[Alrojoca]

This product is sold on amazon, they have a thinner version also, maybe if you call them they can send aome free samples to test




Thermo-Tec : Super Sonic Acoustical Mat


Thermo-Tec Automotive Products
P.O. Box 96
Greenwich, Ohio 44837

800-274-8437 Toll free
419-962-4556 Outside the US
419-962-4013 Fax

Office hours are Monday - Friday
8 am - 5 pm EST

 

[TOOSTUBBORN2FAIL]

Alrojoca, thanks for the info, email has been sent, and I'll make a follow up call monday afternoon.



The testing program is done, just waiting until tuesday when my dad is back in town to load it onto my computer. That means this week I'll be testing the repeatability of the test system, then moving forward with tests next weekend. I'm hoping to be able to post the first full scale test by the following monday.



I had an interested experience last night when removing the deadener from the test panel. So everyone knows, the material I used was SDS' CLD sheet, approximately 30% coverage. Obviously, most people have seen his video on his website, and that is an accurate idea of what the metal SOUNDED like before and after deadener was applied. However, I needed to remove the deadener last night, so I tried to pull it off. I ended up getting the aluminum layer off, while most of the butyl stayed put. Guess what, the metal sheet now SOUNDS almost as it did before any deadening was in place. Now, I haven't measured, and that's something I plan to do, but its pretty compelling evidence that the butyl by itself its not doing much, and that constraining layer construction and thickness is one of the most important factors in how this stuff works.

 

[rton20s]

Thanks for the update. I emailed Knu again to see if they had any interest, as I never heard back. Still no response. I might just have to order a sheet of each of their products.

Interesting observation regarding the effects of the butyl thickness vs aluminum thickness.

 

[TOOSTUBBORN2FAIL]

Ya, I hadn't heard anything back either.


The SDS tile was a monster to get off. Had to use channel locks to grip it past a certain point, and the constraining layer came off in pieces. Hit myself in the stomach a few times with the channel locks pulling towards me when the aluminum tore.

 

[Justin Zazzi]

As I'm finishing this quarter's differential equations and materials science classes, your project is becoming more and more fascinating! After studying mass/spring/damping systems and stress/strain behavior of materials for the past few months, I'm able to see with a new clarity what is happening. Maybe I can contribute a little better now.

Damping ratio can also be calculated by comparing the phase shift between the exciter and the substrate, which is what I'm working on now.

Is this still the method you are using to measure damping factor?

This method works because the damping material resists the change in velocity of the substrate, so the system's displacement becomes out of phase with the external force. This difference in phase builds up over many cycles (over some amount of time) to a maximum steady-state value, and if the external force is removed the displacement of the system will also take time to decay down to some insignificant amplitude. If the frequency of the external force is changing "fast" enough, the difference in phase between the system and the force will not have enough time to react and establish a meaningful value. What I mean is, this process does not happen instantaneously and cannot be measured with this assumption. So pink noise is not a useful signal for this measurement, and a "fast" sweep is not either. My guess is a stepped change in frequency or a very "slow" sweep would be best for this method.

But how "slow" is slow enough? If adding the time for one cycle of each frequency from 1hz to 600hz (the upper limit of your accelerometer), it would take about 7 seconds (thank you wolframalpha). So if your sweep was continuous and increased logarithmically in frequency similar to what REW does, you could have around 4-8 cycles near each frequency along the way (or many more if you start the sweep at 10hz instead of 1hz).

All that aside, how are you going to measure the absolute phase of the external force? The two speakers you're using, the enclosure volume both in front and behind the woofers, the airspace coupling them to the substrate, the group delay of the system and who knows what else seems to make that impossible. If you were to replace the external force with something more directly coupled to the substrate like an electromagnet, that might make this method easier to use. But that line of thinking is moving towards the Oberst Beam method that you responded to a few pages back.

I don't know if anyone else finds this fascinating (doubtful), but I suddenly do! I wish you the best man, and kudos for taking on such a great project. I hope your experiment yields some interesting results.

 

[Justin Zazzi]

.....However, I needed to remove the deadener last night, so I tried to pull it off. I ended up getting the aluminum layer off, while most of the butyl stayed put. Guess what, the metal sheet now SOUNDS almost as it did before any deadening was in place. Now, I haven't measured, and that's something I plan to do, but its pretty compelling evidence that the butyl by itself its not doing much, and that constraining layer construction and thickness is one of the most important factors in how this stuff works.

More evidence that constrained-layer-damping tiles are more effective and worth the money compared to home improvement store alternatives

As anuser mentioned a few pages back the constraining layer stiffness, the adhesive stiffness, and the viscoelastic layer thickness are all very relevant to damping performance. You can record the thickness of the aluminum backings as a relative measure of stiffness, and you can record the thickness of the viscoelastic + adhesive layers too as part of your data.

 

[TOOSTUBBORN2FAIL]

Jazzi, that's exactly the idea behind measuring phase of both. And that was my worry, was how to best measure phase of the exciting force. I haven't yet figured out how to work that it, and it probably wont get fully worked out, which is why I'll also have decay plots. But if it can be relative, it may work in the context of a product comparison. As for how to deal with frequencies with enough cycles to let the phase differences to its maximum value, we will be using tones, very close together. Thinking 1/24th octave. Basically, any tone (and amplitude) I want can be entered into Access, and the program will read that and create the tone. Then it will record the results and log them in another Access file, which can then be imported into excel for any graphing and math needed.


I definitely plan to measure and include comparisons of the constraining layer thickness, butyl thickness, and weight. Basically I'll have 1"x1" samples for measurements, controlled heat (oven), realistic heat (car roof through summer), and the 5.5"x5.5" samples for vibration testing.

 

[Justin Zazzi]

I was thinking more about finding the absolute phase of the external force and that path just looks nasty and full of uncertainties.

Instead, leverage the relative measurements like you mentioned. How about calibrating your system by running one (or several) untreated pieces of sheet metal and using these measurements as a control. Then you can subtract the control data from your product data to see the change in phase relative to no treatments. The change in phase between control and products should only be an increase (positive values) corresponding to an increase in damping, or an increase in performance. This works out nicely in my mind's eye anyways.

I hope your family is doing well.
Enjoy what is left of the weekend!

 

[TOOSTUBBORN2FAIL]

That's exactly what I started thinking about, and why I want to run a number of consistency checks as well. As long as consistency between batches is good, then the relative damping will be shown for this system.


Thanks, you too. Weekends are never short enough.

 

[GOYOP]

TOOSTUBBORN,

Have you looked at the paper I referenced from the Navy testing damping properties for ship hulls? If you haven't you will find some valuable references / ideas.

You mention some issues you are having. Keep in mind there is no possible way to eliminate all the imperfections in a test setting. Actually the best way to test this would be to take a stripped out car body and apply the products one at a time. Then shoot a sweep tone at it from underneath and measure the response inside the body with a super flat reference mic.

With that in mind your setup is a good start. You mention decoupling the sound from the speakers in relation to the microphone. You should be able to do this with some minor changes to your test box. You would then create a baseline waveform that you subtract out from all the samples tests so whatever "spillage" from the speakers to the mic are removed from all the test runs.. There are a variety of ways to subtract the common baseline waveform.

I am happy to help build text boxes, provide test tone sweeps, baseline testing, audio advice, accelerometer issues, etc. Let me know what is bogging you down if you want and I will try to help you with a solution. I do it for a living.

Question - have you considered using a transducer for measuring the results? Or as an extra data set? Depending on the device it should have a negligible effect on the results and it would be the same on each sample. The accelerometer is currently being used for this but there may be more suitable transducers for this application. I am not sure but will look into it if you like.

Keep at it.

Greg

 

[TOOSTUBBORN2FAIL]

Thanks for the input Greg.

I have read though most of the file you linked, and to be honest, I'll need a couple more reads before I get everything down. Some of the math is definitely over my head. My dad also has it, but hasn't had a chance to look it over. He's been trying to get permits for his reverse osmosis setup and acid neutralization for his shop, but its hard when local officials have no idea what in the world those things are. It took 2 months to get the furnaces and evaporator ok'd, and he still needs to add a cooling system for the evaporator.

I hadn't actually though about making the baseline run and subtracting it from the results, but I was planning on making baseline runs without the metal in place to see what the differences were, so that wouldn't be hard to do.

I haven't looked past the accelerometer mostly due to cost. When I first started this, I had a little more money to spare, but rent went up where were at, and the job my wife had lined up fell through. I am in the running for a supervisor position which could give more options for testing (due to the raise), but being that I'm by far the youngest of the candidates, it could come down to office politics.

 

[TOOSTUBBORN2FAIL]

Just to try it, I measured response with the mic, without the metal, and with the metal (butyl rubber still on it). I also took a measurement including phase, with and without. For without metal measurements, the clamping ring was still bolted and torqued, just not with the metal.

Without metal

With metal

Without metal with phase

With metal with phase

 

[Hanatsu]

Interesting that there's practically no phase shift in the usable range of the 'with metal' plot...

Sent from my Samsung Galaxy 3 via Tapatalk.

 

[TOOSTUBBORN2FAIL]

I thought so too, but im not 100% confident with the phase measurements with the metal on. It seemed to struggle to measure phase with the metal, but had no problems without it. Im going to repeat the tests today and see what I get.

 

[Justin Zazzi]

What do you think the measurements with no metal will show you?

Wouldn't it be more interesting comparing bare metal to treated metal?

 

[rton20s]

I think he is just testing how the mic will react to strictly the drivers with no additional resonance from the metal. I believe the bare metal will still become the "baseline" measurement.

On a side note, I placed an order from Knu Konceptz today, so we will have some Kno Knoise and Kno Knoise Kolossus Edition to add to the testing.

 

[TOOSTUBBORN2FAIL]

Dustin is right, theres definitely a reason for testing without the metal. I wanted to get a measurement without it in place, one with it, and another set of both after I isolate the back of the speakers from the room. That way it gives me a better idea of the limitations of the mic due to bleed through. Im also still on track to have the accelerometer up and running this week.

Bare metal will absolutely still be the baseline.



Dustin, let me know when it comes in, I'm looking forward to comparing it to the others.

 

[GOYOP]

TooStubborn,

More ideas that you probably have already considered but maybe will help. You seem super thoughtful about your testing which is commendable. However, any rig you build will not be perfect and can always be improved so don't let it keep you from starting to run tests. You may find that some things that seemed critical are not and other things you never considered will be problematic.

Ideas:
1) In any testing calibration and baseline are key. Even if the test system is flawed you will document the flaws and they will apply to every test.

2) If you are able, the best thing you could do is to EQ the mic for a flat response. This is probably the single biggest source of error. What is the peak to peak voltage you are getting out of it with an 80 - 90 db source signal? Do you have the response curve from the manufacturer? Get the mic response flat and then the peaks, etc, even at resonance will not overdrive your mic. If you do experience overdrive back off the gain a bit.

3) You mention phase measurement - is there any reason you can't just use the source signal against the measured signal? I don't even know of any better method because all methods involve a transducer whether it is the actual speaker or something to measure what the speaker is doing. I was always taught to measure the two signals and compare them.

4) Isolating the back of the speakers should be simple. Close them in with MDF and stuff them with insulation. You will not get it to zero but it will be negligible. Out on the highway there will be a variety of entry points for road noise that are much louder.

It's OK to get it reasonably tight then run a batch. It will show you some important things and then you can decide how much you need to improve the test setup.

Greg

 

[GOYOP]

TooStubborn,

I just thought of something that might be an awesome idea. In the lab we always use the best practices to make our tests as accurate and clean and reproducible as possible. I think we are all agreed on that.

But as far as an input sound source, what if you took a super flat microphone, strapped it to the underside of your car, and went out and measured a variety of road noise samples?? You could do concrete freeway, blacktop highway, a bridge, a tunnel, whatever. Tone sweeps, etc, should be done also but we are trying to dampen road noise. And road noise has unique signature. Of course it would be different depending upon the vehicle size and tires, etc. but any sample would be a good test medium.

My 2 cents.

Greg

PS I can make some sample if you like.

 

[mda185]

Here is a free tool that may make this task easier. It will help identify the difference between two audio signals and makes mic calibration a non-issue as long as you are not overdriving the mic or the digital to analog converter inputs.

Audio DiffMaker

I have used this company's flagship product, Praxis audio measurement software for more than a decade. Bill Waslo, founder of LIberty Instruments, was one of the legends of DIY audio 20 years ago and published multiple articles in Speaker Builder and AudioXpress magazines. He had an earlier audio measurement package that was very DIY friendly in the early days of computer based speaker design software. It was cheap and accurate but unfortunately based on DOS and it faded from use as Windows took over the world. Praxis was pretty expensive when I bought the package that included a calibrated mic and USB amplifier and audio interface box. You won't need any of that to use Audio DiffMaker.