Why is driving a proportional solenoid with PWM more efficient than resistive current control?
PWM holds the switching element either fully on or fully off, so the lost power is low; in resistive control all the unwanted voltage turns into heat.
In PWM the switching element is either fully conducting or fully off; while it conducts the voltage across it is low and while it is off the current is zero, so the power spent in it stays very low. In resistive (linear) drive, on the other hand, the voltage that does not go to the coil is dropped across a semiconductor or a resistor and is turned entirely into heat. That is both an efficiency loss and a matter of heat sinks, enclosure volume and the risk of thermal failure.
Where the loss goes
| Resistive drive | PWM drive | |
|---|---|---|
| Excess voltage | Dropped across the element | Not dropped, shared out in time |
| Lost power | Turns into heat | Limited to the switching loss |
| Cooling requirement | High | Low |
| Fitting in a closed enclosure | Difficult | Comfortable |
A concrete example: think of wanting to give a coil on a 12V supply an average current of half its fully open value. In the resistive method almost as much power as reaches the coil is spent as heat in the element that limits the current. PWM sets up the same average without producing that loss.
The coil itself forms the average current
In PWM the switch turns fully on and fully off thousands of times per second; even so, the coil current is not pulse by pulse. Coil inductance behaves like a low-pass filter and smooths the current, and a steady average current corresponding to the duty cycle flows. In a proportional coil the quantity that determines hydraulic flow is in any case this average current.
Current is controlled, not voltage
As the coil heats up its resistance rises; at the same voltage the current falls, so the flow drifts with it. That is why serious drivers do not hold the duty cycle constant but correct it against the measured current. Thanks to current-controlled drive the coil is held at a constant current independently of temperature. The same drive layer can also carry functions such as ramp, minimum and maximum current limit, dither and short-circuit protection; the card we call a proportional amplifier does exactly this job and converts the voltage signal coming from the joystick into the regulated current the valve needs.
Where PWM is not free
Switching on its own does not solve every problem. If the frequency is chosen too low the current ripple grows and the valve chatters; if it is chosen too high the switching losses and the electromagnetic interference grow. Choosing the PWM frequency is done between these two limits and there is no single correct value; the inductance of the coil and the mechanical response of the valve settle it together. Also, dither is a separate component added deliberately to break the spool friction; it must not be confused with the carrier frequency. A path must also be provided for the energy in the coil to discharge the moment switching stops; that is why a freewheeling diode on coil loads is essential, and where it is missing the fault shows up in the output stage of the driver.
What to do
If you are planning to drive a proportional valve directly with voltage, give up the idea; use a current-controlled driver. When choosing the driver, look for the minimum and maximum current settings, the ramp and the dither setting to be accessible; these are the only way to fit the valve to the system. After installation, measure the temperature inside the enclosure once and note it down. If you are coming over from a resistive solution the difference shows itself in the first shift; if you are not, that measurement becomes the only reference you have when a problem comes up later.
The full topic: Why should it be commanded by an electronic signal?