Car Corner
Continuously Variable Boost Supercharger

September 1, 2004
By Scott Lewis

Last month I told you my feelings about the Mini Cooper S. I like the car a lot. I would need some more time behind the wheel to get used to its lack of low RPM torque. I even told you about an idea I had for a nitrous system. I would like to expand on that just a little. Then I want to move onto a new idea I have been toying with, a Continuously Variable Boost Supercharger, or CVB Supercharger.

The Nitrous Solution

If you recall, last month I suggested a nitrous system that gave two levels of boost. All the dual stage nitrous systems I have read about all look at providing less horsepower at low RPM, low speed situations. Basically the idea is to launch a car off the line without sending the tires up in a cloud of smoke. Once the car has sufficient traction more nitrous is added. This provides the best performance, especially in a drag race type situation.

My idea is the opposite. We have a car (in the case last month it was a Mini Cooper S) lacking torque in the low RPM range when there is little or no boost available from the supercharger. My idea was a nitrous system that would be more for "spirited" driving than for racing. Inject some nitrous right off the line. Not a lot, about 50 horsepower worth, maybe even 40. This would provide some extra punch, without shredding the tires. At least that's the goal here. Then, as boost increases we would REDUCE the amount of nitrous injected into the engine. Basically replacing nitrous with supercharger boost.

I mentioned this as a dual stage system last month. Cutting the amount of nitrous in half at a few pounds of supercharger boost, then cutting the nitrous off when enough boost was available to make the engine come alive.

With a system like this we use just enough nitrous to make the car a lot more fun to drive while using little enough nitrous to make it something you could leave on all the time.

What would be ideal is a system that controls the flow of nitrous more precisely. Why couldn't a computer be used to run the nitrous. You start at some amount of nitrous at full throttle but little or no boost. The system will reduce the nitrous precisely as boost increases, until the nitrous is no longer necessary. I think a system like this should not be too expensive. With some nitrous systems selling for 500 bucks, a system like this should be able to go for about a grand.

Continuously Variable Boost Supercharger

This leads me to the intent of this article. My idea for a different supercharger. Superchargers are belt driven turbines that increase the airflow to an engine above that of normal air pressure. Normal air pressure is 14.7 psi. So A supercharger that is causing the engine to "feel" 20 psi is generating 5.3 psi of boost.

Since a belt drives the supercharger it is tied to the speed of the engine. We will drastically over simplify things here. Let's assume for a moment that a supercharger has to reach at least some level of speed (measured in revolutions per minute, RPM) to actually produce positive boost. Let's say for example that an engine starts to get boost at 1500 RPM (engine speed). 1500 RPMs is 0 boost (as apposed to vacuum). And the engine has a redline of 6000 RPM. If we want to get 6 psi of boost, we have to adjust the speed of the blower to provide 6 psi of boost at 6000 rpm. In belt driven superchargers the boost is reasonably linear. Meaning as RPMs increase so does boost at nearly a constant rate.

There are far too many variables to make a blanket statement here. Yes, some superchargers increase boost exponentially as RPMs increase, but that is not important for this discussion. Here me out.

So we have a test engine. It generates 0 psi at 1500 RPM and 6 psi at 6000 RPM. If we map that out we get the following if the boost is linear:

2000 RPM = 0.7 psi
3000 RPM = 2.0 psi
4000 RPM = 3.3 psi
5000 RPM = 4.7 psi
6000 RPM = 6.0 psi

At 2000 RPM we only have 0.7 psi of boost. If we wanted more low RPM torque out of this engine we would need to increase boost at 2000-3000 RPM. If we tried to get 3 psi of boost at 2000 RPM, we would be looking at generating upwards of 25 psi at 6000 RPM. That would destroy the engine.

Well, what if we could have 3 psi of boost at 2000 RPM, and still have 6 psi of boost at 6000 RPM. That's were the idea for the Continuously Variable Boost Supercharger comes in.

With a Continuously Variable Transmissions (CVT) you have essentially two huge pulleys. These pulleys are cone shaped and expand and contract. When one pulley is large the other pulley is small. When the smaller pulley gets larger the larger pulley gets smaller. You can see this in the diagram that I got from this site. Trust me they do a far better job of showing you how a CVT works. The bottom line is that you get a transmission that changes its "gear ratio" continuously.

Superchargers use pulleys and also have a gear ratio. Why can't we use this same technology for the belt drive of a supercharger. At low RPM the supercharger spins faster, and at high RPM the supercharger would turn slower... relative to engine speed. The pulleys here would start out small on the supercharger and large on the engine. As RPMs increase the pulley on the supercharger gets larger and the pulley on the engine gets smaller. Clearly we want to maintain a reasonable speed on the supercharger at high RPM, so we would maintain some limit to the difference in the speed between the engine and the supercharger.

We could use the same principals as the CVT for the CVB Supercharger. I think it could work. The trouble is getting the pulleys to react at the proper engine speed.

Another possible way to execute a CVB Supercharger would be to drive the belt with something other that the engine. What if we mount an electric motor in the car. We run a belt between the supercharger and the electric motor. The electric motor is computer controlled. It is spun at exactly the RPM we need for the current conditions (engine speed, load, etc.). This is similar to the way computers run cars today. They read a bunch of sensors like throttle position, air flow, temperature and engine speed. They use all this information to calculate exactly how much fuel the engine needs. This is a "drive by wire" system where the throttle pedal is not directly connected to the engine or its throttle body (the device that truly controls air flow into the engine).

Why couldn't the same programming be used to spin a supercharger. We build a "map" to determine how much boost we want at any given situation. If the map says provide 5 psi of boost the electric motor spins fast enough to provide that level of boost, regardless of how fast the engine is turning. We could generate more boost at lower RPMs than we do at higher ones. How about an engine that makes 10 psi of boost at 1500 RPM, but only 3 or 4 psi of boost at 6000 RPM. This could create an extremely torquey engine. With such an abundance of low RPM torque we could gear a car to operate in a lower RPM range for better fuel economy.

With this system it should also be easier to package a supercharger since we don't have to hang it off of the front of the engine. We just need to put it under the hood somewhere.

All we would have to do is beef up the engine internals to handle the boost at low RPM. We could build an engine that acts very much like a diesel... lots of low RPM torque but better RPM range because we don't over boost the engine at higher RPM. In fact, this sounds like a great towing engine.

If we use electronics to control the boost at any point in the operating range of the engine we could easily use a knock sensor to get the most boost before detonation. This would allow a system to make the most use of available fuel. If you put better fuel in the car the boost will be turned up. Put bad gas in and the boost would be turned down to protect the engine.


There you have it, my ideas for a Continuously Variable Boost Supercharger. This could revolutionize the industry. We could build smaller, boosted engines for cars to improve fuel economy without sacrificing performance. We could build great towing engines without the noisy, smelly diesel issues. And we could build some performance cars that have such huge, flat torque curves they could make race cars jealous. The only question is which way do we build it, CVT style or with a computer control electric motor.

If I win the lottery I will try to build such an engine. As it stands now someone else will have to get it off the ground. Just like they did with the satellite radio idea I had back in July 1999.