Ultrasonic flow measurement can provide a non-intrusive way to determine flow by comparing how long sound takes to travel through a fluid in opposite directions. In large industrial pipes, however, reliable time-of-flight (TOF) measurement depends on the complete signal chain: transducer drive, receive amplification, channel switching, waveform acquisition and careful impedance matching. TI’s TIDA-01486 reference design illustrates how these elements can be combined for a transmitter serving two pairs of piezoelectric transducers.
The design uses an MSP430FR6047 microcontroller with an integrated Ultrasonic Sensing Solution (USS) module. Its added circuitry provides drive pulses of up to ±12 V at frequencies up to 2 MHz, and up to 50 dB of total gain. In testing with a 10-cm pipe and four in-line transducers, the reported zero-flow offset drift was around 50 ps, with single-shot standard deviation around 40 ps. Those figures describe that test configuration, not a guaranteed result for every meter or installation.
How time-of-flight ultrasonic flow measurement works
A sound wave traveling with the fluid arrives sooner than one traveling against it. A meter estimates flow from the difference between the upstream and downstream propagation times. The transducers may be installed inside the flow tube, arranged to reflect the signal, or mounted outside the pipe in a clamp-on configuration. Clamp-on systems avoid direct contact with the fluid, but the signal must also pass through the pipe wall, adding attenuation.
Large pipe diameters can also increase signal attenuation. Using two transducer pairs is one approach to accommodate this challenge. The TIDA-01486 board supports two pairs, with transducers operating sequentially as transmitters and receivers. Ultrasonic meters require a direct acoustic path between transducers; air bubbles can significantly attenuate the signal and interfere with measurement.
Signal-chain design: drive, receive and switch
Transmitter and receiver stages
Each piezoelectric transducer can transmit or receive. The design uses four THS3095 high-speed amplifiers to drive and receive signals for the two pairs. These amplifiers support the required high-voltage drive and are also used in the receive path. Using the same amplifier type in both directions helps avoid introducing a differential time delay between transmit and receive paths, including under zero-flow conditions. The shared signal path also means the receive amplifier must tolerate the transmit voltage.
On reception, a non-inverting amplifier stage supplies 20 dB of gain. A 750 mV bias brings the signal into the range needed by the MSP430FR6047, allowing a single-supply analog switch to be used without a negative supply. Clamping diodes limit the signal presented to the switch and microcontroller inputs. The design also uses a capacitor in the feedback path to keep the DC bias from affecting the amplifier’s gain calculation.
Switching between transducer pairs
The switching stage routes the transmit and receive signals to the selected pair and direction. Low switch on-resistance matters because the transducer is a low-impedance load; added resistance can influence the measurement. The design uses TS5A3357 analog switches, specified in the reference design as 5 Ω, and GPIO signals from the MSP430FR6047 control the switches and amplifier shutdown pins. Sequential operation allows the board to select either pair and configure its upstream or downstream path.
Why impedance matching and PCB layout matter
For stable TOF measurements, the electrical conditions seen by a transducer should be as similar as possible during transmission and reception. The design guidance recommends examining the circuit from each transducer pin back into the board and making the TX and RX paths look alike. Any mismatch can contribute to zero-flow offset drift, which may vary with temperature.
Both component choices and physical layout influence that symmetry. The reference design recommends using 1% resistors, particularly for the 200 Ω components, and capacitors with 5% tolerance or better. It also calls for matched trace lengths between corresponding channels, symmetrical routing and a ground plane beneath signal traces where possible. The board guidance further advises avoiding vias on transmit and receive signals and removing copper fills beneath op-amp inputs and outputs.
Matching paths from the multiplexers to the microcontroller header is also useful: fixed channel-to-channel delay differences can otherwise require calibration. These considerations are not merely cosmetic. In a precision timing system, small electrical differences can appear as timing offsets.
ADC-based processing for robust TOF estimation
Rather than measuring only a single threshold crossing, the ADC-based approach captures the received waveform for upstream and downstream measurements. Digital processing then correlates the waveforms to estimate their relative timing. The MSP430FR6047’s USS module supports high-speed acquisition, while its Low-Energy Accelerator (LEA) can perform signal processing independently of the CPU.
Correlation can provide filtering that suppresses noise and interference, including line noise, and the reference describes lower measurement variation than approaches without this processing. Because the algorithm uses waveform information rather than relying solely on signal amplitude, it is also less sensitive to amplitude changes caused by flow conditions, transducer variation or aging. Captured amplitude information can help assess the signal envelope and support techniques such as automatic gain control.
Temperature is another consideration. Sound velocity in water changes with temperature, so a flow calculation that depends on acoustic velocity may need temperature compensation. The reference design also discusses using absolute upstream and downstream TOF, alongside their difference, as an alternative approach that can avoid requiring a temperature sensor to determine sound velocity. The method depends on accurately extracting absolute TOF from the received signal.
Practical implications for system designers
The TIDA-01486 demonstrates that ultrasonic flow measurement is a coordinated system-design problem. Engineers evaluating a similar architecture should consider:
- Pipe geometry, transducer placement and signal attenuation when selecting in-line or clamp-on arrangements.
- Transmit voltage, receive gain and protection requirements for the specific transducers and acquisition inputs.
- Switch resistance and matched electrical paths between transmit and receive operation.
- Waveform acquisition and processing requirements for the desired timing stability.
- PCB symmetry, component tolerances and temperature effects that can influence zero-flow offset.
In the reported test, the drive level and external gain were adjusted for the particular pipe and transducers, underscoring that reference-design settings may need adaptation. The practical takeaway is to treat the mechanical path, analog front end, digital processing and board layout as one measurement system. Matching those elements—and validating them with the intended pipe, fluid and operating conditions—is central to achieving repeatable time-of-flight measurements.