The exhaust resonance of my Race exhaust between 2500 and 3000 or so rpm was getting on my nerves when cruising so it was time to do something about it. It's worth noting that the drone appears quieter now that my car's travelled 4000 miles compared with when it was new, presumably as the system gets coated in exhaust particles which appear to have a damping effect.
I've included more detail below than I normally would in case someone fancies doing something similar.
The first step was to identify the problem frequencies and a spectrum analyser is needed to measure sound and display a graph, hopefully so the drone frequency can be identified. The drone should be measured in the cabin rather than outside the car at the exhaust tips, due to resonance potentially affecting the noise levels we hear due to the harsh acoustic environment of the carbon tub surfaces which will bounce sound around the cabin.
I'm not an acoustic engineer and don't have proper recording kit available, so just like in many areas of life, my smartphone was harnessed into use. Reading that the iPhone's microphone clips frequencies below 150hz, I plugged in a better quality mic to bypass this and allow measurement of a wider range of frequencies. To analyse the sound, SpectrumView IOS app was used for its good reputation, cheap price (the basic app is free) and ease of use. The method used was to drive on a level road using a light throttle, and record measurements from 2k5 to 3k rpm in 100rpm steps. This was repeated 20 times. The resulting graphs were enlarged on a 24" screen to effectively increase the resolution and make it easier to identify precisely the drone frequency(ies). These were noted, outliers were ignored, and the mean was seen to be 192hz. The median value was 190hz so correlated closely with the mean, which I've taken to justify reasonable accuracy and repeatability in my readings.
The resultant graph (below) shows a sound level of around 82db when cruising at 75mph. Now, I should add a caveat here that I've used non-professional kit (my phone, a phone app, and a nothing-special mic) to measure this, and the readings may or may not be widely inaccurate. However, to illustrate how loud 82db is, the Health and Safety Executive in the U.K recommend: "The level at which employers must provide hearing protection and hearing protection zones is now 85 decibels (daily or weekly average exposure) and the level at which employers must assess the risk to workers' health and provide them with information and training is now 80 decibels." To put things in perspective, the above HSE warning would only be relevant if you were regularly driving within the drone frequency period for some hours. But it explains why we find the drone annoying, and why I want to remove it.
Photobucket | The safer way to store your photos
The graph shows a noise spike at around 190hz of about 13db. A difference of 10db is a doubling in perceived sound level, so this explains why the drone is so noticeable.
To remove the drone I decided to add a single Helmholtz resonator. Helmholtz resonators have been used to attenuate sound for many years, and with careful calculation and fabrication they can act on quite specific frequencies. On cars, you'll find them on induction tracts (my GTV Cup has 3 as standard) as well as exhausts. They're widely used in loudspeaker enclosures too.
There's a Wiki page if you want to read up about the idea, though I prefer the document here https://www.acoustics.asn.au/conference_proceedings/ICSV14/papers/p406.pdf.
Now that the frequency that causes the drone is known, the resonator volume can be calculated to counteract it. The calculation shown in the above paper is complicated but luckily, there's an app for it: iDuct, it's free, simple to use, and makes a complicated calculation very easy indeed.
As an example of using the app, to remove a frequency of 192hz requires a 1 litre enclosure and a neck of 32mm dia x 30mm long. Using the app you can alter any of the three parameters and see how the frequency changes.
The original plan was to fit a single resonator, connected at the 'Y' piece of the exhaust, something like this:
but this piece at the Y is a pipe joint which I didn't want to disrupt so instead, 2 resonators were fitted - one on each branch of the Alfa Race exhaust. Standard diameter exhaust pipe was used for the cylinder, 75mm for the enclosure and 30mm for he neck. I was disappointed to find that the standard exhaust is fabricated from mild steel. So careful measurement and fabrication left me with two cylinders. Their volume was measured by filling with water, and the volume entered into iDuct to make sure the correct frequency would be targeted. The neck length was cut to the correct length according to the iDuct calculation.
The finished article looks like this. It's difficult to see the brace added at the solid end of the cylinders.
Once fitted and settled in, the spectrum analysis was repeated with the following result:
Notice that the drone spike has been removed by the resonators though the rest of the graph looks a similar shape, frequency and sound level as before. The drone has gone and I can hear the radio whilst cruising. The engine note has been lowered in frequency throughout the range. The pops and crackles are still there, though the farting noise has been removed too. Acceleration through canyons (or valleys and tunnels in my case) produces a much harsher sound than before which sounds like ARGHH!! I'll record a sound file for posting.
I've included more detail below than I normally would in case someone fancies doing something similar.
The first step was to identify the problem frequencies and a spectrum analyser is needed to measure sound and display a graph, hopefully so the drone frequency can be identified. The drone should be measured in the cabin rather than outside the car at the exhaust tips, due to resonance potentially affecting the noise levels we hear due to the harsh acoustic environment of the carbon tub surfaces which will bounce sound around the cabin.
I'm not an acoustic engineer and don't have proper recording kit available, so just like in many areas of life, my smartphone was harnessed into use. Reading that the iPhone's microphone clips frequencies below 150hz, I plugged in a better quality mic to bypass this and allow measurement of a wider range of frequencies. To analyse the sound, SpectrumView IOS app was used for its good reputation, cheap price (the basic app is free) and ease of use. The method used was to drive on a level road using a light throttle, and record measurements from 2k5 to 3k rpm in 100rpm steps. This was repeated 20 times. The resulting graphs were enlarged on a 24" screen to effectively increase the resolution and make it easier to identify precisely the drone frequency(ies). These were noted, outliers were ignored, and the mean was seen to be 192hz. The median value was 190hz so correlated closely with the mean, which I've taken to justify reasonable accuracy and repeatability in my readings.
The resultant graph (below) shows a sound level of around 82db when cruising at 75mph. Now, I should add a caveat here that I've used non-professional kit (my phone, a phone app, and a nothing-special mic) to measure this, and the readings may or may not be widely inaccurate. However, to illustrate how loud 82db is, the Health and Safety Executive in the U.K recommend: "The level at which employers must provide hearing protection and hearing protection zones is now 85 decibels (daily or weekly average exposure) and the level at which employers must assess the risk to workers' health and provide them with information and training is now 80 decibels." To put things in perspective, the above HSE warning would only be relevant if you were regularly driving within the drone frequency period for some hours. But it explains why we find the drone annoying, and why I want to remove it.
Photobucket | The safer way to store your photos
The graph shows a noise spike at around 190hz of about 13db. A difference of 10db is a doubling in perceived sound level, so this explains why the drone is so noticeable.
To remove the drone I decided to add a single Helmholtz resonator. Helmholtz resonators have been used to attenuate sound for many years, and with careful calculation and fabrication they can act on quite specific frequencies. On cars, you'll find them on induction tracts (my GTV Cup has 3 as standard) as well as exhausts. They're widely used in loudspeaker enclosures too.
There's a Wiki page if you want to read up about the idea, though I prefer the document here https://www.acoustics.asn.au/conference_proceedings/ICSV14/papers/p406.pdf.
Now that the frequency that causes the drone is known, the resonator volume can be calculated to counteract it. The calculation shown in the above paper is complicated but luckily, there's an app for it: iDuct, it's free, simple to use, and makes a complicated calculation very easy indeed.
As an example of using the app, to remove a frequency of 192hz requires a 1 litre enclosure and a neck of 32mm dia x 30mm long. Using the app you can alter any of the three parameters and see how the frequency changes.
The original plan was to fit a single resonator, connected at the 'Y' piece of the exhaust, something like this:
but this piece at the Y is a pipe joint which I didn't want to disrupt so instead, 2 resonators were fitted - one on each branch of the Alfa Race exhaust. Standard diameter exhaust pipe was used for the cylinder, 75mm for the enclosure and 30mm for he neck. I was disappointed to find that the standard exhaust is fabricated from mild steel. So careful measurement and fabrication left me with two cylinders. Their volume was measured by filling with water, and the volume entered into iDuct to make sure the correct frequency would be targeted. The neck length was cut to the correct length according to the iDuct calculation.
The finished article looks like this. It's difficult to see the brace added at the solid end of the cylinders.
Once fitted and settled in, the spectrum analysis was repeated with the following result:
Notice that the drone spike has been removed by the resonators though the rest of the graph looks a similar shape, frequency and sound level as before. The drone has gone and I can hear the radio whilst cruising. The engine note has been lowered in frequency throughout the range. The pops and crackles are still there, though the farting noise has been removed too. Acceleration through canyons (or valleys and tunnels in my case) produces a much harsher sound than before which sounds like ARGHH!! I'll record a sound file for posting.