Since Gordon Moore, co-founder of Intel, postulated the law that bears his name about doubling the number of transistors every two years, there have been dramatic improvements in the computational capabilities of electronic devices. Reducing component sizes coupled with increased demand for computing power has resulted in ever-increasing power densities, requiring optimized and advanced cooling configurations to maintain a safe operating temperature. Electronic thermal management is a separate topic and beyond the scope of this blog. However, I would like to discuss one of the consequences of increased electronic performance – noise!
Anyone who works during the summer in an office is no doubt accustomed to their computer or laptop fans spinning as the number of applications running in parallel increases, or even more fun, getting to run advanced Computational Fluid Dynamics (CFD ) or Simulation using the Finite Element Method (FEM). Fan noise, while simply considered an unavoidable inconvenience, is the result of a complex interaction between the fan itself and the airflow it generates. For this reason, it can sometimes be called flow-induced noise or aeroacoustics.
Despite the need to properly cool all these electronic components in the cramped space of a modern laptop, people have come to expect the noise generated to be non-intrusive. Noise-canceling headphones can help you here, but they are far from an ideal solution during a hot summer day. What's more, acoustic performance has become one of the key indicators of high-quality laptop brands – combining the gentle hum of fans with a set of clear, vibrant, well-placed speakers that play your favorite tunes.
This puts a lot of pressure on the engineers who develop these systems. Let's discover how state-of-the-art acoustic simulation tools can help dedicated engineers predict acoustic performance earlier, faster and more reliably.
The inevitable “bad” sound…
When analyzing the noise signature of a fan, there are typically two components, tonal noise and broadband noise, as shown in Figure 1. Tones can be clearly visible as higher sound pressure levels resulting from periodic interactions of the incoming air. with the fan blades (blue circles). The broadband noise component is caused by random loading forces on the blades that can be induced by factors such as turbulence ingestion or boundary layer development (green line).
Figure 1. Free-field response detailing tonal peaks (blue circles) and broadband noise (green line)
Considering that fan noise is the result of the interaction between aerodynamic flow and acoustic wave propagation, both airflow and acoustics need to be simulated. Acoustic wave propagation can be included directly in a CFD simulation already used to evaluate the design's cooling performance, but this – although possible – can present significant challenges. These challenges are mainly caused by the significant differences in length scales between the acoustic waves and the flow. This means that high-order physics schemes and exceptionally long calculation times are required, so this approach is not always viable.
Hybrid approaches have been developed in response to this, in which sound generation and sound propagation are separated. CFD data is used to reconstruct sound sources due to flow effects, while acoustic simulation models are used to propagate sound waves caused by these sources. This offers the advantage of enabling more efficient low-order flow simulations and taking advantage of efficient acoustic resolution technologies.
Figure 2 illustrates how a model is prepared for an acoustic analysis showing the air finite element mesh around the laptop (2A.), the internal mesh (2B), the connections between the inside and outside, the ventilation openings (2C) and the inlet grille under the laptop (2D). The next step in an acoustic analysis is to define the source region – this can be obtained from CFD or directly from the test data.
The equivalent acoustic source is calculated using Simcenter 3D and introduced into the FE model. Once resolved, the sound field generated within the laptop and radiated from it can be analyzed. Simcenter 3D allows the acoustic engineer to understand how sound leaves the laptop, its direction, and also reflections from the immediate environment.
Figure 2. FE mesh for acoustic simulation of a laptop: 2A the air around the laptop; 2B the internal mesh of the laptop geometry; 2C the laptop openings and 2D the entrance grille.
Notebook OEMs need to understand the sound generated by the cooling architecture and investigate ways to minimize the impact on the user, such as directing noise away from the user through a rear-facing jack. Additionally, sound engineers can understand how the laptop screen shields some noise at various screen angles and user positions, as well as in the closed position if docked or connected to other displays.
… and the much sought after “good” sound
As mentioned in the introduction, the sound quality of a laptop is considered an indicator of the brand's high quality. Therefore, it is pertinent for the engineer to understand the behavior of the speaker and how it works in the laptop chassis. To optimize sound and maximize quality for the user, the engineer must start with the notebook's stand-alone speaker and work through subsequent integration into the notebook chassis to the notebook's behavior in a realistic user environment – see Figure 3.
Figure 3. Acoustic simulation steps to ensure the fidelity of results at the operational level when the product is in use.
At the speaker level, as in all simulations, a geometry is defined from which the FE model is created and merged. This structural vibration model of the speaker is then coupled to a small volume of air near the speaker membrane. Specific acoustic radiation conditions are applied to the external surface to allow prediction of far-field sound radiation characteristics. Simplified 1D models based on the Thiele-Small model are used as inputs for the coil loads. These models contain all relevant electromagnetic coupling effects in the speaker driver, and their input parameters are easily obtained from the supplier (or from simple measurements).
After solving the model, the sound radiation can be analyzed and post-processed to provide directivity data, impulse responses, and distortion data. Figure 4 provides a graphical representation of this typical workflow.
Considering the performance of the speakers and the association with the perceived quality of the laptop, the engineer would be interested in quantifying the intensity of the acoustic source and the uniformity of the speaker's sound field. The image on the right of the video above visualizes the radiated sound waves.
The next step is to understand how integrating the speaker into the laptop affects acoustic performance. In a laptop, the behavior of the speaker is strongly influenced by the coupling of the speaker membrane with the volume of air behind it and by the viscothermal effects that occur in the grilles that cover and protect the speakers from dirt and dust. The speaker model in the video is therefore extended to also include the back of the speaker membrane to model the interaction with the rear cavity inside the laptop and the effect of the grille and air volume between it and the speaker membrane, as shown in figure 4.
Figure 4. Visualization of a possible laptop speaker configuration illustrating the air volume or rear cavity behind the speaker and speaker grille.
The grid effect can be explicitly simulated by:
Model the fluid in the holes and apply specific visco-thermal properties of the fluid or
Using simplified equivalent transfer admission relations.
Figure 5. Comparison of speaker performance before and after installation.
Figure 5 illustrates the effect that installation conditions can have on speaker performance after you integrate it into your laptop. For the isolated speaker, the left image in Figure 6 shows that the speaker source intensity is uniform above one kilohertz. In comparison, the image on the right of Figure 6 illustrates a performance degradation above one kilohertz. There is very low sound radiation between the frequency range of four to six kilohertz and this is explained by the interaction between the speaker membrane and the resonance found in the rear cavity. This is a more realistic assessment of the performance of the speakers in the laptop and provides design engineers with valuable information to further optimize their product.
The final step in the process is to evaluate how the laptop will perform in the intended user environment – a typical office, for example. However, doing this with a finite element model would require significant time and computing power. An alternative method is to use Ray Acoustics, one of the advanced acoustic solvers available in Simcenter 3D. This technology is based on ray tracing, allowing us to effectively simulate the propagation of sound in wide spaces, over long distances and at high frequencies, much faster than finite element or boundary methodologies ever could.
The model discretization and solution times are independent of frequency, making it perfect for solving problems where the geometry is larger than the acoustic wavelengths. Simcenter 3D provides frequency and time domain results as output from this solution. To simulate the office environment, three main modeling features are available in Simcenter 3D:
Edge and surface diffraction – useful for typical dividing walls in an office environment
Curvature effect correction – accurately captures discretized and mesh surfaces
Absorption – surface and air absorption
Particle tracking – takes into account late reverberations and diffuse reflection effects typically found indoors
The lightning acoustic simulation model can directly calculate sound quality parameters such as reverberation times, clarity values or sound transmissibility indices. Simcenter 3D can also directly incorporate binaural effects into the acoustic response without the need to model the human head – essentially obtaining the sound pressure levels that the listener's left and right ears experience.
Figure 6. Typical office environment with visualization of sound propagation – individual ray paths are visualized and the binaural impulse response
Figure 6 (left) illustrates a typical office environment with all reflective or absorbent surfaces discretized using a simulation mesh (gray surfaces). The microphone surfaces near the laptop and the subject's head are set to visualize the sound fields, as seen in the upper right corner of Figure 6. Ray tracing models provide information about how different speaker combinations perform. spread to the person's ear. A clear view of how much sound is radiated to the user and how much is being reflected from different surfaces can be unlocked using these ray tracing visualizations.
Putting it all together – a cacophony or a symphony?
All simulation steps discussed provide quantitative and visual information about the acoustic performance of the individual component through to product integration and incorporation into the real-world environment. However, despite everything, isn't it better to be able to hear the results of the simulations?
Simcenter 3D Acoustics offers a sound processing and auralization tool that takes simulation results and combines them with measured sounds, like music, to create acoustic scenes you can hear!
Only fan noise
Fan noise + low-cost speaker
Fan noise + high-end speaker
The videos above allow you to test three different scenarios from a laptop:
Fan simulation noise
A piece of music played in the office environment using a low-cost speaker
The same song played on a high-end speaker.
In the first scenario, both tonal and broadband components of the noise are present – auralization allows you to check how loud the sound sounds and how irritating or distracting it can be. In scenario two, one can investigate how some music can mask the sound – the fan noise is masked, but some low-frequency components of the music are missing. Scenario three employs a state-of-the-art speaker, providing much clearer and richer sound to the music.
Simcenter 3D Acoustics allows you to understand the sound quality and noise of your electrical components. With these features, you can design around inherent noise and give the user of your products a more pleasant listening experience.
Interested in optimizing the acoustic performance of your electronic products? Schedule a meeting with CAEXPERTS now and discover how our state-of-the-art acoustic simulation tools can help you predict, understand and improve the sound behavior of your devices quickly and reliably. Don't let noise get in the way of the perceived quality of your products - let's work together to ensure an exceptional listening experience for your users. Schedule your meeting today!