Ear pain/Pressure help

[Yonki]

From what I gather, there are at least 3 known sources:

-cooling system or other component that runs when vehicle is parked (unique to Model Y?)
-resonance from body/chassis (when driving, known issue with several other vehicles)
-hatch buffeting (when driving, known issue with Model S)

I am considering a measurement mic that can pick up infrasound. Measurement Microphones — Earthworks Audio
If the problematic sound is infrasonic, headphones or earplugs wouldn't really be of much use.

I agree - my best guess is it's an interaction of heatpump/cooling system mounting, car resonance, and I perhaps hatch buffeting, but we drove the Y for a bit yesterday with the hatch up, and the noise seemed just as loud. Individually none of them are that bad, but get two or more happening at once...

I'm also trying to figure out how to digitize the noise. The best mics I have need a 48V phantom supply, so I'm thinking about setting up a laptop, my Pro Tools 003 box (from a 12V inverter) and a mic in the car. I'll certainly let everyone know if I pull it off. Service appt is Wed so I have a hard deadline...

 

[DaveORD]

I wonder if an accelerometer on glass would work?

Measuring Shock and Vibration with Accelerometers | Dewesoft

 

[Yonki]

I wonder if an accelerometer on glass would work?
Measuring Shock and Vibration with Accelerometers | Dewesoft

That's a really good idea, might be a more appropriate/useful measurement than audio. I have an accelerometer and a Raspi to talk to it...if I can find it (just moved, garage is still not sorted out). Will try to do that test too.

 

[polyphonic54]

RME makes a laptop powered interface with preamps, but inverter is your best bet at this point in time. Do you have a shock mount for the mic?
I recorded my car (not a Model Y), and it showed a lot of booming noise (30-50Hz) compared to a control vehicle. I clearly heard that booming noise, though not everybody can perceive it even in my vehicle. Standard mics typically have a frequency response between 20-20kHz. Curious to see what you come up with. If you send me some audio I'll run it through a spectral analyzer.

 

[Yonki]

RME makes a laptop powered interface with preamps, but inverter is your best bet at this point in time. Do you have a shock mount for the mic?
I recorded my car (not a Model Y), and it showed a lot of booming noise (30-50Hz) compared to a control vehicle. I clearly heard that booming noise, though not everybody can perceive it even in my vehicle. Standard mics typically have a frequency response between 20-20kHz. Curious to see what you come up with. If you send me some audio I'll run it through a spectral analyzer.

I have a shock mount. Happy to send you the audio once I get it, but I've already got iZotope's RX 7 so I'll be doing some analysis of my own!

Unfortunately I'm missing a Firewire adapter so I won't be able to cobble all this together until it arrives on Monday.

 

[cab]

On my Model S, I just placed my phone on the hatch cover and recorded as I drove around. When I played it back on my pc, what I heard was slight static with the worst pressure pulses (sort of like when a singer puts their mouth right on the mic or you get wind when recording outdoors).

 

[polyphonic54]

Ah, well I was just going to throw it into RX7 as well

I would just suggest driving the same stretch of road, at the same time of day, both with the Y and a control car. I find that night time is a lot better for this.

Gonna do the same if I get the infrasonic measurement mic.

 

[scott_mccool]

you need a pressure sensor capable of reading at 500 Hz or so if you realistically want to see a wave at 250 Hz or lower. I bet most are set to update once per second or less. They do exist. Just not cheap. And then read the signal on an oscilloscope and not a voltmeter.

I was using a barometer that measures at 1 Hz. I'm going to borrow my bosses barometer that measures at 20Hz. I was hoping that would be fast enough but your post makes it sound like that won't work.

 

[scott_mccool]

From what I gather, there are at least 3 known sources:

-cooling system or other component that runs when vehicle is parked (unique to Model Y?)
-resonance from body/chassis (when driving, known issue with several other vehicles)
-hatch buffeting (when driving, known issue with Model S)

I am considering a measurement mic that can pick up infrasound. Measurement Microphones — Earthworks Audio
If the problematic sound is infrasonic, headphones or earplugs wouldn't really be of much use.

I feel it in my Model Y and also felt it in a Model 3 I had as a loaner. Since the Model Y and Model 3 have different shapes and AC systems, I am thinking it has to be the colling system for the batteries.

 

[David.85D]

I was using a barometer that measures at 1 Hz. I'm going to borrow my bosses barometer that measures at 20Hz. I was hoping that would be fast enough but your post makes it sound like that won't work.

You need to sample faster than the frequency you are trying to measure. If you knew for sure the frequency was 15 Hz, then measuring above 30 Hz would do it.

the instrument makers confuse this issue by switching between highest measurable frequency and sampling rate. Mostly for marketing reasons. Look at spec sheet and try to figure out if they are saying measures to 20 Hz or samples at 20 Hz...

An instrument that samples at 20 Hz should be able to resolve pressure frequencies up to 9-10 Hz. This is still useful, but leaves a gap between that and easily heard frequencies - say - 50 Hz plus. And yes, some humans can hear down to around 20 Hz, and you may be one, but your experience will vary.

 

[scott_mccool]

You need to sample faster than the frequency you are trying to measure. If you knew for sure the frequency was 15 Hz, then measuring above 30 Hz would do it.

the instrument makers confuse this issue by switching between highest measurable frequency and sampling rate. Mostly for marketing reasons. Look at spec sheet and try to figure out if they are saying measures to 20 Hz or samples at 20 Hz...

An instrument that samples at 20 Hz should be able to resolve pressure frequencies up to 9-10 Hz. This is still useful, but leaves a gap between that and easily heard frequencies - say - 50 Hz plus. And yes, some humans can hear down to around 20 Hz, and you may be one, but your experience will vary.

Thanks, Makes sense, but I have no idea what Hz I'm trying to measure. I don't know the physics. I just know that my ears hurt after driving the car. I don't hear anything, my ears just hurt. Your previous post suggested around 500hz. Is that what you think it is? Or was that just an example?

 

[cab]

So, I have to ask (as I guy who headed down this rat hole with my S), what’s the end game of the measurement process? More specifically, if you record it...then what? I don’t think Tesla is going to acknowledge it or lemon the cars, so you are likely stuck with trying to do something about it yourself.
The likely choices still seem to come back to:

1. Sound deadener tiles on various sheet metal surfaces.
2. Vibration isolators added to some components (I.e. a/c compressor mounts, suspension mounts or whatever)
3. Vibration dampers added to some structural components like cross members, suspension arms
4. Sound absorbers / blockers added in various places (I.e. perhaps in rear hatch area, or in between glass roof and a glass roof sunscreen or behind door panels, etc.)
5. ???

I’m just to sure the measurements will reveal the magic clue...and then you have to decide how far you are willing to go.

 

[David.85D]

Thanks, Makes sense, but I have no idea what Hz I'm trying to measure. I don't know the physics. I just know that my ears hurt after driving the car. I don't hear anything, my ears just hurt. Your previous post suggested around 500hz. Is that what you think it is? Or was that just an example?

if you are not hearing it, it would have to be 0 to maybe 50 Hz. My guess on the range of the meter is to make sure you have the ability to measure this. The 2x frequency rule is based on the minimum over sampling to calculate a FFT. Other methods have other limits.

Here's something fun to try - if you aren't a vain person... Get your best pair of head phones out and hook them up to a Mac/PC (phones tend to suck) then use this webpage to generate tones to see where the limits of your hearing are. I can sense some pressure at 20Hz-30Hz but not really hear it as a pure tone until 40Hz. On the top end, I can't hear anything above 12 kHz. P.S. If you turn the volume up to hear the low frequencies, turn it waaay down before you crank in the mid frequencies.... You might also find a frequency that hurts the way your car does and you know what frequency you are sensitive to. Take it easy if you think you have a sensitivity...

BTW, I say to use headphones as most of them cover a really broad dynamic range. The speakers attached to my Mac don't pass anything below 60Hz, but I can hear it clearly on the headphones.

Online Tone Generator - generate pure tones of any frequency

 

[David.85D]

So, I have to ask (as I guy who headed down this rat hole with my S), what’s the end game of the measurement process? More specifically, if you record it...then what? I don’t think Tesla is going to acknowledge it or lemon the cars, so you are likely stuck with trying to do something about it yourself.
The likely choices still seem to come back to:

1. Sound deadener tiles on various sheet metal surfaces.
2. Vibration isolators added to some components (I.e. a/c compressor mounts, suspension mounts or whatever)
3. Vibration dampers added to some structural components like cross members, suspension arms
4. Sound absorbers / blockers added in various places (I.e. perhaps in rear hatch area, or in between glass roof and a glass roof sunscreen or behind door panels, etc.)
5. ???

I’m just to sure the measurements will reveal the magic clue...and then you have to decide how far you are willing to go.

This is a valid point - I don't think you will convince Tesla to do anything, but you probably have a number of choices if you were interested in DIY sound deadener / sound isolation if the offending sound is 30 Hz or so, and very few choices if the offending sound is 2 Hz. Even expensive air-cushioned isolators pass a lot below 2 Hz...

The other side of this, is that if they can show, say, there is a loud 20 Hz noise/vibration in their car that isn't in similar other cars, and can show the SC how to measure it, then its something the service center can track down and fix.

 

[scott_mccool]

So, I have to ask (as I guy who headed down this rat hole with my S), what’s the end game of the measurement process? More specifically, if you record it...then what? I don’t think Tesla is going to acknowledge it or lemon the cars, so you are likely stuck with trying to do something about it yourself.
The likely choices still seem to come back to:

1. Sound deadener tiles on various sheet metal surfaces.
2. Vibration isolators added to some components (I.e. a/c compressor mounts, suspension mounts or whatever)
3. Vibration dampers added to some structural components like cross members, suspension arms
4. Sound absorbers / blockers added in various places (I.e. perhaps in rear hatch area, or in between glass roof and a glass roof sunscreen or behind door panels, etc.)
5. ???

I’m just to sure the measurements will reveal the magic clue...and then you have to decide how far you are willing to go.

I want to be able to measure whatever is causing this so I can prove to Tesla that there is a problem. The service center does have the problem that most of the Technicians can't feel it. So they can't fix it because they are pretty much flying blind. It's make an adjustment, givie it back to me, I find they haven't fixed it and I have to wait 6 weeks for another appointment. If there was a way to measure it, then they would have a better chance of fixing it.

And if there was some way of measuring it and proving it, then I would have a better chance of using the Lemon Law. As it is now, Tesla could argue that I am making it up. I'd like to be able to prove it.

 

[scott_mccool]

if you are not hearing it, it would have to be 0 to maybe 50 Hz. My guess on the range of the meter is to make sure you have the ability to measure this. The 2x frequency rule is based on the minimum over sampling to calculate a FFT. Other methods have other limits.

Here's something fun to try - if you aren't a vain person... Get your best pair of head phones out and hook them up to a Mac/PC (phones tend to suck) then use this webpage to generate tones to see where the limits of your hearing are. I can sense some pressure at 20Hz-30Hz but not really hear it as a pure tone until 40Hz. On the top end, I can't hear anything above 12 kHz. P.S. If you turn the volume up to hear the low frequencies, turn it waaay down before you crank in the mid frequencies.... You might also find a frequency that hurts the way your car does and you know what frequency you are sensitive to. Take it easy if you think you have a sensitivity...

BTW, I say to use headphones as most of them cover a really broad dynamic range. The speakers attached to my Mac don't pass anything below 60Hz, but I can hear it clearly on the headphones.

Online Tone Generator - generate pure tones of any frequency

Thanks for your help. How do we know if it is a 10Hz infrasound causing the issue, or some random pressure waves that are too quick to be captured by my 1Hz barometer? I know the problem is made worse then the 4 hatch attachment points aren't all touching. I don't believe that is a harmonic or infrasound or anything, I go over a bump and get a single pressure wave from the hatch "bouncing"

 

[David.85D]

Thanks for your help. How do we know if it is a 10Hz infrasound causing the issue, or some random pressure waves that are too quick to be captured by my 1Hz barometer? I know the problem is made worse then the 4 hatch attachment points aren't all touching. I don't believe that is a harmonic or infrasound or anything, I go over a bump and get a single pressure wave from the hatch "bouncing"

if you had some way to record the pressure data - data recorder or oscilloscope, you could see a spike like that in the amplitude data. Just a half cycle of a wave probably wouldn’t show up on a frequency response curve.

It’s hard to measure... trial and error...

if you really think it’s hatch related, stuff every pillow and blanket you own back there to quiet it and see if it helps. Narrowing it down like that is very useful.

 

[Yonki]

Still waiting for a 1394 adapter before I can take some measurements with a studio-grade microphone, but I did have some fun tonight with my iPhone and this Spectrometer app. It was more useful than I expected.

In the dead of night, inside my Model Y, inside my garage (car was not in motion), I noticed that I can hear a constant, quieter version of this noise/pressure when I set the temp higher (so it's trying to heat the car) and the compressor is running. So I took three measurements:

Compressor on, car trying to heat the cabin:

Compressor off, car trying to heat the cabin:

All environmental controls off:

When the compressor is on, there's a large peak around 30Hz of 37.8dB SPL. When you turn the compressor off, the sound level drops to about -12dB. So the energy at 30Hz increases by about +50dB when the compressor is on.

When you turn everything off, at 30Hz you see about -20 dB. So the compressor on is about +58dB above the noise floor.

+50 to +60dB above ambient at 30Hz is noticeable if it's otherwise reasonably quiet, but you could probably just tune it out if it was constant at that level. My theory is that the cabin and or some other large part of the car (frame or suspension or...) also resonates somewhere around 30Hz, and when you drive over even a small bump or a non-perfectly smooth road it's like you're hitting a ~30Hz bell that's already being stimulated by a weighty compressor vibrating around that frequency. And that would create a lot of variation in the energy around 30Hz, with aperiodic, basically random, spikes of much higher amplitude as the big cabin "bell" interacts with the compressor and suspension movement, and that randomness produces something that sounds a little like and is as irritating as Helmholtz Resonance.

It's not that easy to dampen low-frequency, high-mass resonances. So that may be why Tesla isn't owning up to this - it may be hard or impossible to do much about it. Reducing compressor vibration or changing its frequency would probably be the easiest thing to do. Not sure if re-tuning the entire chassis would be feasible on existing cars (they can and probably are making changes to solve this in a future design change).

One thing I'd like to know is if anyone who is sensitive to this noise/pressure has been in more than one Model Y, and if the problem was the same or if different models resonate less. I hope some did not have the problem - there's more hope for a fix if it's variable between cars.

Would also be interesting to hear if changing/retuning the suspension makes a difference. Maybe that's how we get out of this boominess.

If anyone else wants to experiment along these lines, the iPhone app is only $3 (free + one in-app purchase). The same author has an app that looks at the accelerometer that I played with, but couldn't get much useful information out of. Except that the vibration is front-to-back (as opposed to side-to-side or up and down). But someone else might be able to better use the app.

 

[akballow]

Still waiting for a 1394 adapter before I can take some measurements with a studio-grade microphone, but I did have some fun tonight with my iPhone and this Spectrometer app. It was more useful than I expected.

In the dead of night, inside my Model Y, inside my garage (car was not in motion), I noticed that I can hear a constant, quieter version of this noise/pressure when I set the temp higher (so it's trying to heat the car) and the compressor is running. So I took three measurements:

Compressor on, car trying to heat the cabin:
View attachment 621372

Compressor off, car trying to heat the cabin:
View attachment 621371

All environmental controls off:
View attachment 621373

When the compressor is on, there's a large peak around 30Hz of 37.8dB SPL. When you turn the compressor off, the sound level drops to about -12dB. So the energy at 30Hz increases by about +50dB when the compressor is on.

When you turn everything off, at 30Hz you see about -20 dB. So the compressor on is about +58dB above the noise floor.

+50 to +60dB above ambient at 30Hz is noticeable if it's otherwise reasonably quiet, but you could probably just tune it out if it was constant at that level. My theory is that the cabin and or some other large part of the car (frame or suspension or...) also resonates somewhere around 30Hz, and when you drive over even a small bump or a non-perfectly smooth road it's like you're hitting a ~30Hz bell that's already being stimulated by a weighty compressor vibrating around that frequency. And that would create a lot of variation in the energy around 30Hz, with aperiodic, basically random, spikes of much higher amplitude as the big cabin "bell" interacts with the compressor and suspension movement, and that randomness produces something that sounds a little like and is as irritating as Helmholtz Resonance.

It's not that easy to dampen low-frequency, high-mass resonances. So that may be why Tesla isn't owning up to this - it may be hard or impossible to do much about it. Reducing compressor vibration or changing its frequency would probably be the easiest thing to do. Not sure if re-tuning the entire chassis would be feasible on existing cars (they can and probably are making changes to solve this in a future design change).

One thing I'd like to know is if anyone who is sensitive to this noise/pressure has been in more than one Model Y, and if the problem was the same or if different models resonate less. I hope some did not have the problem - there's more hope for a fix if it's variable between cars.

Would also be interesting to hear if changing/retuning the suspension makes a difference. Maybe that's how we get out of this boominess.

If anyone else wants to experiment along these lines, the iPhone app is only $3 (free + one in-app purchase). The same author has an app that looks at the accelerometer that I played with, but couldn't get much useful information out of. Except that the vibration is front-to-back (as opposed to side-to-side or up and down). But someone else might be able to better use the app.

As scientific as you made all seem. I am not to convinced about dB readings from an iphone.

 

[Ldub22]

Still waiting for a 1394 adapter before I can take some measurements with a studio-grade microphone, but I did have some fun tonight with my iPhone and this Spectrometer app. It was more useful than I expected.

In the dead of night, inside my Model Y, inside my garage (car was not in motion), I noticed that I can hear a constant, quieter version of this noise/pressure when I set the temp higher (so it's trying to heat the car) and the compressor is running. So I took three measurements:

Compressor on, car trying to heat the cabin:
View attachment 621372

Compressor off, car trying to heat the cabin:
View attachment 621371

All environmental controls off:
View attachment 621373

When the compressor is on, there's a large peak around 30Hz of 37.8dB SPL. When you turn the compressor off, the sound level drops to about -12dB. So the energy at 30Hz increases by about +50dB when the compressor is on.

When you turn everything off, at 30Hz you see about -20 dB. So the compressor on is about +58dB above the noise floor.

+50 to +60dB above ambient at 30Hz is noticeable if it's otherwise reasonably quiet, but you could probably just tune it out if it was constant at that level. My theory is that the cabin and or some other large part of the car (frame or suspension or...) also resonates somewhere around 30Hz, and when you drive over even a small bump or a non-perfectly smooth road it's like you're hitting a ~30Hz bell that's already being stimulated by a weighty compressor vibrating around that frequency. And that would create a lot of variation in the energy around 30Hz, with aperiodic, basically random, spikes of much higher amplitude as the big cabin "bell" interacts with the compressor and suspension movement, and that randomness produces something that sounds a little like and is as irritating as Helmholtz Resonance.

It's not that easy to dampen low-frequency, high-mass resonances. So that may be why Tesla isn't owning up to this - it may be hard or impossible to do much about it. Reducing compressor vibration or changing its frequency would probably be the easiest thing to do. Not sure if re-tuning the entire chassis would be feasible on existing cars (they can and probably are making changes to solve this in a future design change).

One thing I'd like to know is if anyone who is sensitive to this noise/pressure has been in more than one Model Y, and if the problem was the same or if different models resonate less. I hope some did not have the problem - there's more hope for a fix if it's variable between cars.

Would also be interesting to hear if changing/retuning the suspension makes a difference. Maybe that's how we get out of this boominess.

If anyone else wants to experiment along these lines, the iPhone app is only $3 (free + one in-app purchase). The same author has an app that looks at the accelerometer that I played with, but couldn't get much useful information out of. Except that the vibration is front-to-back (as opposed to side-to-side or up and down). But someone else might be able to better use the app.

I’m trying to wrap my head around how a car can produce 30 Hz waves. Those waves are 344 feet long! The whole idea of infrasonic waves produced in a space the size of a vehicle is hard to imagine. They are hard enough to produce with advanced speaker components. Detecting them with any consistency would be futile. What seems more likely is pressure gradients. Increased pressure, not in wave form, due to overly sealed environments. If the recirc is on, if the pressure equalization valves are shut, etc. But some sort of infrasonic speaker that only bothers a few select people? In a small space? Hemholtz resonance is more likely, but the force created would have to only bother the hypersensitive.