This article is sponsored by DEWESoft. In this Voices interview, Design World spoke with John Hiatt, business development manager and applications engineer at DEWESoft, about the measurement technologies engineers use to analyze rotating machinery and complex dynamic systems. He shared insights on how modern data acquisition systems and analysis techniques help engineers diagnose performance issues, improve reliability and understand the behavior of rotating equipment.
Design World: Could you start by telling us about your role at DEWESoft and the company’s core product offerings?
John Hiatt: I am a business development manager and applications engineer at DEWESoft. I describe myself as an applications engineer because I am more on the technical side rather than commercial.
DEWESoft is a data acquisition company. We take sensor signals, which could be anything from temperature and pressure to vibration, strain or sound, condition those signals, digitize them and then analyze them. Our systems cover a very wide range of measurements, from one sample per second to about 15 million samples per second, and from microvolt-level signals up to 2,000 volts.
We also provide a software platform called DewesoftX, which engineers use to configure acquisition systems, visualize signals and perform both real-time and post-processing analysis. One unique aspect of our platform is that the software license is tied to the hardware. Once the data is recorded, the license travels with the data file. That means anyone can open and analyze the file without needing additional software licenses.
Our systems are used across many industries because almost every engineering field measures physical quantities related to performance. My focus tends to be on dynamics, sound and vibration, and stress-strain analysis; basically, signals that change with time.
How has modern order tracking improved the way engineers analyze rotating machinery behavior?
Order tracking allows engineers to relate dynamic signals like vibration or acoustic signals to a rotating component.
A simple example is a driveshaft in a car. If the driveshaft is rotating at 3,000 RPM, that corresponds to 50 Hz. With order tracking, you can immediately identify the first order, which represents a once per rev vibration signal. If there is an imbalance in the driveshaft, that will show up clearly at the first order line in the order spectrum.
What makes order tracking powerful is that it helps engineers quickly identify which rotating component is generating a particular vibration or noise. For example, you can determine whether the issue comes from a tire, an engine firing frequency or a driveshaft imbalance, because all these components create different sound and vibration frequencies related to rotational speed.
Rotating systems generate dynamic forces at multiples of rotational speed, such as 1x, 2x, 3x, and sometimes even fractional multiples. Order tracking makes it much easier to see which component is causing a problem.
In traditional FFT-based vibration analysis, you collect time-domain data and transform it into a frequency spectrum. Order tracking works differently. The system resamples the data to the angle domain based on rotational position. The Fourier transform is then applied relative to angle domain data rather than time.
This becomes useful when the rotational speed changes quickly. In FFT analysis, rapid changes in speed can smear the frequency spectrum because the signal varies within the processing window. With order tracking, the analysis is tied to rotation angle/revolutions, so it tracks those changes more accurately.
Companies can also use order tracking to diagnose issues. Let’s say an engineer is getting warranty complaints about a whining noise in a vehicle. The first step is to instrument the vehicle with microphones and speed sensors for components such as the engine and driveshaft.
By performing order tracking, engineers can determine which rotational order dominates the noise. If an axle gear has 13 teeth, for example, you might see a strong 13th-order component. Once that order is identified, engineers can use audio replay of the original and filtered signal to verify the issue. At this point, the analysis shifts from identifying the component to determining whether the issue is due to component quality or system-level dynamics.
What common challenges do engineers face when analyzing rotor dynamics, and how can advanced measurement tools help address them?
One of the biggest challenges with rotor dynamics is that the component you want to measure is rotating. For stationary components, measurement is straightforward; you can attach an accelerometer and record vibration. But when the component itself is rotating, attaching sensors becomes much more complicated.
You could use slip rings or telemetry to transmit the signal from a rotating sensor; however, slip rings affect rotational dynamics and telemetry can have transmission issues. Because of that, engineers use non-contact measurement techniques, such as eddy-current proximity probes. These probes measure shaft displacement at the bearing housing. They allow engineers to observe how the shaft moves inside the bearing clearance while the machine is running.
The challenge is mainly in the setup. The probes must be mounted correctly near the bearing support and calibrated with the proper gap voltage so they can measure the proper shaft distance from the probe. Once the measurement setup is correct, the analysis itself is straightforward.
When it comes to rotor dynamics analysis, the system uses two probes positioned 90 degrees apart. By plotting one probe signal against the other, engineers can create an orbit plot, which shows how the shaft moves within the bearing clearance. If the orbit were perfectly circular, it would indicate purely uniform rotation. But in reality, the orbit is usually elliptical and may rotate as speed changes. The orbit shape and orientation can reveal forward or backward precession.
Other diagnostic plots include centerline plots showing shaft centerline relative to the bearing clearance circle, order cuts 1x or 2x, sub synchronous orbits and full spectrums.
How are engineers using high-resolution data acquisition and synchronized pressure measurements to understand combustion performance?
Combustion analysis involves capturing extremely fast events. Inside an engine cylinder, combustion is an explosion that occurs in a very short period. If an engine is running at 600 RPM, that corresponds to 10 revolutions per second. In a four-stroke engine, a cylinder fires every other revolution, so multiple rapid pressure events happen within each engine cycle.
Because these events occur so quickly, the data acquisition system must sample at high rates at about 100 to 200 kHz and maintain precise synchronization across all channels. Another key requirement is having a reference signal tied to the crankshaft position.
Combustion analysis is performed in the angle domain, meaning the data is resampled according to crank angle. Engineers analyze combustion events in increments of fractions of a degree, around 0.1 degrees.
The output of combustion analysis is the pressure-volume (PV) diagram, which plots cylinder pressure against cylinder volume throughout the engine cycle. To generate this diagram, the system needs engine geometry and the crank angle position. From the PV diagram, engineers can calculate metrics such as mean effective pressure, combustion work and cylinder-to-cylinder variations. The analysis also allows engineers to detect phenomena such as knock and combustion repeatability.
In the entire measurement setup, signal integrity is critical. Cylinder pressure sensors operate at extremely high temperatures and must capture very rapid pressure changes. The data acquisition system must ensure that all channels share the same sample clock so that measurements occur at the same instant. Even small timing misalignments between channels can distort combustion analysis.
How has field rotor balancing changed with modern instrumentation, and what best practices help ensure long-term reliability?
In the past, balancing could be done manually, by adding weights to a rotating component until the vibration improved. Today, the process is much more automated. Engineers measure vibration using an accelerometer on the bearing support and use a once-per-revolution tachometer signal to establish a phase reference.
From that measurement, the system calculates vibration magnitude and phase relative to rotational position. A trial mass is then added at a known location. By measuring how the vibration changes, the software calculates the influence coefficient that tells how the system responds to mass changes.
Once that coefficient is known, the software can determine the corrective mass and location required to balance the rotor. The workflow becomes simple: take a baseline measurement, add a trial mass, calculate the imbalance and apply a correction.
A few best practices that engineers must follow to ensure long-term reliability is to sweep through operating speeds before balancing. This helps identify regions where phase behavior is stable. Balancing near resonance should be avoided because phase changes rapidly in those regions, which can make corrections less reliable.
Engineers also want to verify that the balancing correction improves vibration levels across the operating speed range, not just at the specific balancing speed. Good tachometer signals and, if you have multiple balance planes, understanding whether these planes are coupled or not is important in understanding the correct balance process to use. Typically, this will be single plane or multi-plane balancing.
What are the most effective measurement approaches for identifying torsional vibration issues before they cause serious damage?
Torsional vibration is the fluctuations in rotational speed rather than the average speed itself. For example, when an engine cylinder fires, it produces a torque pulse. That pulse causes the rotational speed to increase slightly and then decrease again. Those fluctuations propagate through the driveline and can excite torsional resonances in the system. If the excitation frequency aligns with a torsional resonance, it can create noise, vibration or even durability issues.
In automotive applications, torsional vibration can produce low-frequency booming sounds that drivers perceive as uncomfortable. In more severe cases, the repeated torque oscillations can lead to fatigue failures in drivetrain components. That is why engineers try to measure these speed fluctuations and identify potential resonances early in the design process.
The most accurate method uses rotary encoders, which generate many pulses per revolution, sometimes thousands. These pulses allow engineers to measure small variations in rotational speed very precisely. However, installing encoders directly on the drivetrain can be difficult and may affect the rotational dynamics.
An alternative approach for engineers is to apply striped optical tape (AKA Zebra Tape) around a shaft. An optical probe detects the stripes as the shaft rotates, which generates multiple pulses per revolution. Another method is using magnetic pickups that detect gear teeth as they pass the sensor.
Regardless of the method, the key requirement is generating a multi-pulse-per-revolution signal, so engineers can analyze speed fluctuations rather than just average RPM. Once the speed signal is captured, engineers analyze the oscillations in speed using spectral or order-based analysis to identify torsional resonances and design appropriate damping solutions. Guidelines are for at least 2x the number of pulses as the highest torsional order you want to measure. The more pulses the better the results, but often more pulses are not practical.
DEWESoft provides a complete solution for Rotating Machinery sound and vibration development as well as condition monitoring for early fault detection, and has no Software Support or analysis seat costs. You can find more information about our Rotating Machinery data acquisition and analysis solutions at dewesoft.com.
