The inductance of the motor is the issue. A motor can be modeled as a resistor and an inductor in series. At DC or zero frequency only the resistance limits the current. That limit is often 5-10 times higher than the rated running current of the motor. When you apply a step change in voltage to a series LC circuit, the current begins to ramp up. If you switch it again before the current has reached the DC limit, the average current will be lower than the DC value. It is somehat more complicated than this because when the motor is turning it also generates an internal voltage called the Back EMF (electromotive force) which also tends to reduce the current. The time it takes to approach the DC limit is a function of the resistance and the inductance of the motor winding.
To get PWM to work for a motor you need to have the frequency high enough that the current never gets close to the DC limit before the controlsignal switches again. Generally the motor specifications do not include the inductance so you need to make maesurments to determine the minimum frequency. You did that with your 555 circuit. Based on what you reported it appears that your minimum PWM frequency is about 5-10 kHz. If you cannot run your Digital I/O at that frequency, you will not get accepatble results from your PWM.
It is not a matter of limiting the maximum voltage or current at the amplifier. It is a matter of timing.
Lynn
