Microphony

[BE718]

Well, not if the claims that microphony is a non-problem are true, obviously...

Anyway, I thought the point here was that we weren't simply measuring the vibration, we were going to be measuring the output of the equipment when subject to, and not subject to, the vibrations from a musical signal. I realise that has been delayed due to non-availability of kit, but isn't it therefore premature to pronounce any vibration currently measured as being tiny. It may be tiny, but we don't yet know if it is insignificant.

There is logic to your statements here of course, but as I mentioned you need to understand how small this vibration is.

As I mentioned further up the thread I have used very sensitive amplification systems in jet engine test cells where the acoustic pressure levels were over 120 dB without problem. Another poster was talking about the qualification testing he has to do on military audio equipment vibrating kit with shakers without any problems becoming apparent.

So am I expecting to see anything in the amp noise floor? Nope.

 

[Hipper]

I can go back and do the measurements again if you really want, but I will tell you what the result will be - rigidly coupling does not isolate vibration, there is no mechanism to do so.

Just think about it, its pretty much common sense.

OK, something you might relate to. If you replaced the tyres on your car with solid metal ones, do you think you would feel less or more vibration in the car?

Perhaps I misunderstand your graphs.

I understood you were making an 85dB 120Hz noise to test the vibration response. You get a 120Hz response of less dB plus a harmonic at 240Hz.

My question was, what is the other recorded responses at 100Hz and below? For example, on the first graph you get some small peaks at roughly 55, 75 and 90Hz. Is it for example noise inherent in your measuring device, or, more interestingly, is it external background noise?

That leads to the question, 'can you measure background vibration on your floor for example?'

 

[BE718]

Thanks for doing those measurement, very interesting . I don't suppose you could try redoing them but measuring the difference in baffle vibration with the speaker mounted on spikes and on the Focal pods? Perhaps a range of frequencies (or/and the centre of the side panel of the cabinet). Sorry, I realise that's a big ask (I agree with your conclusions though).

Yep, I fully realise the measurements were very limited and I was waiting for someone to come along and ask for just what you have The measurements were just to demonstrate the specific point about rigid coupling.

To do that effectively I need one of our Bruel and Kjaer systems and sweep across the range. Thats going to have to wait for another day I'm afraid.

 

[BE718]

Perhaps I misunderstand your graphs.

I understood you were making an 85dB 120Hz noise to test the vibration response. You get a 120Hz response of less dB plus a harmonic at 240Hz.

My question was, what is the other recorded responses at 100Hz and below? For example, on the first graph you get some small peaks at roughly 55, 75 and 90Hz. Is it for example noise inherent in your measuring device, or, more interestingly, is it external background noise?

That leads to the question, 'can you measure background vibration on your floor for example?'

I was playing a sine wave at 120Hz at 85 dB(A) at 1m and measuring the vibration in velocity (mm/s).

I did not play any stimulus at lower frequencies. If you really want I can do so, but as I said the rigid coupling issue will remain.

Those other peaks are mainly from the speaker itself i.e not measurement or background noise. I think the driver may need tightening

A note about accelerometers. the one used here is a low output variant, 10mV/G which are inherently more noisy at low frequencies. I do have a larger 100mV/G low frequency one I can use to measure again.


Again, do you realise just how low those other components were? We are talking 0.003 mm/s just how can I put this into context for people?

OK, I will take some measurement of the floor

 

[Hipper]

Again, do you realise just how low those other components were? We are talking 0.003 mm/s just how can I put this into context for people?

Of course this is the problem as I'm not an engineer of any sort.

The highest peak was 0.0045 mm/s or so. I presume in laymans terms that means the censor moved back and forward at that rate. Or it's largest oscillation was 0.0045mm, or 45 microns.

The distance between tracks on a CD track is 1.6 microns.

 

[Sue Pertwee-Tyr]

Again, do you realise just how low those other components were? We are talking 0.003 mm/s just how can I put this into context for people?

One way might be to show the velocity of the driver itself, then we will have an idea of just how attenuated the cabinet vibration is. I know when my speakers are working, while I can feel the vibration of the driver, I can't see it, and under strobe lighting the movement appears to be on the order of 1mm or so, so if the driver is moving 1mm, and the cabinet is moving 0.005mm, that is 2%. That wouldn't be trivial, if expressed in dB, would it?

 

[BE718]

Of course this is the problem as I'm not an engineer of any sort.

The highest peak was 0.0045 mm/s or so. I presume in laymans terms that means the censor moved back and forward at that rate. Or it's largest oscillation was 0.0045mm, or 45 microns.

The distance between tracks on a CD track is 1.6 microns.

Vibration is a little tricky to get you head around at first, we measure it in acceleration (G), velocity (mm/s) or displacement. They are all intrinsically related

So yes the sensor moved at that rate, but that is not the physical displacement. You would need integrate the velocity to get displacement.

Acceleration is essentially a flat frequency response if you want to think about it in those terms, velocity integration is like a 6 dB/octave filter so double in frequency you half amplitude. As you double integrate to get displacement you hald it again with doubling of frequency. So they both dramatically reduce with increasing frequency

We very rarely use displacement though.

Heres is the plot.. Its so low I questioned the calibration, but it a appears to be correct. I have also been using a 100mV/G transducer recently so worst case if I screwed up it would be x10 out. Even then still so low its not worth thinking about.

 

[Sue Pertwee-Tyr]

Ah, realised my error too.

 

[Michael J]

Again, do you realise just how low those other components were? We are talking 0.003 mm/s just how can I put this into context for people?

After five and a half minutes the measured surface will have moved almost 1mm in total, half one way and half the other. Best not move our ears while we listen to the music...


Using speed of sound 340 m/s and a frequency of 17KHz for easy maths and because golden ears, one "audio cycle" of such has wavelength 0.02 m = 2 cm! Just for dimensional context.

 

[clivem2]

Given 3 scenarios for speakers:
- no spikes or pods
- spikes
- pods / damping

it seems reasonable they will all behave differently in terms of vibration / movement of the cabinet. Experience suggests there's no single right way to couple or damp a speaker. Part of the picture must be the interaction of the driver with the cabinet. A cabinet which moves fractionally will look different to a driver and may sound different to said driver being operated with a totally rigid mounting. Might the leading edge of notes be slightly altered if the cabinet moves as a result of strong driver excursions?

 

[YNWOAN]

Pure theory dictates that the cabinet must be moving very slightly and it is clearly vibrating - which is itself movement. However, that movement is very tiny indeed. Instead, the amplitude of some cabinet vibration frequencies can be pretty pronounced and as the drive units are coupled to the cabinet that means that the vibration is also coupled to the drive unit and is likely, at least in part, to be added to the signal they replicate (but at a delayed and distorted rate).

There is a bit of a conundrum here in that the drive unit vibrates (to produce sound) and that vibration goes into the cabinet where it is modulated by the cabinets material and construction method - ideally you don't want that filtered vibration to find itself back into the drive units. Close coupling the speaker to the floor (via spikes) creates a mechanical earth for that vibration but, inevitably, it means that the floor is now energised. Supporting the speaker on a compliant material means that the cabinet vibration will dissipate more slowly but not via the floor - it also means that the drive units will 'see' that vibration differently than if the cabinet were rigidly coupled.

 

[BE718]

Given 3 scenarios for speakers:
- no spikes or pods
- spikes
- pods / damping

it seems reasonable they will all behave differently in terms of vibration / movement of the cabinet. Experience suggests there's no single right way to couple or damp a speaker. Part of the picture must be the interaction of the driver with the cabinet. A cabinet which moves fractionally will look different to a driver and may sound different to said driver being operated with a totally rigid mounting. Might the leading edge of notes be slightly altered if the cabinet moves as a result strong driver excursions?

In principle a speaker should be rigid as far as Im concerned. You dont want it moving back and forth , however having said that what force is a driver exerting? Not an awful lot A speaker cabinet may well have resonances at certain frequencies which can potentially colour the sound.

 

[BE718]

just to try and get some context for people

Kettle nearly boiling - the waveform is probably fmore representative due to the dynamic nature of the signal. The FFT wont follow that.

Nespresso coffee machine......... bbbbbbrrrrrrrrr!

 

[clivem2]

In principle a speaker should be rigid as far as Im concerned. You dont want it moving back and forth , however having said that what force is a driver exerting? Not an awful lot A speaker cabinet may well have resonances at certain frequencies which can potentially colour the sound.

There is something at play though as different coupling often results in different sounds. YN's post above presumably is very relevant. Yes the movements are small but these small movements are important, let's face it most cones don't move by much but they don't all sound the same, just because a movement is small doesn't necessarily mean it's not important.

 

[BE718]

There is something at play though as different coupling often results in different sounds. YN's post above presumably is very relevant.

I suppose this is getting a bit off topic as it is meant to be about microphony.

The only thing playing for Stevens stand is damping, the plastic will behave differently to metal. However music is usually a continuous vibration input and its still very rigid in a vertical plane. As I mentioned previously the resonant frequencies will also be different but that would have to be quantified and may not be a good thing.

 

[radamel]

However music is usually a continuous vibration input.

Why do you say that?

 

[BE718]

Why do you say that?

I was generalising, please dont get pedantic and say particular music x stops and starts with silence in between......

 

[radamel]

I was generalising, please dont get pedantic and say particular music x stops and starts with silence in between......

Clearly a bad generalization IMO then.

 

[BE718]

Clearly a bad generalization IMO then.

Not at all. How much of the music you listen too is of a continuous nature?

 

[YNWOAN]

The only thing playing for Stevens stand is damping, the plastic will behave differently to metal. However music is usually a continuous vibration input. As I mentioned previously the resonant frequencies will also be different but that would have to be quantified and may not be a good thing.

In terms of material damping then yes to a point, but acrylic is no more a broad band isolator than MDF (for example). The way the stand is structured means that it's not terribly rigid and therefore lossy - but again this wil be centred around a narrow frequency range.