Our last blog post, covered why we decided to make an IEC baffle, and the practical constraints we face in our office. In this post, we will focus on the creation of the baffle, as well as testing its effectiveness.

Creation Process

After establishing the design goal and identifying our constraints, it’s time to start building the baffle. We’re using 1/4-inch hardboard for the baffle due to its flat, hard, and smooth surface, which is ideal for sound reflection. Additionally, we opt for smaller wood boards to be attached to the back, acting as a frame to provide much-needed stiffness to the baffle. Weight is an issue with the light materials we’re using in the baffle design, so it is crucial to minimize movements and vibrations through alternative means. Furthermore, we are mounting the entire baffle on a large TV stand with wheels, allowing us to easily move it around the office.

For speaker mounting, we made a rectangular cutout on the baffle using a Dremel tool and installed some hold-down clamps around it. This allows us to mount various loudspeakers and loudspeaker enclosures to the baffle with 3D printed mounts. A fast 3D printer capable of producing these types of inserts in just a few hours, greatly improves our efficiency and turnaround time in testing speakers. Custom 3D printed mounts also ensure maximum flatness and precision on the front side of the baffle.

Speaker Mount Renderings and Final Prints

Testing and Improvements

In order to test the structural resonances and damping characteristics of the IEC baffle, we make several mechanical measurements. To do this, we chose three random points on the baffle, attached an accelerometer to each point and then tap the baffle surface near the speaker with a hammer. The impulse response from the accelerometer is recorded in to the computer for analysis. To understand the resonance frequency and damping, we use a spectrogram of the impulse response. The horizontal axis is frequency, and the vertical axis is time in milliseconds. Colors represent the relative magnitude of acceleration at that point.

We can see, the baffle has many resonance modes, especially below 400Hz. Many of the microspeakers we test at Amplify Labs can’t produce these frequencies to excite these resonances so for the most part this should not be an issue. For the larger speakers that can reproduce these frequencies, we need to reduce these resonances by improving the material damping characteristics.

In order to improve damping, we experiment with putting automotive damping material on the backside of the baffle.

We made the same measurements after the application, with the results below.

Please scroll left and right to compare before and after at the three measurement points. Similar result can be observed at all three points.

Here we can see a slight decrease in damping time, about 100ms, in the frequency range from 200Hz to 400Hz. We’d like a more substantial change, but this is better than no improvement. In the future, perhaps we can use constrained layer damping to further improve the baffle.

Speaker Testing on Baffle

After construction, it’s time to mount some speakers for testing. To compare speaker measurement results with and without the baffle, we chose a 1-inch speaker enclosure.

We measure this speaker enclosure in the near field, in free field, and on the IEC baffle and compare the results.

We can see that the shape of the free field result (red curve) generally follows the “baffle step” curve shape — meaning that baffle step is indeed an issue in this type of measurement. The sharp dip at 800Hz is from the resonance of the measurement setup. If we look at where the amplitude starts to drop, we see it is around 3kHz. By measuring the distance from the center of the loudspeaker to the edge of the small circular baffle, we can get a quick estimation of when the baffle step should occur. The distance is 3.5cm, corresponding to a diffracted time delay of 0.102ms. From this, we can calculate that max constructive interference between the direct and diffracted sound will happen at 1s/(0.102 ms*2) = 4.9kHz. This seems to align closely with the peak we see at around 5kHz in the free field measurement. Destructive interference will happen when the time delay corresponds to 1/4 period or less, which corresponds to a frequency (at 1/4 period) of 1 s/(0.102ms*4) = 2.5kHz. Below this frequency we see destructive interference between he direct and diffracted sound, and thus the “step” is formed. In our measurement, the magnitude of the “step” is around 5dB, slightly less than the theoretical 6dB. This is likely caused by reflections in the room also picked up by the measurement mic.

We also compare the baffle measurement to a nearfield measurement, which does not suffer from room reflections or baffle effects. The nearfield measurement technique is not full range, and its upper frequency limit is determined by the effective cone diameter D: F_max = 10950/D where D is in centimeters. For this speaker, we determined that  F_max = 3.1kHz. As we can see, the baffle measurement follows the nearfield measurement very closely below 3.1kHz while providing more accurate information at higher frequencies above 3.1kHz.

Conclusions

The IEC baffle we conceived and built in our audio lab is making our measurements more accurate and repeatable. At Amplify Labs we prioritize making world-class audio designs for our customers. Our data driven approach ensures that we make the best design decisions possible for each project, based on the requirements of our customers and their unique goals for bringing a product to market. Noted acoustic engineers and loudspeaker designers Philip Newell and Keith Holland said, “Measurement has driven analysis, and analysis has driven design.” This is an idea that our team puts into practice regularly. As we continue to grow and work with exciting new clients, we will continue to use this approach to deliver best-in-class products to our customers and the world.