I Made A Quieter Air Purifier.
The physics of sound, and the PC fan CR box v1.
DIY air purifiers could help reduce disease transmission at lower cost.
But, they are loud. And so people turn them off. I’m not aware of a formal study that has measured the noise of an air purifier against its usage, but there’s probably a relationship - based on observing how people in my house interact with the CR box.
If this is true, and if we are trying to maximise total air cleaning, we don’t actually care about how effective air purifiers are in isolation. Rather, we care about how quiet they are too. We’d rather have a weaker air purifier that people leave on for ten hours, rather than a noisy one that they only turn on for an hour. But to maximise this tradeoff, we need the data on:
noise vs effectiveness (measured in CADR) and
noise vs usage.
We would then choose the air purifier that has highest usage times effectiveness. But again, we don’t have noise versus usage data.
In any case, there is no point in a strong air purifier in a classroom of kids if it gets turned off by the teacher. We need to get the air purifiers quiet, so that people use them more. This is my attempt to do that.
The standard DIY Corsi-Rosenthal box uses a Lasko box fan. As of now, it’s pretty loud, even on its lowest setting. I made this previously and analysed it here.
I am trying to get this DIY CR box quieter. So, I built a prototype with PC cooling fans instead. PC cooling fans are an industry where quietness is actually something they try to optimise. This is the prototype:
To do this, I bought some Recticel insulation foam and bought 5 arctic pc fans. I cut the Recticel foam in to a square to match the size of my four MERV 13 filters. Then I cut out 5 square shaped holes with a box cutter and then taped the pc fans inside. I hoped that the foam would be good at damping sound as well. I then stuck the lid on top of my filters.
To power the PC fans, I attached them to a fan hub which was powered by SATA power. And then I connected this to an adaptor that takes mains as input, and outputs SATA.
How does it perform?
Not badly!
I used the same 20 match test that I did with the Lasko CR box to generate PM2.5. According to the data, it’s close (and maybe even a bit better) to a Lasko CR box on the lowest setting without a shroud. But, it’s worse than a CR box on the lowest setting with a shroud. The blue line shows the air cleaning of this prototype, alongside the shrouded and un-shrouded CR boxes. The y-axis are the PM2.5 levels, and the x axis is time. Overall, it’s not bad.
But Is it quieter?
Yes.
Using my Macbook Air built-in microphone, I tested the sound properties of the shrouded Lasko CR box against the PC fan CR box.
I measured the sound 50cm away. On the top middle graph, I’ve plotted the sound levels as a function of frequency. Across the range, especially the mid to high frequencies, we find that the pc fan CR box is quieter. It is quieter overall by about 3 decibels. The other plots show the comparison at different octave levels, and other technical metrics. Its worth noting that there is significantly less high pitch noise.
Also, I’ve noticed Lasko CR box is more tonal - there is a droning brrr at a clear frequency which makes it more jarring to listen to. Imagine listening to an endless low piano note versus white noise. Qualitatively it’s just more comfortable to listen to the PC fan CR box.
This is the sound from my PC fan CR box
And this is the sound of the original Lasko CR box. You can hear the droning and higher volume!
Why are Fans Loud? (Technical)
The literature around fan loudness is scattered. This is my current model based on my imperfect understanding Lighthill’s theory of acoustics (Lighthill, 1952). As well as (Guedel, 2005).
Fan noise comes from
The noise from the blades interacting with the air.
The noise from the motor.
And the mechanical noise from bearings, the shaft, and other mechanical parts.
The noise from the blades interacting with air is the most complicated to model. First order, the noise consists of two parts. First, we have regular ‘fluctuating flows’ as the blades of the fan predictably rotate in the case. Second, we have turbulent flows.
The noise from the blades interacting in the air comes from a ‘droning tone’ called loading noise, as well as white noise called broadband noise.
The droning tone has frequencies associated with it, called ‘blade passing frequencies’. The reason this exists is because fan motion is periodic, and so any interaction is going to cause some harmonic motion.
What dictates the profile loudness? That has to do with how the pressure fluctuations cause sound waves to radiate.
Imagine swirling your finger in water. The act of a fluctuating force creates create waves. Sound is just air waves, and so Lighthill applied this to the movement of solid bodies in air - they’re the same thing except air is a lot less dense!
When the blade rotates, it creates a pressure differential between the front and the back of the fan. These creates a force associated with each blade. The lines of force move as the blade rotates. This creates fluctuating momentum. And just like wiggling your finger in water, this creates waves. The waves look like a dipole radiation pattern below in the case of a basic axial fan.
Does this look familiar? This is the same kind of dipole you would get for a positive and negatively charged particle next to each other, like a hydrogen atom in an electric field.
Lighthill argued that sound waves from such sources could be decomposed into a monopole, dipole and quadrupole source. The loading noise is dipole like in nature.
But the quadrupole noise is the rest. You have broadband noise, which is not tonal and exists across the frequency spectrum. This comes from turbulence. For my understanding, this turbulence mainly comes from the tips of the blades near the casing, as well as something called ‘self noise; which comes from the interaction of the blade’s turbulence with the blade itself.
References
1. Lighthill, M. J. (1952). On Sound Generated Aerodynamically. I. General Theory.
Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences,
211(1107), 564-587. https://doi.org/10.1098/rspa.1952.0060
2. GUEDEL, A. (2005). Aerodynamic Noise of Fans. Contributed Report 01,
Air Infiltration and Ventilation Centre, International Energy Agency.
CETIAT (Centre Technique des Industries Aérauliques et Thermiques).
Current box fans are absolutely horrible, they are built down to a price with no regard whatsoever for CFM or noise. If you can find an old 3 bladed box fan, I am pretty sure they would by quieter than your PC fan setup, and probably more more air to boot.