formats radared best with mild pitch props (lower gearing), while lower revving "torquey" formats radared best with steeper props (taller gearing). I agreed, telling him that any time we made a quantum increase in overall horsepower and torque on a pwc engine, we usually got the best radar speeds with the tallest prop it would pull.

Ken went on to explain that his testing was leading him to believe that those rules no longer applied...or at least did not apply to the pumps in the XP. He said that the engines in their race boats had enough power to pull very steep prop pitches up to peak rpm. However those steep props were yielding lower radar numbers than other milder pitched props they had tested. He seemed to be struggling to put the whole puzzle together in a way that made sense. He said that he believed that prop pitches and nozzle diameters had stopped becoming "gearing" type variables. He said that he was beginning to view props and nozzles as tools that could control the "water velocity" of the pump. He said that when the horsepower numbers are big enough, you can spin nearly any pitch prop. He correctly pointed out that horse power does not make boats go faster..."thrust" (or water velocity) makes boats go faster. He said that it appeared that the best speeds (in his tests) were being made by a certain "range" of pitches and nozzle diameters that were able to generate the "maximum water velocity". There is no doubt in our minds that Ken eventually put his puzzle together, and used the information to contribute to the Westcoast team's 1996 successes.

All this information may seem "old hat" and remedial to many hydro dynamic engineers. However, the horsepower to weight ratios (as well as the horsepower to prop diameter ratios) of pwc's have increased so much, so quickly, that this concept has not yet taken hold for many pwc engine builders. Furthermore, while the hydrodynamic specialists are quick to say "we knew that all along"...none of them have offered any definitive information that consumers can use to make an educated purchase.

For the last 6 months, we have been involved in "hair splitting" testing of the 782cc GSX platform for endurance racing. We didn't get very far down the road before our test results had Ken Stuart's words ringing in our ears. Unlike Ken's tests, we're not trying to maximize one format. We were doing simultaneous testing with several different formats. We already had a high revving format that had good water speeds. However we were hoping to find a, lower revving, more fuel efficient format that could run competitive peak speeds. As testing went on, the lower revving format had to be revved slightly higher to make good speeds. At this same time, the high revving format was yielding better speeds with a taller prop and larger nozzle. By the time all the possible combinations had been exhausted, we had put together our own puzzle. While we can't say that this information will apply to all high output pwcs (or even all high output Sea Doos) it certainly does apply to the 785 GSX.

Most closed course racers use the Skat Trak "swirl" props because of their excellent hook up in the "foamy" water of closed course events. In endurance racing there is little foamy water, and plenty of long semi smooth straights where sheer speed is needed. As a result, all our testing was done with the Solas impellers that waterspeed better in the smooth.

Maximum "On Water" RPM ---- Optimum Prop ---------------- Exit Nozzle Range

6900 - 7000 --------------------------- stock----------------------------- 87 - 85

7050 - 7150 --------------------------- Solas X1------------------------- 87 - 86

7250 - 7300 --------------------------- Solas Xo------------------------- 87 - 86

7450-7500 --------------------------- Solas Xo------------------------- 86 - 84

During these tests we never got better speeds with any (GSX) format that spun over 7500 rpm. The XP hull machines could benefit from peaks that were about 50 rpm higher, and/or nozzle diameters that are about 1mm smaller. On the other end of the spectrum, we never found a setup that yielded it's best speeds with the steeper X2 pitch. We suspect that exceptionally torquey formats that peak at 6900 could use the X2, however those rpms couldn't generate the speeds we needed.

The trait that really pointed to these combinations was not just peak water speed. Anytime we strayed away from the combinations on this chart, we seemed to take a noticeable loss in peak water speed AND high speed hook up ("pump loading" as the test riders called it).

As a side note, our initial reason for running these tests was to determine what made the fastest 92 octane pump gas format, and what made the most fuel efficient format. The fastest pump gas format is the 6900 - 7000 format listed on top. The most fuel efficient was, surprisingly, the 7450 - 7500 format. This format won on fuel consumption only because it matched or exceeded the speeds of the other combinations while running at 60% throttle. At 100% throttle...it guzzled.

We realize that this information lacks a "scientific" explanation. We will openly admit that we don't have one. However it has been our experience that our customers are more interested in definitive recommendations...and less interested in the science. We certainly invite the hydrodynamic folks to complete the theoretical picture (of what this chart says) in an understandable way.

About High Humidity - During our last year of on water testing with the Rave engine machines, we have observed some unusual, and hard to explain, variations in "on water" performance. Several times during the year, we had our daily testing interrupted by uncommonly bad weather of one kind or another (not so unusual). However when we resumed testing, a day or two later, we found that we had experienced some very significant rpm losses on machines that were unchanged from the last test date. More importantly, we did not experience similar losses on reed valved machines that were being tested at the same time. For a while, we though that there was simply a problem with the boat. But it always seemed to clear up a few days later...all by itself.

Later on, in passing conversation with some pipe builders, they described sudden and mysterious rpm losses of their unchanged test boats from day to day as well. After spending some time trying to find a common thread to all these disconnected (yet similar) losses, we could only find one common factor...sudden and temporary increases in humidity. That is, the large and sudden humidity increase that comes with a heavy rainstorm in an otherwise "not so humid" area. We understand that low barometric pressure also comes with such rainstorms. However our air density gauge (which doesn't "see" humidity) showed no huge losses in air density that mirrored the power losses we were experiencing.

Since making this connection, we have yet to have another series of storms that would allow us to confirm our beliefs. This leaves us in a position to express some important, but yet to be proved, thoughts.

Our testing seems to have shown us that the performance of high output Rave motors seem to be more affected by sudden increases in humidity than are comparable reed valve motors. It's hard to know if this is somehow a function of the rotary valve inlet system design, or just a function of the Rave motors being hurt worse because they make more power per cc than our reed valve test motors. In either case, we would not encourage Rave owners to race any of their reed valved buddies for pink slips on an uncommonly humid day.

This also could mean that owners of modified rave boats, who live in constant high humidity areas, may not realize the full horsepower increases being claimed by their performance shops. This is not a big deal while racing at home, because all your competitors have the same humid air. But it does mean that these owners could expect their rave engine machines to run considerably faster when they go racing in a less humid environment.

NOTE: As a rule we try not to talk much about any given problem until we have a sound solution for it. We certainly intend to continue working on this problem...however we have to concede that nobody (besides God) can do much about the humidity where you live. We hope to figure out some ways to make the Rave motors less affected by high humidity...but, for right now, that information does not look to be close at hand.

About Mid Range Evaluation - Conveying the exact strength of mid range power, in definitive terms, has always been a problem for both owners and engine builders. While we do not claim to have the "ideal" measurement procedure, we have one that works pretty darn good. We call it a "Full Speed Left" test. As it implies, the machine is ridden to full speed on glass water...then suddenly turned to full left. Since the front face of the impeller blades, and the full length of hull, are being pushed at maximum load against the glass water, this test induces the maximum possible (repeatable) load against the engine. While the turn is in progress, the rider observes (on a digital tach) the drop in rpm . In a perfect world, there will be a nominal difference between straight line, and full left rpms. In the real world, there is usually a significant drop in rpm. While engine builders disagree on how much of a drop is acceptable, we are firm on our limits. In the testing of all our recreational Group K Rotax engine kits, we consider a drop of more than 4% of total peak "on water" rpms to be unacceptable. For the higher revving race motors, we accept 5% losses. Our Super Stock I engine kit (XP) turns 7270 rpm, with an "FSL" rpm of 7150...that's a very acceptable 1.7 % drop in rpm. Our Super Stock II XP peaks at around 7540 rpm, with an "FSL" of 7210 rpm (a difference of 4.5%).

It looks simple on paper, however it's not nearly so easy in practice. Most riders want to install the prop that gains them the best peak water speed...that's easy to find. However that prop will often yield an FSL rpm loss much greater than 5%. The natural choice, for riders who demand maximum cornering acceleration, would be to slightly increase nozzle diameter, or install a slightly milder pitch prop. While these mods will usually increase both the FSL rpm and the peak rpm...the gap between the two will be reduced. Unfortunately the radar speed will also be reduced, while the octane requirements (for the higher rpm peak) are increased. The rider must then choose between running a slower speed (not acceptable for most) or bolster up the mid range power. Here again, nearly everything that increases mid range power will tend to reduce peak water speeds. After much testing with our kits, we have often found that there are two, or more, different impeller pitch/nozzle diameter combinations that can yield the same peak waterspeeds. However they will yield very different FSL rpm losses. We try to cover as much of this testing as possible...so our customers don't have to. Just the same, we often get calls from customers who want some engine modifications, but want to retain the aftermarket prop that they have already purchased. They want to know, "Will the prop I have work?" The answer..."Well, yes it can work...it just can't work as good as the right combination." So often we see customers who are looking for more midrange power, when they often need only to access the midrange that is being hidden.

On the flip side of that equation are customers who make a modification that seriously hurts mid range power, and increases the FSL rpm gap. In a perfect world if you install any part, for example a well designed pipe, you should expect to see an increase in peak rpm as well as a decrease in the FSL rpm drop. On laydown rave engines this seldom happens. Some pipes, intended for more rpm, may require the use of a milder pitch prop to accomplish a performance increase in both areas. That's okay, as long as the eventual FSL rpm percentage loss is still acceptable...and the new peak rpm doesn't require a quality of fuel octane that you can't afford.

About Break In - The piston rings on nearly all Japanese made piston rings are manufactured with a thin layer of Teflon on the bore contact surface. This Teflon coating permits the rings to seal very quickly to the bores, hence reducing the break in period. Rotax rings have no such coating. The hard chrome bore surface of the top rings, and the bare iron surface of the lower rings, require alot more break in time before they completely seal to the bores. This is why Rotax owners often report that their new machines get noticeably faster as they get run longer. Owners that install new rings in older machines will experience the same effect. The point of this is, if you assemble a modified rave motor with fresh rings, do not expect to see normal performance levels until 4 - 5 operating hours have passed.

About ECWI - IT WORKS!! This device, offered by Factory Pipe Products, makes a huge difference in acceleration from 3500 - 6800 rpm. It works as well on a stock pipe as it does on the modified ones. At $250 (for the solenoid and the electric box that talks to it), this is one of the greatest "all gain-no lose" modifications we have seen in a long time. We expect that these devices will soon become common fare on most modified (and some stock) machines.

Pump gas pipes - There is a bevy of aftermarket pipe builders preparing to release various exhaust systems for the Rotax 782cc engine boats. Since this is perhaps the biggest ever "single model" opportunity, the stakes (and the tension) among these pipe builders is very high. All of these pipe builders have done the lion share of their discovery testing on stock engines in a dyno room. We believe this is sound thinking, because these pipes must work on the most basic of engine formats. However we see a storm brewing for some of these pipe builders.

The forth coming batch of aftermarket pipes will likely increase the power of a "stock" boat into a "92 octane safe" rpm range...that's great. However, whether these pipe builders like it or not, most laydown rave owners (who are seeking more power) have already begun their modification project with the stock pipe in place. These slightly modified machines are already turning the rpms that the new batch of aftermarket pipes can lift a stocker to. When these pipes (which all come with high rpm ignition modules) are installed on otherwise modified machines, the result (with most) will be a slight increase in peak rpm along with a slight loss in bottom end. Most of the experienced Sea Doo technicians out there will immediately think "This thing is over propped!" After switching to a lower pitch prop, the machine will immediately become a violently fast, fire breathing beast...that turns well beyond 7100 rpm. (race gas territory)

This scenario is unique because there has never before (in the pwc industry) been a pipe whose mere installation resulted in the mandatory need for expensive race fuel. Of course, the builders of these pipes could simply mandate the use of the taller "less acceleration, less rpm" prop. But no rave owner in his right mind will want to use that taller prop once he has ridden the (rocket ship) milder pitch setup.

None of this information means that the aftermarket pipes are "no good". However it does mean that many of these pipes will be no good for owners who "must" run 92 octane fuel. At Group K, we have no intelligent solution for this potential problem. However we are anxiously watching and waiting to see how the pipe builders intend to solve it.

We have seen no after market bolt on part that can match the 92 octane power output of our Sleeper kit (7050 rpm / $650) Our testing has shown us that any aftermarket pipe ,on top of this Sleeper setup, turns 7300+ rpm (with a slightly milder prop)...that, again, is race gas territory.

We expect that the pipe builders will all come up with their own solutions. We urge 92 octane customers to ask these pipe builders specific questions about rpms, and octane (not to mention recommendations for prop pitches and maximum compression)...before you buy.

Nozzle diameters - While R&D makes exit/ steering nozzles in the 83/85, 85/87, and 87/89 sizes... we have found the need for 1mm increments on our race formats. Buying all these nozzles is a bit expensive, so we have done our best to make the closest possible recommendations. However for the racer who plans to ride in many different water and race course conditions, 1mm increment nozzles are strongly recommended. To that end our testing has uncovered one very important variable.

The steering nozzle diameter "must" be 2-3 mm larger in diameter than the exit nozzle. If the steering nozzle is under 2mm larger, peak rpm and acceleration ability will suffer significantly. If the steering nozzle is larger than 3mm over the size of the exit nozzle, the precision of steering is affected. DO NOT assume that the aftermarket nozzles you have purchased are accurately sized. Measure them, and bore them as needed to maintain the 2-3 mm difference...it is time well spent.

Impeller pitches - All the impeller manufacturers have adopted their own pitch measurement systems. Unfortunately, no two of these measuring systems seem to be the same. This situation means that the pitch numbers offered by each maker are useless information to potential prop buyers (a situation that the prop builders ought to remedy). Just the same, engine builders must come up with some sort of system by which "they" can compare different pitches from different manufactures. At Group K, we turn a blind eye to the pitch numbers, and instead base our comparisons on how each prop "loads down" a particular engine format at low rpms and high rpms. It may not sound very precise (it's not), but it's more meaningful than anything being offered by the prop builders. Our system uses the stock prop as a "0" base, and goes plus for higher pitches..minus for lower pitches. Note : as more props are available and tested, we will include them on our chart.

Solas X2 / +3 Solas X1 / +.2

Stock XP /0 Stock GSX / -.1

Nu Jet / -.5 Skat Trak swirl 17/22 / - 1.5 Solas Xo / -2 Solas X / -4

About the "High Compression" combinations - At the '96 Havasu World Finals we shared a display booth with Factory Pipe Products and Rocket ship Racing. The purpose was to jointly launch the availability of the Fischetti Replica race boats, the FPP Spec 1 and Spec 2 pipes for the 785 rave, and the Group K 785 rave Super Stock engine packages. During our week in the booth, we had an opportunity to speak with many 785 rave owners about their modified 785's. By the end of the week, we realized that we had dramatically under estimated the number of rave owners who already have cylinder heads that yield 200+ psi compression (which our testing has led us to believe is very excessive). Despite the continuing sales of 200+ psi heads, we are still convinced that the "softer compression (170 psi)", format will eventually become the prevailing high performance technical trend.

Many these owners spoke about how many starter failures they have, as well as complains of random rear piston seizures. Our test leads us to believe that these are also related to the excessive compression. We spoke with one racer who claimed to have an ignition problem that was compression induced. He said that every time he installed his 240 psi head, the engine would begin to sputter and pop at 7100 rpm. We explained to him that increased compression requires more voltage to cross the plug gap. He had simply reached the voltage threshold of his boat's ignition. Operating the engine at length with this setup would eventually cause various ignition components to fail prematurely.

We believe that one of the original purposes of the rave valve is to permit a wide powerband without having to resort to high compression ratios. 200+ psi compression ratios seem (to us) to defeat this primary design intention. And we will reiterate, "Every time we exceeded 170 psi in our tests...our boats ran slower...NEVER faster!"

A supporting footnote for this approach is the 785 Polaris race boats. The 1996 race team engines have run their best with about 125 psi indicated. The new SLX Pro, with it's claimed 138 hp, does not require more than 135 psi. We had not noticed very many 200+ compression Sea Doos consistently "blowing off" the 785 Polaris team boats this year.

About oil mixing and injection - While many engine builders don't care to use the stock Rotax oil injection system...we do. This system does a great job of providing "rpm adjusted" oil input, as well as circulate the lubrication supply for the rotary valve drive. We recommend that it be retained for all recreational and off shore competition applications that turn under 7300 rpm.

On our off-shore Sleeper and Super Stock I kits, we use a combination of the oil injection and a 60:1 oil premix. This yields a combined ratio (at full throttle) of about 25:1. The offshore engines using this mix do not smoke visibly more than any other machines. However the additional oil helps greatly to improve piston and ring sealing. This is an important function for race engine formats (like ours) that use compression ratios that are much lower than the aftermarket norm. The presence of this additional oil results a measurable increase of indicated compression readings without actually increasing the mechanical compression ratio. This mix also offers the additional lubrication that enables us to fit slightly tighter piston clearances. These tighter clearances also increase indicated compression with out increasing the mechanical ratio. There are many builders that claim a richer oil mix is a handicap. They claim reduced fuel octane, reduced combustion strength and a host of other maladies. WE DISAGREE! In our 25 years of racing two strokes, we have never seen leaner oil mixes out perform richer mixes on race engines.

An added benefit of this dual oil mix is protection from an accidental loss of oil supply to the pump. We are pleased (and ashamed) to say that we experienced a few situations this year when the additional premix saved an engine that would have otherwise fried itself. It bears noting that the dual mix would be too much oil for recreational machines that are trolled often. However any rave engine that gets constant, heavy duty, use will benefit from this dual mixing.

Helpful Hints and Other Stuff

Rave Valve Diaphragms - Recently a customer had brought us his '96 XP 785 Sleeper complaining of a big loss of peak rpm ability. All the engine internals seemed to be in good condition, and the engine did not respond to a wide range carburetor adjustment.

After checking all the normal stuff, we decided to see if the rave valves were getting full movement. One quick way to confirm this is to run the machine (on the water) after completely removing the black plastic rave adjuster caps and springs (on the cylinders). While the bottom end performance is not very good, the valves are certain to attain the pull up position. With these caps off, our customer's boat immediately picked up the 200 rpm it had been short of. Reinstalling the caps resulted in the 200 rpm loss returning. After plenty of inspection and parts changing, we isolated the problem to the rubber diaphragms on the rave valves. These diaphragms were well secured on the top and bottom, and visually "looked" good. However when we replaced them with new diaphragms, the engine ran (and revved) normally with the black adjuster caps and springs in place.

Apparently the rubber rave diaphragms can become weakened in a way that allows them to look okay...but not work okay. This particular machine had a lot of hours on it, so we figured it is just a matter of time before it happens on most rave engines. If you have a 785 rave boat with lots of hours on it...it would pay to change these diaphragms before you set out for the season.

Rossier Pipe Tests PART 2 - While we were pleased with the 92 octane tests of the Rossier pipe on the '96 XP hull (with the Sleeper kit), we realized that the 785 models needing the most help (performance wise) are the heavier and longer 785 Sea Doos. In particular, the badly under powered 1997 spring seat 785 XP. We installed the Rossier pipe on a '97 XP 785 (with the Sleeper kit and stock prop) to see if we got the same performance numbers offered by the lighter '96 XP. It bears noting that the rev limit eliminator module for the '97 ignition is a little over twice the price as the similar module for the '96 electrics. ($228 vs $99).

On the water, the peak rpms of the '97 XP Sleeper were the same as the '96...7300 rpm. Unfortunately the loss of peak rpm, that came with extended full rpm operation, was also present. One big difference, however, was that this '97 XP setup had noticeably less authority of acceleration off the corners (compared to the '96). This weaker acceleration was on the borderline of being acceptable for recreational riding, and was definitely not acceptable for competition of any kind. We considered this a problem that had to be resolved.

In an unrelated conversation with Gary Robison (of Novi), we mentioned this acceleration problem. Gary noted that he had solved such problems with slight increases in carb throat diameter. We explained that we had never see any such increases in the same tests we had done with our Sleeper kits (using the stock pipe of course). Gary agreed with us with respect to the stock pipe tests. However he stated that the installation of higher rpm biased exhaust pipes can change the inlet tract resonation in such a way that larger throats can offer a genuine benefit. He also denoted that the stock inlet manifold opening is usually 41.5 - 41.8mm in diameter. To this end, Novi does a modification (for about $240) where they bore the stock carbs to 42mm, and fit them with larger butterflies. We immediately got a pair of these 42's, and installed them on our '97 XP Sleeper/Rossier pipe test machine. The increase in acceleration was phenomenal. The larger heavier spring seat 785 bolted off the corners every bit as good as the lighter '96 boat had. There was no increase in the 7300 rpm peak, however it seemed that the boat didn't loose near as much peak rpm ability during the extended full throttle runs. All in all, this package made the '97 XP a very powerful machine.

After running these tests, we began to wonder if the 42 Novis might offer a similar kind of gain (with the RE pipe) when installed on our lighter '96 test boat. Within a few hours, we had the RE pipe and the 42s mounted on our '96XP ready to hit the water. Again, the peak rpm remained around 7300...but this machine was fast...very, very fast. The acceleration was easily acceptable for competitive level closed course racing...on 92 octane. The acceleration of this setup was so violent that the boat was a little hard to control. We installed a set of (smaller) R&D 85/88mm nozzles with hopes that this would tame the boat down a little. We expected a reduction of peak rpm...but there was none. Instead the XP ran right up to 7300 again...but picked up 1.5 mph doing so. The acceleration seemed to be a little more controllable...but only a little.

The end result of this test certainly delivered more than we expected. This '96 XP setup (Sleeper/RE pipe/Novi 42s/ R&D nozzles/stock prop) is the quickest and fastest pump gas pwc that we have ever built. Depending on weather conditions, it radars between 64.3 and 65.2 mph. Every one of our test riders considers it the ultimate recreational pocket rocket...and the fastest pump gas boat that they have ridden.

Rossier Pipe Tests - After 2 years of selling Sea Doo 785 Sleeper kits, we have been overwhelmed by kit owners who are looking for "a little more" speed while staying 92 octane pump gas safe. The deteriorating "high performance friendliness" of pump gasolines has made this request ever harder to accommodate. The area most kit owners are interested in approaching for this increase is that of aftermarket exhausts. We have tested repeatedly with the FPP spec 1 and 2 pipes (that we use in our race gas kits) in an effort to come up with a "faster than a Sleeper kit" pump gas setup. To date we have been unsuccessful in this area.

After several conversations, Charlie Rossier (of Rossier Engineering), convinced us to try one of his pipes along with our Sleeper kit 785. Charlie swore that his pipe would allow for the extra speed we were looking for, without inflicting the rear piston scoring that had accompanied so many other such tests with other pipes. We had tested the Rossier pipe on our race gas engine packages, but ended up choosing the FPP spec 2 pipe for it's higher rpm abilities on those 110 octane setups (about 150 - 200 more).

For these pump gas setups, Rossier Engineering normally subscribes to the Skat Trak "swirl" props (for their excellent rough water hook up). Unfortunately these, great accelerating, props normally make for a loss of 2 - 3 mph of peak speed on smooth water. Our customers were clamoring for more smooth water speed, so using the Skat swirl was not an option. Just the same, we installed the Rossier pipe on our in-house '96 XP Sleeper boat. This machine runs high 62's to low 63's (mph) with a stock prop and 85/88 nozzle @ 7090 rpm. To be a success, the Rossier pipe would have to generate speeds above this set up.

We started our testing with a stock prop and stock nozzle set (later testing would show this to be the best setup for peak speeds running the RE pipe on this boat). After all the jetting work was completed (setting the high speed richer until the rpms dropped), we considered our XP Sleeper safe to run on long, wide open, glass water test runs. The cold rpm numbers (stock prop & nozzles) were always around 7300...very fast (about 64.5 mph). However the rpms began to drop noticeably after 2-3 minutes of w/o glass water operation. After 6 - 8 full throttle miles (and minutes) the rpms would drop to a sustainable low of 7100 - 7120. This "hot number" rpm yielded a peak speed well under the smaller nozzled "stock pipe" Sleeper. Despite the huge loss in hot tach numbers, we were still very encouraged by the potential of the pipe's cold tach numbers. We were also very encouraged because this was the first setup we had run, at these rpms on pump gas, without scoring a rear piston. However for the Rossier pipe-ed setup to be truly better (speed-wise), we needed to sustain the hot numbers at least in the low 7200's.

In an effort to reduce the loss of hot rpm numbers, we tested with lower compression, milder ignition timing, various props, nozzles, etc. However, anything that improved the hot tach numbers, also reduced overall performance. Any pump mods that pulled the rpms down, hurt the low end acceleration to an "unacceptable" level.

After exhausting the mechanical possibilities, we attempted to resolve the temperature problem with some octane improvement options. Using ordinary over the counter octane boosters had almost no effect at all. A 50/50 mix of 100 octane avgas/92 octane pump gas allowed the engine to stabilize at 7160 - 7180. A 100% batch of 108 octane race fuel allowed for sustained 7300 rpm operation (without any rpm loss at all). In short, nothing affordable worked very well.

At this point it bears noting that all our ( 92 octane) tests, of extended full throttle on glass water, were intended to portray the heaviest possible use. Very few customers actually operate their machines in this way. With this in mind, we conducted tests to see how much rpm could be recovered if the machine was allowed to cool down via low speed operation (after the 10 minute w/o pass). We again ran the motor from the cold 7300 rpm down to the fully heat-soaked stabilized 7100 rpm. Without shutting the motor down, we then ran the boat at low to medium throttle for about 5 minutes in an effort to reduce the engine temperature. After this cooling period, we once again pushed the engine up to full rpm. This means of engine cooling allowed the rpms to run in the 7230 - 7250 range for a minute or two before the rpms once again descended into the low 7100s. No matter how many times we repeated this process, the cool off period would always allow for the short spurt of returned middle 7200's.

Summary: Unlike other exhaust systems we have tested to date with our 785 Sleeper kits, the Rossier pipe appears to offer safe (non piston scoring) operation into the 7200 - 7300 rpm range on 92 octane fuel. As long as the rpms are above 7200, this pipe allows the Sleeper to run slightly higher radar speeds than the best stock pipe Sleeper setup. This 7200 rpm range can remain within reach (with 92 octane) so long as the engine is not run "constantly" at full load. Owners who seek to run their engines at 7300 (indefinitely) can do so by simply using 105+ octane race fuel.

The loss of rpms that comes with engine heating is not a function of any design problem with the pipe. It is a function of the sheer heat being generated by the higher horsepower output. The only way to eliminate this problem is by cool down running or race gas use.

NOTE: All these tests were conducted on a '96 XP boat. We intend to repeat these tests on the larger, heavier, GSX and '97 XP hulls as time permits. We will post those results as an addendum to this entry.

Coffman Exhaust Pipe - We had an opportunity to test a production Coffman pipe on one of our GSX endurance racer boats that was using our modified stock exhaust pipe (our modified stock pipe allows for 200 rpm over the stock unit). With this modified stock pipe, our test unit turned 7280 rpm with an Xo prop and a stock 87mm exit nozzle. After installing the Coffman pipe, the same boat accelerated much harder in the bottom and mid range, not to mention having the power to pull a 2mm smaller exit nozzle to the same 7280 rpm (larger nozzles turned more rpm, but did not make more speed). This would equate to about 1.5 mph gain over our modified stock pipe(that's pretty good). The Coffman pipe was easy to install and easy to tune in. The headpipe fitting uses a Mikuni main jet for the adjustment of water being injected into the pipe. We got our best results with a #115 jet. Given that the water injecting point is so small in diameter, a water filter would be mandatory to avoid accidental blockage (no filter came with the pipe). Talking about jetting, this pipe also caused our GSX engine to want bigger pilot jets (85's from 77.5's) than it has ever needed before (even with the 7500 rpm FPP Spec 1)

The stamped aluminum body was extremely lightweight, and easy to get in and out. Our endurance racers had some reservations about the long term abuse that this pipe body might be able to endure...but only time will tell.

What everyone wants to know is, "how does it compare to the FPP pipe". Well this same machine, fitted with the FPP spec 1 pipe, turned 7500 rpm with the same prop and nozzle combo. From this respect, the FPP pipe yielded higher speeds...but it has to turn 7500 to do it. Installing a taller prop to pull the FPP setup down to 7300 hurt waterspeeds and rough water hookup (see Impeller Selection and Water Velocity above). If your a closed course rider who wants the added reliability of the more conservative rpm peak, the Coffman can be a good choice. However, we don't see the Coffman design being tunable to get the strong 7500 rpm that the FPP makes.

It should be noted that neither of these rpm ranges can be considered anything but "race gas only" territory. If you want to run 92 octane pump gas...forget about all of them (our modified stocker included).

Pump Exit Nozzle Diameters - During the prop and nozzle testing mentioned above, we noticed that our best rpm numbers and radar speeds were made with exit and steering nozzles that had a 3 mm diameter difference. When we finished the testing, we called R&D to let them know. They had apparently learned the same thing, and had already started production of 85mm/88mm and 82mm/85mm nozzle sets. For owners of earlier nozzle sets, we recommend that you measure the exact diameter of your exit nozzle, then bore your steering nozzle exactly 3mm larger...it works.

R&D Trim Tabs - The R&D trim tabs (aka "side wings") are a great benefit to XP owners who do alot of riding in exceptionally rough water. These tabs help keep the nose of the boat down, not to mention a big reduction in proposing. The down side of these tabs is a 2 - 3 mph loss of peak water speed on smooth water. Since many XP owners are paying lots of money to "gain" 3 mph from their engines, a loss of this magnitude is not acceptable. We recently prepared a Super Stock 2 kit on an XP for Karin Pautrel to run in an endurance race. We wanted to use the R&D tabs, but we didn't want to lose the speed.

We laid our stock trim tabs against the bottom of the R&D replacement, and scribed the profile of the stock tab onto it. We then machined the bottom of the overhanging area of the R&D tab at an upward 3 degree angle. The results were impressive. The hull maintained excellent rough water stability, along with less than a 1 mph loss on smooth water. We are made to understand that R&D now inventories these trim tabs with this same 3 degree cut. We strongly recommend them to anyone considering a trim tab purchase. As for the older tabs, we won't have the shop time to start machining them for customers until July (yes we are that busy).

Precision Top End Assembly of the Laydown Rave Motors - Squish clearance, the distance between the piston crown and cylinder head at top dead center, is an important specification that is minded by the Rotax factory as well as all high performance engine builders. The Sea Doo manual offers a fairly wide tolerance for squish clearances, however most engine builders prefer closer clearances to help stave off detonation. Most builders measure the squish clearance by sticking a piece of solder through the spark plug hole toward the outside edge of the bore...then momentarily touching the start button so the piston makes several strikes at the solder. This allows the solder to be crushed to the exact thickness that shows the "squish clearance". At Group K we make a point to take these "squish clearance " measurements over both ends of the wrist pin. This minimizes the piston's ability to "rock" and show an inaccurately large clearance. We also take these measurements over both ends of the wrist pin, on both cylinders, to assure uniform clearances. In a perfect world, all four squish measurement will be within .002" - ,003". Unfortunately, we have been observing everything but uniform clearances. This is a matter we do not take lightly, since accurate squish clearance is the only way to assure correct deck height setup (a function of the base gasket thickness).

Immediately, we though that the variations were caused by incorrectly cut squish bands on the cylinder head...not so. The bands had a uniform depth and perfect center to center location. We then suspected variations in the locations of the cylinder mounting holes in the crankcases, but they also measured perfectly. In the end, we found two items that caused the variations. The first was that the distance between the two cylinder bores can vary greatly, depending on the assembly procedure. That is, there can be almost 1 mm variation in the center to center distance of the bore diameters based solely on the random location during assembly. Secondly, the cylinder head itself has a great deal of movement leeway (both left to right, and forward to back). All these variations of fit take place because the cylinders and cylinder head are not located to one another with "dowel locating pins" (as most Kawasaki engines are). The absence of these locating pins allows for alot of movement leeway, of both the cylinders and the cylinder head. That results in broad variations in squish clearance measurements. Installing dowel pins is a questionably wise solution because of the great risk that the exhaust manifold faces would not be perfectly parallel (that would cause bore distortion when the exhaust manifold is torqued on).

For now, our best solution is to recommend a "Precision Assembly Procedure" that will make for minimum variations in squish measurements.

Our procedure is based on the following information. The cylinder base bolt patterns in the cases are located 132mm (5.196") apart. The standard bore diameters of the cylinders are 82.0 - 82.08 mm (3.228" - 3.231"). This means that the inner distance between the two bores of the torqued on cylinders should be between 49.91 - 49.98 mm (1.965" - 1.968") These last measurements can be taken quickly and easily with a set of dial calipers after the cylinders have been torqued on. In most cases, we have found that you can hit this spec range if you install the cylinders as close together as possible. It bears noting that we have seen engines where the water jackets of the cylinders touched before we could reach this specification range. We literally needed to belt sand some material off the aluminum cylinder casting to get the correct spec. Cylinders like this are certainly the exception (not the rule), but they do exist.

Once the cylinders are installed with the acceptable center to center distance, it becomes apparent that the cylinder head itself can move about 2mm (.080" in any direction over the tops of the bores. This location will also have a profound effect on squish clearance measurements. We got out best results by matching the edge of the head casting, all the way around, as closely as possible to the edges of the cylinder castings. While this doesn't sound very precise, it works surprisingly well. After torqueing on the head, you can take your four squish clearance measurements to determine if the head needs to be moved slightly one way or the other...it's not really as tough as it sounds.

Does everyone one who assembles a Rave top end "have" to go through this whole procedure...not necessarily. However we strongly recommend this procedure to anyone preparing a high output Laydown Rave motor. If you find a wide variation in the squish clearances of your freshly assembled engine...it could certainly avert some problems before they get started.

About the Factory Pipe Products Pipes - All of our early development work for our 785 laydown rave racing engine packages was performed with the use of an early version of the FPP Spec I pipe (the one photographed on page 45 of the Feb '97 Splash). This pipe, used without the ECWI, made great overall power and acceptable "full speed left" tach numbers. However it did appear to need a relatively low impeller pitch to accomplish all this on our engine packages. These low impeller pitches resulted in rpms (and octane numbers) that was acceptable for our Super Stock II formats, but a little on the high side for our Super Stock I kit. The S/S I kit is intended to be a 100 octane format that could be quick enough to be a competitive closed course racer, yet still withstand the long term abuses of off shore racing. For this, we needed an rpm peak of no more than 7250. We found that by modifying the components of the stock pipe body, these rpms, along with impressive midrange acceleration, could easily be attained. In fact this arrangement had so much overall torque that we were able to use the higher pitched Solas X prop with an 85mm nozzle. This new pipe/prop setup revved 150 rpm less than the Spec I/Solas x setup, yet allowed for the same peak waterspeeds. This modified stock pipe cannot turn the 7500 rpm's needed for the S/S II, but we think it is an ideal choice for our S/S I kit. Since we do no testing of parts on otherwise stock machines, we have no information for the performance (or prop choices) of the FPP pipes on a stocker. However, it bears noting that Factory Pipe Products developed their pipes on a stock engine...not a Group K kit. It is no surprise that our test results differ somewhat. As for our current S/S II kits, we recommend exclusively the FPP Spec II pipe. On our test boats, the Spec II out performed the Spec I in every way. We suspect that the Spec II pipe worked much better in this application because it was developed on modified engines as well as stockers.

Loose Rave Valve Tops - The Rave valve guillotines are attached to the actuating diaphragms via a plastic threaded cap. The top spring rests between this threaded cap and the top cover. The Sea Doo manual does not give a torque spec for the tightness of this threaded cap, however we presume that the torque must be relatively light to avoid peeling the threads out of the plastic cap. If this threaded cap unscrewed, the guillotine would stop itself before running into the piston. However, during the process of unscrewing the guillotine would protrude noticeably into the exhaust port (at the full up position). That is, it would not be moved into the full up position flush with the exhaust port roof. Among other things, this could cause a noticeable loss of peak rpm ability. Knowing this, we gently tightened the threaded plastic caps on our test machines...and forgot about them.

A couple of weeks later we preparing to port a pair of rave cylinders that had just been removed from a customer's boat. After seeing that the guillotines did not retract all the way up, we noticed that the threaded caps were loose. This customer never noticed any unusual noises or poor performance, so it was impossible to know how long the caps had been loose. For curiosity's sake, we checked the threaded caps of two other cylinder sets being prepared for porting...they were loose too! Then we started checking all the test boats we had in the shop...the caps on every single one was loose (even the one we had assembled and tightened two weeks earlier). As a result of this experience, we recommend that these plastic threaded caps be checked regularly to assure that they are tight. As an added measure to "slow down" this inevitable loosening, we also recommend to put a zip tie at the bottom lip of the rubber rave valve diaphragms.

1997 Ignition Boxes (11/16/96) - Sea Doo calls their cdi brain box an "MPEM module". The 1995 XP 800 and the 1996 XP have MPEMs that are slightly different from one another. To accommodate this circuitry difference, the makers of aftermarket high rev modules for the laydown rave machines have been forced to make their rev modules specific to each of those two year models. For 1997, Bombardier has, once again, changed the circuitry in the MPEM units in the laydown rave machines. The folks at Micro Touch (who brought this situation to our attention) have already embarked on development of a rev module for the 1997 machines, however a problem looms for some 1996 owners as well. Apparently the last 1000 1996 GSX models were assembled with this new 1997 MPEM unit. That means that none of the existing aftermarket rev modules will work on these machines. Your Sea Doo dealer has a service bulletin (#96-34) that denotes all the affected GSX models.

Base Gasket II - After the above experience, we started to look very closely at the thicknesses of our base gaskets. The Sea Doo factory gaskets are supposed to have .1mm (.004") of thickness for each hole stamped into the center pattern. However we have found many gaskets in our inventory that were too thick or too thin by .15mm (.006"). This may not be a big deal on a stock engine...but "we" certainly think it's a big deal on our Sleeper and Hammer engine kits. In any case, we urge all rave owners to check their squish clearances, and measure the thickness of your base gaskets. What you see (in the hole pattern) may not be what you get.

Base Gasket Thickness - "Jon's Story" - One of our early test boats was an XP that was to be raced in region one offshore. We shipped the kit to the owner, who had it assembled at a local shop. After the break-in, the test rider reported a little better top end, along with average bottom end power. We were concerned because all our other test riders had reported big increases in bottom end as well as peak rpm. After a few unsuccessful attempts to resolve the problem with carb tuning, he came to our shop with the machine. After a quick test ride, we found this XP to have mediocre acceleration and a 6900 rpm peak (that's 150 rpm short of the norm). Back at the shop we checked out every possible problem. Besides the indicated compression being just a little low, the only other inconsistency was that the squish clearance seemed to be about .010" (.25mm) too thick. At the time, we didn't think this was the problem...but it was the only thing we saw that we could "fix". We removed the .024" (6 hole) base gasket and replaced it with a .016" (4 hole) base gasket. The next morning we gave the boat a short break-in...then gassed it. The difference was unbelievable. The boat pulled viciously up to 7040 rpm...just like all our other test boats. No one was more shocked than us, that .008" of base gasket thickness could turn a weakling into rocket. After this experience, we specified for our assembly instructions that a .038" - .042" squish clearance must be maintained on all Group K modified top ends. Since then, we have not experienced this problem again. (Note : This XP won the 90 minute 1200 pro overall at the Havasu Global Offshore Finals)

Fuel consumption - Since offshore races are often won by quick fuel stops, we sought to get the maximum mileage from our GSXs. The design of the GSX fuel filler leaves a large air space locked in the top of the fuel tank. We found that by loosening the top clamp on the fuel pickup, we were able to purge out all the air, and gain about 1.5 gallons of capacity. This procedure should be done before the race, on a work stand, out of the water. The chaos of a pit stop does not make it time effective for a race stop. However the end result is a first tank that can last about 110 minutes (on a 7200 rpm GSX "Hammer"). This permitted our GSX to run the 90 minute events with "no" pit stop, and require only a quick "splash" of fuel for the longer events. (NOTE : This GSX won the 90 minute 785 pro overall, and the 2 hour 1200 pro overall at the Havasu Global Offshore Finals)

Oil injection pump screws - Two screws in the top of the stock oil injection pump are responsible for holding on a tin plate that doubles as a "check valve cover / cable end holder". One of our low hours GSX machines had these screws come loose and fall out. The result was oil spewing out of the top of the open pump (this of course resulted in a fried motor). We immediately checked the screws on our other two GSXs, they were loose too. We secured them with Loctite, and have had no problems since.

95 Rave guillotine - The Rave valves on the 1995 XP 800 are different from the later style valves used on all 1996 rave engines. The main difference is the design of the shaft where it connects to the guillotine. The design of the earlier shafts makes them susceptible to breaking. When this shaft breaks, the rave guillotine drops down to crash into the piston (it's not a pretty sight). We would urge any 1995 rave owners, who still have the original rave guillotines, to replace them with the 1996 style. 1995 valves have the number "350" cast onto the face...1996 valves have "352".

Group K
4597 Calle Del Media
Fort Mohave, AZ. 86426
928-763-7600
email: groupklemm@aol.com