Sonex cockpit noise

This isn't a "tip or trick", merely a report of sound level measurements I made while trying to reduce the noise level in my Sonex. The cockpit noise has always bugged me. My old Skyfox was noisy, but not so noisy that speech on the intercom was difficult. I have tried 4 different headsets and 5 different microphones, ending up with the most expensive Bose A20 set. They all worked OK for protecting the ears from the noise but none of the microphones have proved capable of cancelling cockpit noise so speech on the intercom is difficult. What really prompted me to try reducing noise was when I found that my radio calls were not being heard clearly. I eventually tracked this problem down to a dodgy press-to-talk switch, but in the mean time I started trying to see what sort of sound deadening really worked. I had read a few accounts by builders who had fitted extra stringers or sound damping panels in the aft fuselage, both of which sounded like good ideas to me, but other than "I think it made a bit of a difference", nobody could tell me how big an improvement the mods made. In all cases they built the plane that way so they didn't have any comparison. I decided to get a little more methodical and bought a sound level meter, a cheapie from Jaycar, and try different things out, one step at a time. I wasn't so worried about the absolute measurements, just whether any changes I made had a useful benefit.
My engine is a Jabiru 3300 with straight-out exhausts, so it's not exactly the quietest set-up, but on the ground it doesn't seem noisy. From over-flying I have had comments of "Wow, that thing sounds like it's on steroids" and "It sounds like a bigger plane" but not "Jeez, that's a noisy little stinker". Here's the results:


dB A at Engine speed, RPM
Comments
Date
Condition
1000
1500
2000
2800
3250
Readings up to 2000 rpm are taken on ground.
21/11/2014
As is - floor mats partly up firewall, foam panel behind luggage bay
84
95
103
102
108

21/11/2014
As is + foam filling luggage bay to seat belt level
84
92
103
104
106

22/11/2014
As above + foam filling complete bay behind seat
84
94
103
104
105

22/11/2014
As above + blanket double layer on glare shield
84
91
102
102
-

3/12/2014
As is plus + sound damping on forward fuselage sides
84
94
100
101.5
103

14/12/2014
As above + damping on glare shield
84
93
101
102

2800 Reading was actually taken at 2850
14/12/2014
As above + replace mats on firewall with larger panels of damping
85
93
101



14/12/2014
As above + foam filling complete bay behind seat
84
92
99
101

2800 Reading was actually taken at 2850
28/12/2014
As above + rubber seal strips along canopy sides



100.5


18/05/2015
As above but with "soundproof" board blanking rear fuselage. Board was made of 5-ply with different types of insulator on each side.
Varied

95
101

2000 reading taken in flight
22/05/2015
As above
82
88
100


2000 rpm reading taken on ground

The final sound damping panel is shown in the photo below. I'm sort of glad it didn't work because it would have been a pain to make a useful working version of this.
sound damping panel.jpg
First, a little explanation of the inconsistencies in results: It wasn't always possible to get a result for 3250 RPM because this engine speed results in an airspeed that is way above the structural cruise speed, and in summer it's rare to get days without turbulence. I tried to hold the meter in the same location but this wasn't always possible. The actual engine speed also varied a bit - most times I was happy to get within 20 or 30 RPM of the indicated number. There is a big time gap between the readings taken in late 2014 and the last two in May 2015. In early February I was about to do "one last" noise test flight when I had a taxying prop strike which resulted in my Sonex being out of action for three months while the engine was bulk stripped and checked. During that time I also made some upgrades on the engine including a later model exhaust and obviously a different prop.

The decibel scale is log scale, so an increase of 3dB is equivalent to a doubling of the sound energy, and conversely a 3dB drop would indicate a halving of the sound energy, but apparently our ears hear things a bit differently. For us to hear a halving of the sound level would require about a 10dB drop.

As you can see from the table, nothing made such a dramatic difference and there is enough inconsistency to imply that some sound deadening actually makes things worse, at least at some engine speeds. The only real improvement was at about 1500 to 2000 RPM and possibly at near full throttle, but at cruise speed, around 2800 RPM, the improvement was negligible.

One area where my aircraft is a little different to other Sonexs is the fit of the canopy to the turtledeck. When the canopy is closed, the rear edge fits hard against a small supporting angle attached to the forward turtledeck former (see photo below). It could be that this is transmitting vibration into the canopy causing extra noise. I'll try to get some noise readings in another Sonex some time to compare with mine.
Turtledeck trim.jpg
A source of noise I didn't investigate beyond trying to stop air leaks (including even having a friend tape up all the canopy joins) was the sound caused by air rushing over the canopy. With power off at high speed there is a significant wind rush sound - try it with your headset off some time. It could be that the tip vortexes from the powered propeller add significantly to the noise. The only real fix for that would be a thicker canopy.

Peter Anson