Started 16th January 2021
under construction
Imagine striking a bell and listening to the reverberations go on and on, only gradually fading away. It turns out the ultrasonic transducers used in low cost ranging modules exhibit the same behaviour. This makes measuring short distances difficult.
Playing with the HC-SR04 it came as a surprise to me that whilst the drive signal had only 5 cycles the receiver oscillations viewed on the 'scope had many more cycles. But reading the datasheets [1] for ultrasonic transducers, 'ring time' is well known.
The obvious way to stop the ringing is to add a resistor across the transducer - damping - thinking about moving coil loudspeakers a short circuit would surely stop ringing instantly. Unfortunately it is not that simple.
Commonly the Butterworth-van Dyke circuit model (left) is used for ultrasonic transducers. L, C and R represent the mechanical properties of the transducer, whilst Cp is the electrical parasitic capacitance. The only electronic way of changing the ringing behaviour is to connect a device across the transducer terminals or equivalently Cp. Connecting a wire, shorting out Cp, one is left with an LCR circuit which maximises ringing. Connecting a high value resistor has little effect due to the parallel lower impedance of Cp, but somewhere between infinity and zero there is a value which results in a minimum ringing time.
If the transducer is driven from a voltage source with close to zero impedance, then when not actively driving the transducer the source will represent a near short circuit and do little to stop ringing. Components to reduce ringing can be connected in series with the signal source and transducer. Which raises the problem reducing ringing may degrade desirable behaviour. Ringing occurs in both transmit and receive modes, the problem is the same, but the way of connecting the transducer can be different.
There are a large number of papers and patents devoted to reducing ringing in ultrasonic transducers.
In the micro-controller world it is common to generate a number of pulses and then stretch the final pulse to twice the usual width, providing an out of phase "braking" pulse to reduce further oscillations.
References: