It is possible, by reducing the applied voltage to an induction motor, to vary the speed somewhat. The trick, of course, is that the speed variation is very non-linear. Additionally, you don't want to cause the motor to stall out (since very bad things, such as it flaming out, may occur).
I'm told that some people have been able to add a suitable resistance in series with the motor to reduce the applied voltage. One approach is to use a suitably sized incandescent light bulb as the resistive element.
Obviously, wiring something like this will require access to line voltage, which can be LETHAL. Please only attempt this if you know what you're doing, and can do it safely, such that there is no danger to yourself or others.
There may also be some triac based motor speed control devices which will work with an induction motor. These may be a safer, although probably more expensive, alternative.
Thanks. Right now I think I'd rather put up with it instead of just putting a resistor inline. I don't want a rheostat.
For me it's that there's a speed regulator in there already, but I don't know the principle of how that one works. I was hoping that modifying the existing regulator was possible. But if that's too hard-core, I won't do it.
It isn't so much a matter of a speed regulator, but having different circuits for different speeds. Modifying a speed basically means rebuilding the motor.
Oh, so the speed regulation is built in the motor, and the switch is just signalling the stuff inside? I was assuming that the speed control was external to the motor, and that led me to misunderstand Dave's comments.
Most of the speed control is determined by the way the motor is built. Most fans use a type of induction motor. Most of these have two sets of windings, with their corresponding pole pieces. The speed is mostly determined by the number of poles. But, by playing games with the way the two sets of windings are connected, it can make it appear that the number of poles is different, which will cause the speed to be different.
There are also tricks where one of the windings is fed with a capacitor, to produce a phase shift in one set of windings (and, thus, in one set of poles), which affects the rotational speed.
However, since these motors are induction motors, they don't run at a synchronous speed [1]; there is a bit of "slip" between the synchronous speed and the actual rotor speed. By reducing the line voltage, this slip can be increased somewhat, which represents a slowing of the rotation of the motor.
[1] On a 60 Hz power system [2], for a two pole motor, synchronous speed would be 3600 RPM. For a four pole motor, synchronous speed would be 1800 RPM. For a six pole motor, synchronous speed would be 1200 RPM (And, of course, you can have more poles, which corresponds to a lower rotational speed, limited only by the complexity of the stator design and how many poles that can be carved into it.). Of course, factoring in the slip at normal voltage, the rotational speed would be closer to 3550 RPM, 1750 RPM, and 1150 RPM, or something like that depending upon the particulars of the construction of the motor. By dropping the voltage, the slip could be increased, although it's difficult to predict by exactly by how much (5%? 10%? 25%?).
[2] For a 50 Hz power system, the synchronous speed would be 3000 RPM (2 pole case), 1500 RPM (4 pole case), and 1000 RPM (6 pole case). This brings up another method of speed control. If, instead of connecting the motor directly to the power line, the power is rectified and then fed to a variable frequency inverter, such that any frequency can be produced, then it is possible to vary the speed of the motor over quite a wide range. Unfortunately, the construction of such variable frequency inverters is not for the feint of heart, given that there are some secondary design considerations which come into play.
In any case, since you can't change the number of poles in the motor, at least without cutting/stamping a bunch of new pieces of steel for use as the stator, and then winding the coils appropriately, you're pretty much stuck with the stator design as it is [3]. So, the only options are to either use the variable speed inverter approach, or to reduce the voltage applied to the motor to enhance the slip. The two most common ways of reducing the voltage is either with a large power resistor, or with a phase controlled triac.
[3] Yeah, I've always wanted to redesign/rebuild a motor, but I don't have the skills needed to cut/stamp the steel laminations for the stator. I could probably rewind an existing stator, given enough of the appropriately sized magnet wire and enough time, but that wouldn't do much to change the motor characteristics, since the number of poles would be mostly fixed by the steel laminations of the stator. There used to be motor repair shops where skilled technicians could rebuild/repair motors, but most of these have disappeared, given the modern throw-away philosophy. :-(
Dave
P.S. Check my numbers. It's been a LONG time since I've done much with induction motors.