Tom Wilson
December 14, 2009

Horse Sense:
We were still two blocks away from Westech when we were surprised to hear the thunder of what sounded like a 7,000-rpm locomotive over our '91 Mustang's stereo. Turning down the tunes and opening a window, we discovered the distant roar was just our test engine warming up on the dyno.

So, you bolted a Kenne Bell Twin Screw supercharger to your GT500 and managed to detonate a connecting rod through the block by way of tuning exuberance. What's your next move? Get a bigger blower, of course.

Like any good hot-rodder, Mark Meiering ordered a robust aluminum short-block and bolted the newest, largest Kenne Bell supercharger atop it. This made Mark the first to fiddle with Kenne Bell's just-announced 3.6-liter water-cooled supercharger, and when the combination was tested on one of Westech's Superflow engine dynos we were invited to tag along for the first-ever proof-of-concept and tuning ride.

Considering the heretofore largest-of-the-large Kenne Bell blowers-the 2.8-liter-could slap down over 800 hp to the rear tires of an otherwise stock GT500, just what is the 3.6-liter supposed to do? Make more power, of course, but via increased efficiency rather than increased boost. By the way, there are 2.8 and 2.8H (High Pressure Ratio) Kenne Bell superchargers. In this article we casually refer to the "2.8" blower simply to denote its displacement; in all cases we're referring to the 2.8H version. All 3.6 blowers are considered high-pressure ratio, but because that's the standard 3.6, there's no need for the "H."

When Jim Bell, the big rotor at Kenne Bell, decided to build the 3.6-liter supercharger he was looking for several gains. First, a larger blower would be useful for larger-displacement engines. These are typically big-inch Chrysler and Chevy units, not the pip-squeak 281ci 4.6-liter mills, or even the middling 5.4-liter Four-Valves. They make good power with the 2.8-liter in all but the wildest applications. Of course, many 5.4 owners are a little over the edge, so the larger blower could come in handy for a larger percentage of Shelby owners than we might first imagine.

Second, Jim knew he could increase the supercharger efficiency, on all engines, because the larger blower would turn more slowly. Plus, he could make a few mechanical upgrades. This would mean that for the same boost the supercharger would take less power to drive and thus shuttle more power to the Mustang's wheels rather than waste it driving the supercharger. So, while many Mustangs don't really require a larger blower, they do benefit from a more efficient blower.

Also, high-boost screw blowers suffer from non-linear heat distribution. A screw blower is a compressor and unavoidably heats the air, and because the air enters one end of the supercharger and exits at the other, a screw blower runs cooler at its inlet end and hotter at its outlet end. On moderately boosted street engines this effect isn't enough to bother with, but as the boost goes up, so does the air temperature-it's just a fact of physics and not a sign of poor blower design.

Just how hot is hot? At crazy high boost 100-degree incoming air becomes 350-degree outlet air-that's a huge heat gain for a twin-screw blower. Jim says he horsebacks a 10-degree gain per pound of boost. Thus, when boost arcs into the 25-pound stratosphere, the hot blower rotors have expanded enough to crash into each other, destroying themselves and sending gritty metal bits downstream into the engine. The rotors always touch first at the gear end (discharge end) of the blower because it is considerably hotter than the inlet end, which you can think of as air-cooled.

Kenne Bell's soon-to-be patented cure is to water-cool the front, or discharge end, of the big 3.6-liter compressor. The inlet end remains air-cooled by the incoming air, while the water takes away excessive heat at the discharge or gear end. Kenne Bell denotes the water-cooled blowers by the initials "LC "for liquid cooling.

Photo Gallery

View Photo Gallery

Architecturally, the cooling water could be part of a self-contained system, shared with the engine cooling water or borrowed from the charge cooling ("intercooling") water-to-air system. Kenne Bell elected for the latter, tapping into the charge-cooling system. This keeps piping and hoses to a minimum and doesn't require yet another heat exchanger in front of the radiator. The charge cooling water runs coolly, too, so it gives the blower maximum chill.

Another change is the use of larger bearings and straight-cut rotor gears. The larger, heavier rotors require the larger bearings, and also drive the need for straight-cut rather than the helical gears found in other Kenne Bell superchargers. The new drive gears are housed in a new billet front bearing bulkhead; not the old cast bulkhead of the 2.8-liter.

The gear issue comes during deceleration. Helical gears induce a thrust or twist on one rotor, and with the 3.6's heavier rotors this effect is sufficient to drive the rotors out of time, more quickly crashing them together when they get hot. Straight-cut gears don't induce the thrust and migrating rotor timing is thus not an issue. Furthermore, the straight-cut rotor gears are even stronger, and, of course, they are noisier than helical gears. Jim Bell says the 3.6-liter is quieter than a straight-cut geared centrifugal, and we'll have to take his word on it as the only 3.6 we've heard to date was on the test engine outlined in this article, and it was running 1,000 hp worth of open exhaust in a cinder block enclosure. Rob Zombie could have been demon speeding in the data room when this super beast was bellowing and you wouldn't have heard him.

Yet another high-boost consideration of the 3.6 are its oil seals between the front bearings and the rotors. These seals can be caught in a real pressure differential between the ear-popping rotor cavity and atmospheric pressure in the drive chamber, so for the 3.6 Kenne Bell ported the drive chamber to equalize the pressure on the seals and hugely reduce the chance of their failure. This is called Seal Pressure Equalization in KB-speak.

Something Jim Bell would like us to remember is his newest blowers use his latest rotor profiles-the fine detailing of the rotor shapes-for maximum efficiency. Furthermore, he says his 4x6 rotor pack is inherently superior.

Jim Bell also reminded us that his 2.8 blower already requires notching the factory hood reinforcement structure for clearance; the bigger-yet 3.6 will demand about a 2-inch-rise aftermarket hood or lowering the engine. Interestingly, Jim says this is a deal-breaker for Corvette owners, but hardly a hiccup for the modification-mad Shelby GT500 crowd. "I already have a big hood!" is their response.

Thanks to the 3.6's size and appetite for air the inlet casting required enlargement, so it's actually a new, slightly larger casting detailed to smooth airflow into the 3.6. It's still called the Mammoth, however. Other than those changes the 3.6 blower is a 2.8 kit with the larger supercharger.

Clearly the new 3.6-liter Kenne Bell supercharger is a superior belt-driven mouse trap when it comes to making massive power on GT500s. If mega-power is your thing, the 3.6 seems a great way of getting it.

Of course, hooking this sort of power to the pavement without breaking anything is between you and your chassis builder; or your right foot where you learn to lead a sensitive life and consider tire-spin a cheap fuse for driveline breakage.

Pricing for the 3.6 blower is $6,899, well ahead of the 2.8H option, which remains available as the best GT500 choice for 16 pounds or less (pump gas) applications. More details on the 3.6's construction and dyno test are in the photos and captions; here's to big boost!

The Engine
John Mihovetz at Accufab gets credit for the beastly 5.4 under the 3.6 blower in this test. Commissioned by Mark Meiering to power his street/strip GT500 and assembled by Accufab tech Fred Grochulski, the engine starts with an exotic Ford GT aluminum block machined for wet-sump oiling and a standard starter, but from the head gaskets up its standard GT500 fare.

Internally it's all good stuff, of course. There's a stock eight-bolt crankshaft swinging forged, Manley rods and JE pistons with a healthy 10:1 compression ratio. Stainless steel rings are used.

The GT500 heads-the castings off Mark's Shelby-were CNC-ported. There's nothing trick in the porting department says the Accufab crew; just really small ports with places where the CNC didn't touch. These were hand-finished, so they are better than stock, but not radical. The valvetrain, except for John's trick (smaller than you think) camshafts, is all stock Ford.

Mark's stock intake manifold is used-that's what the Kenne Bell bolts to-with the supercharger kit taking over from there. One change for dyno duty is the throttle body is mechanical, with a throttle cable rather than the GT500's electronic unit. Fueling came from 95-lb/hr injectors on Evolution rails.

Headers have replaced the GT500 cast-iron manifolds. In the Shelby Mark runs 1 7/8-inch long-tubes, but on the dyno the engine was fitted with 2-inch primary beauties. They seemed to work great but won't fit in the car.

Two things intrigued us about the basic engine. The first was the stainless-steel piston rings. John likes them for durability in these hot-rodded blower applications, and notes they require a bit of brutality to break in. It takes immediate cylinder pressure to push the rings into the cylinder liners and get the hard rings to "cut-to-fit," and that cylinder pressure comes from WOT and blower boost. If you pussyfoot around during break-in, the hard rings will cut down the soft cylinder liners and never seat, giving you a lazy oiler of an engine. As a result, this engine went from freshly assembled to WOT on the dyno for break-in purposes.

Our other question was how in the world does John, who was doing all the electronic tuning with admittedly high-buck Motec software, get away with combining gasoline, 10:1 compression and 25 pounds of boost? "Everything has to be just perfect," was his quick response, noting the dyno session used expensive 116-octane C16 leaded race fuel, and that 16 pounds of boost was the limit with 93-octane pump gas.

By "perfect," John was referring to the electronic tune, which he carefully developed with extensive, expensive real-time monitoring and environmental tools (wideband oxygen sensors, dyno-controlled cooling water, and so on). More fundamentally, John has found the Four-Valve Ford's combustion chamber to be detonation resistant if a flat-top piston is used-he pushes the piston up in the cylinder with a relocated piston pin rather than putting a raised dome on it. "You could never do this with a Two-Valve," said John, because the Two-Valve lacks the Four-Valve's centrally located sparkplug and even airflow characteristics.

John also gave us a powerful clue to modular power production by noting he's found raising the compression ratio absolutely fundamental in his Four-Valve experience. He tested this theory by building otherwise identical engines for his own turbo'd 4.6 racer, one with 8.5 compression and the other with 12:1. The higher compression gained over 300 hp!

On the Dyno
This was a Kenne Bell test, but the crews from Westech and Accufab were there and hands-on as well. It was something of an all-pro event. Besides simply proving and tuning the engine for owner Mark Meiering, the session was designed to test the new 3.6-liter blower relative to the proven 2.8-liter Kenne Bell supercharger.

Jim Bell always takes his testing seriously, and for this test he wanted to duplicate conditions as precisely as possible between the two superchargers. This meant locking down the ignition timing and adjusting the air/fuel ratio as tightly as possible while also controlling air and coolant temperatures. Above all it meant getting the boost pressures equal between the tests, which meant a fair bit of pulley and harmonic damper changes as the two blowers have rather different outputs, as you'd expect.

The whole point was to both verify the 3.6 delivers more useful power than the 2.8, and that the power increase is from superior supercharger efficiencies rather than greater boost.

With the engine bolted to Westech's number one Superflow dyno the festivities began with 2.8-liter blower. It was run with a variety of pulleys and thus boost pressures to document its usable working range. To pick one run out of a hat the engine made 994 hp at 19.1 pounds of boost, with the air-charge temperature starting at 103 degrees at the beginning of the run and terminating at 136 degrees at the end of the run. This was a little high for the 2.8 blower, which John pointed out, typically starts at 110 degrees and ends at 125 degrees at the "normal" street/strip boost levels.

John carefully monitored the air-charge temperature so the runs would start as close as possible to the same temperature; there were no trims in the engine software to vary the fuel or spark relative to the air temperature. The same was true with the engine coolant temperature; the idea being to hold all possible variables constant so only the power or supercharger would change. The lofty goal was 1-percent repeatability, or 10 hp in a 1,000hp engine.

To this end the ignition was locked at just 19 degrees of advance throughout the testing. Clearly, at 16 pounds of boost you could run 23 degrees of ignition timing and gain power, but you cannot run that much timing at 22 pounds of boost. To change the timing from low to high boost would introduce another variable, so you're stuck running the lowest necessary ignition timing throughout the entire test.

As John pointed out, at 16 pounds boost you could likely get another 30 hp with 23 degrees of timing compared to the 19 degrees run throughout this test. The same could be done with the 3.6 supercharger at the same boost level as well, so the 19 degrees of ignition timing didn't affect the power differential between the two blowers, which was the point of the test.

Once verified at various boost levels, the 2.8 blower was removed and the 3.6 installed. Aside from grunting the heavy blowers on and off the engine-definitely a two-man job-and connecting cooling water to the front of the 3.6 blower, the change was like an intake manifold swap.

Four blower pulleys and two crank pulleys (harmonic dampers) were run on the 3.6. Boost varied from a mere 13 pounds to a rootin' tootin' 25 pounds. Of course, at 13 pounds the 3.6 blower was just mumbling along, whiffing out 989 hp, while the 25-pound pulley set resulted in 1,126 hp.

Examining the data, the supercharger dyno runs that most closely match, 2.8 and 3.6, are compared below:

2.87.1 x 3.1215,92913.1859692116
3.67.1 x 4.0012,42513.3904721122
2.87.1 x 2.7518,07216.6932787125
3.67.1 x 3.5014,14017.1989822131

Remember the ignition timing was locked at 19 degrees and the air/fuel ratio was manually adjusted to 11.5:1 to reduce the spark and fuel variables. Clearly the larger 3.6 blower didn't have to turn as fast to make the same boost, plus its efficiencies from the straight-cut gears and other tweaks delivered greater boost at less supercharger rpm. And by the way, 18,000 rpm is about as fast as Kenne Bell wants to turn its superchargers, and that's about 6,000 rpm more than suggested by the twin-screw patent-holder, Lysholm in Sweden. But all good hot-rodders dabble in the high-teen blower rpm range for sprint applications such as drag racing.

To put a round number on it, the 3.6-liter blower gets an extra 45 to 55 hp to the flywheel compared to the 2.8 version. Jim Bell's rule of thumb is the 3.6 makes the same power as the 2.8 at 2 pounds less boost. He also says his data on his in-house supercharger dyno and Dynojet chassis dyno indicates the 3.6-liter requires about 75 hp less to drive it than the 2.8 blower at 23 pounds of boost. Such exact numbers are tough to determine, but in any case the 3.6-liter is clearly the better choice for bigger boost levels.

Looking a little closer at the 3.6 supercharger tests we have the following data:
3.67.1 x 4.0012,42513.3904721122--
3.67.1 x 3.5014,14017.1989822131+9
3.67.1 x 3.2515,26019.51,035883137 +6
3.67.1 x 3.0016,52021.51,086952146 +9
3.68.0 x 3.0018,62025.01,1241,032175+29

Here we see an evenly spaced progression as boost increases, until we get to well into the 20-plus-psi boost range where the charge air temperature (the figures listed are obtained at the exit of the charge cooler) starts running away. A partial culprit, Jim Bell figures, is the throttle body. It becomes a restriction over 1,100 hp, proven by 1.7 inches of vacuum at high rpm on the 1,124hp run. Earlier runs showed negligible vacuum at redline. That makes sense, in addition to the blower simply running out of its range when spun over 18,000 rpm.

And what's this 4.2 blower? It's another, larger-yet version of the 3.6 supercharger architecture for '07-'10 Shelbys, '03-'04 Cobras, and '05-'10 Mustang GTs. Longer by 0.61 inches than the 3.6, the 4.2 is otherwise identical with the latest rotor profiles, water-cooling, and seal pressure equalization. Fans with a serious power streak should look for the 4.2 a bit after the 3.6 hits the market.

Belts And Pulleys Maybe one of the more amazing facts about the 3.6-liter supercharger and the big power it supports is just how much Kenne Bell gets out of an 8-rib pulley and belt combination. They made over 1,000 hp with an 8-rib during the test we attended, and in fact, the 10-rib hardware only came out for the final 20-psi pulls.

This is helpful as 8-rib belt selection is plentiful, but 10-rib belts are limited in the variety of available lengths and can be difficult to correctly size to an engine.

In fact, Kenne Bell says its 6-, 8-, or 10-rib drives can be used according to the following boost schedule:
6 8-15 pounds
8 8-26 pounds
10 8-32 pounds
The Price Of Power With its new internals and larger size the new 3.6-liter supercharger and the forthcoming 4.2-liter simply cost more than their 2.8-liter little brothers:
2.8 $5,999 $6,649
3.6 n/a $6,899
4.2 n/a $7,099

All 3.6 kits are polished, so there is no "standard" black finish available as there is with the 2.8 blowers. Kenne Bell says the company will begin shipping 3.6 kits January 2010.