An idea for decoupling speakers put on a separate stand and avoiding speakers rocking back and forth:

Make a mold in plywood for the speaker, drill 2 holes through the sides (a bit away from where front and back of the speakers will be) and insert plastic tubes before concrete is poured into the mold. Spray some water on top of the concrete, cover with plastic film and let it harden for some days.

Insert wires or 2 rods through the tubes. Add 2 turnbuckles and wires on each side.

Put a shelve on top of the speaker with some rubber sheeting in between. Add 4 threaded eyelets to the shelve. Tighten the threaded turnbuckles / wires so the vibration dampers / decouplers are compressed to correct residual height for a low resonance frequency. Check the correct residual height with caliper, adjust so the speaker doesn’t lean in any direction.

This should create a push and pull system. As the speaker is tensioned to the stand with its high concrete mass, it ”becomes 1 mass”. With is substantial mass compared to the woofers low mass and its back and forth acceleration, any wobbling should be minimized. A potential problem could be resonance in the wires, which may act as tightened guitar strings.

If it is an unusable and stupid idea, I hope Northward would chime in.

Turnbuckle: Turnbuckle - Wikipedia
Eyelet: https://www.forankra.se/zif/pi/1182

The part that JayPee posted a photo of is a prototype / proof of concept of a speaker stand decoupler. We're at V2.0 now and it looks a lot slicker... because I.P. is not yet fully registered with V2.0 I can't show photos yet.

It's a high tension push/pull (dummy loading) so it can be optimized for a certain load range. If your speakers are "only" 25kg, the system is still working with more than 200kg of total pressure for ex. So it's very low in f(n) and stiff enough to get rid of rocking modes in normal use.

Dummy Loading is the equivalent of adding mass to the system.

The systems we use in-wall are very different,more advanced, working on all XYZ axis.

Thanks for your input Adhoc. Will take the time to correctly read and understand it!

I just read about Ps+Sylomer® 1 AMC Preload. It is an interesting product since you can pre-load it to reach a lower f(n).

Still reading and learning! thank you guys!

The pre-load is not about getting a lower f(n), it's about fitting the spring in narrow areas where the final spring compression is known (estimated). So you compress the spring/ cage to where you think it will deflect, install it and once it is under load you remove the pre-loading bolts and it releases the system.

If it's not loaded enough the whole structure will go up at this stage, physically and in f(n). If it's the correct load it will not move.

If you don't release the bolts the spring s will be bypassed by the bolts which are conducting vibrations from top to bottom plate. So no decoupling.

In short, it's not about getting a lower f(n). It's about fitting tight spaces and needing to work on site at post-compression / final loads height. Once released, the spring will behave strictly as it is loaded. No dummy loading here.

Roger that. Makes sense...

Thanks Nortwards. My mental napkin sketch might be one possible way to compress the springs to ”a good range” but some device is needed to keep a static load, once the ”guitar strings” are cut. Not impossible for DIY but for a commercial item it should be elegant, versatile within certain ranges plus fool proof.

...I'm reading more about damping ratio, how to calculate it.

I'm reading the Sylomer's PDF Thomas mention in this post (about form factor bit and the elasticity modulus etc).

Pffiouuu so much parameters must be taken into account.

I'm afraid it's too much for me. At leat I read and read and read for long time. But...I don't have that time.

I will probably go for an easy but not perfect solution though

OK I read even more. Learnt a lot. Being lost a lot.

I read about Sylomer properties on AMC tech data as Thomas suggested. Form factor, elasticity modulus etc. Ok but,...it's hard to digest. I was lost.

So I also look at Spring + sylomer products, and check the tech. data (from AMC mainly).

Only f(o), deflection and loads are mentioned.

No static load. No dynamic load. No damping ratio.


"Only" this: respectively NATURAL FREQUENCY CURVES and LOAD DEFLECTION CURVES.



Is it because Sylomer in this product is only used for upper frequencies, so it doesn't have to be calculated precisely? (That's my thought but I probably wrong).

Let's say the mass supported by each spring is 12,5kg.
I will get a deflection of 15mm if I use AMC 15 ref.
Which leads to a f(o) of (about) 4,2Hz according to this tech. data (see above).

Am I wrong so far?

So if it's ok...will it work? :

-Speaker 15kg
-Static load (to add mass. Mass to be determined/calculated so the 3 or 4 springs will be evenly loaded for a mass 12,5 kg on each spring))
-Springs + Sylomer
-Support/Stand (heavy, rigid, non resonant)

The isolation should start at 5.93 Hz=√2x4.2

What should/could be improved/calculated?

Sorry for these newbie questions.

Maybe you are confusing isolation with damping? Two different concepts, but related. The isolation is due to the spring being "springy", and is basically related to just one other factor: gravity. (Unless you also apply pressure mechanically). The weight of the speaker is due solely to it's mass times gravity, and the static deflection of the spring is due solely to the stiffness of the spring and the weight of the load (once again, baring mechanically applied pressure).

It is ONLY the static deflection of the spring that sets the resonant frequency of the system.

Look at the math, and it becomes very clear:

The equation for the natural resonant frequency of single degree of freedom system with static deflection is simple:

F = 1/2 PI x SQRT (k/m)

where:
k is the stiffness of the spring, and
m is the mass that is resting on the spring, causing static deflection

In turn, the static deflection equation is also simple:

kD = mG

Where:
k is the stiffness
D is the static deflection
G is gravity
m is mass

If you don't know what the stiffness of your spring is, you can flip the equation around to find out:

k = mG/d

(in words; the stiffness of the spring is equal to the mass multiplied by gravity, divided by the deflection)

Substitute all of the above into the first equation, and you get:

F= 1/2 PI SQRT (G/D)

The same as I said above: The resonant frequency depends ONLY on gravity (G) and the stiffness of the spring, and in turn the stiffness has a direct relationship to the static deflection (D), which in turn depends on the weight of the object.

So, in simple terms: since gravity is constant, and PI is a constant, the ONLY thing you need to do to find the resonant frequency, is look at the static deflection.

And that's also dead easy to do! Get a sample of the "spring" you plan to use, put your speaker on top, and see how much it "squashes down"! That will tell you what the resonant frequency will be. So if your sample is 25mm thick, and when you put your speaker on top it "squishes down" to just 20mm thick, then your static deflection is 5mm. Thus you can predict the resonant frequency. You want that to be not more than half the lowest frequency that the speaker produces, and ideally not more than third.

OK, that's the basis, but it gets more complex: obviously if you over-compress the material, it wont be as springy any more: and by the same token, if you don't compress it very much at all, then it won't be very springy. so you need to check the linearity of the material, and find out what range of compression (deflection) it will work for (remain springy). For most materials, you should be OK in the range 15% to 25% compression, but that's just a general guidelines, not written in stone: check with the manufacturer.

So, you need to find a material that will allow you to get the right amount of static deflection for the frequency you want, but where that amount of deflection is within the linear and useful range for the material.

For most materials, you'll find that you need a very thick "spring". Thin ones don't compress enough, or need to be loaded with a LOT of additional mass, to get the necessary static deflection (when gravity alone isn't enough).

OK, so all of the above is about the spring and the frequency and the deflection and the isolation: but damping is an entirely different matter. A spring by itself is "springy"! (duh....) It bounces! But you don't want your speaker to be bouncing around on the pads, so you have to stop the "bounce" without affecting the springiness. Enter damping. Damping absorbs some of the energy in the "bounce".

Think of it this way: if you have a brick hanging from a tough elastic band, and you pull the brick down then let go, it will bounce up and down on the spring for a long time, at the frequency given by the static deflection. Now take that same system and put it at the bottom of a swimming pool: if you pull the brick down now and let it go, it will rise to where it was before, but it won't bounce up and down much, if any: it just returns to the static deflection position and stops. Because the water "damps" the motion (in addition to making the brick rather "damp" too! It's unfortunate that "damp" has two unrelated meanings... I need a better analogy... ).

That's where you seem to be right now: worried about the damping.

So most people think: "Cool! I'll just use a very springy material that also damps the hell out of things!". Bad idea. Because there's no free lunch. It turns out as you increase the damping on a system, you also increase the transmissibility of the system above resonance... that's just a fancy way of saying that it doesn't isolate so well for frequencies above the resonant frequency. The more you damp it, the less it isolates.

To understand this, it's first necessary to go back to what I said before about needing to get your resonant frequency down below half (and preferably below one third) of the lowest significant frequency put out by the speaker. The reason for that "factor of two, or three" advise is simple. At the resonant frequency, a system does not isolate at all, and in fact, it amplifies (if it didn't, musical instruments wouldn't work very well!). That resonant system continues to resonate at frequencies above and below it's own resonance. And for convoluted mathematical reasons, it refuses to isolate until exactly 1.414 times the resonant frequency. Above that, it isolates. Below that, it amplifies. It does not matter how springy, or what spring, or what damping, or what the material is, or how much you load things, or what the temperature is, or anything else: the factor is ALWAYS going to be 1.414.

Why 1.414? Because that's the square root of 2, and if you go back to the equations above, you'll see that all of this springy stuff is calculated with a big "square root" sign in there. I'm not going to bore you with the derivation of that "Square root of 2" thing, but it's easy to remember that your resonant system will amplify below 1.4 times the resonant frequency, and will isolate above that frequency.

Of course, it's not that there's a sudden cut-off at 1.4, with everything below amplifying terribly, and everything above isolating excellently! Rather, there's a gradual curve that crosses over from "amplify" to "isolate" at exactly 1.414 times the resonant frequency.

So far so good. Now for the interesting part: the SLOPE of that curve is defined by the damping. If there is no damping, it is very steep on both sides of the 1.414 point: So you get high resonance, high amplification below that, and high isolation above it. With lots of damping, the amplification is much lower.... but so is the isolation!

And that's the problem. If you use a material that has very high damping, then sure, it won't resonate very well below the 1.414 point, but it also won't isolate very well above the 1.414 point.

The graphs below illustrate that clearly. Below the horizontal line marked as "transmissibilty = 1", you get isolation. Above that line, you get amplification. Exactly on that line, you get neither: sounds just travels through exactly as it was, without being either amplified or isolated. Obviously, what you want is to make sure that the peak of the resonance above the line is as low as possible (so that it amplifies very little, even at resonance), and also you want to make sure that all of the frequencies your speaker produces are as far below the "less than 1" region as you can get them.... which means you want a system that moves the "1.414" point as far over to the left as possible. If you have a highly damped system, that curve below the line rises up towards 1, and flattens out... so you don't get good isolation until maybe 3 or even 4 times the resonant frequency. In an undamped system, that curve drops steeply and becomes more vertical, so you get good isolation even at just twice the resonant frequency... but in that case, the resonance is very strong, which you DON'T want.

In other words: Murphy will get you either way!

With low damping, you don't need to tune resonance so low, but then you have very high amplitude at resonance (high "bounce" in your spring)... and with high damping, you get very low amplitude (not much "bounce"), but poor isolation.

The only solution here, is to tune your resonant frequency very low if you want to use a highly damped material. The higher the damping, the lower you need to tune it.

Summary: tune your system using static deflection. That's what determines the resonant frequency. You need a thick spring to get low resonant frequency. You need damping in your system, to reduce the amplitude of the resonance, and stop the "bounce".... but having high damping means poorer isolation.... so you need to tune lower than you thought.

I'm not sure if that helped you more, or if I just managed to confuse you more! Hopefully, the former.


- Stuart -

Many thanks. It's a lot of infos to read/digest and bring together.

Will print your message and read it carefully.

Cheers Stuart!

I'm off for couple of days my brain just explosed

What Soundman 2020 wrote can be seen very evidently in the diagram here: https://www.elesa-ganter.com/static/..._choice_EN.pdf

The diagram is for vibration isolators made of solid rubber but the principles are the same. -You'll need certain load and deflection for the vibration damper to work as wanted. Too much load and it bottoms out and transmits the vibrations, too low a load and its resonance goes up in frequency.

Circular or rectangular rubber pucks work too but check out how much deflection which is needed to reach say 90% / 20 dB isolation at 20 Hz, about 6 mm or 1/4"(!). They are totally useless in other words for light weight speakers or gear. If the "device" is made conical or as a thin walled rubber ball with a flange, then some types may be useful for reaching a low resonance frequency. You'll see in enclosed catalogue that only the conical DVA 6 in soft 40 Shore rubber would be of any use for light weight speakers, all others shapes and dimensions require a substantial load to be of any use as vibration dampers versus low frequencies.

Sorry about that! I sure hope your medical insurance covers exploded brains! I'll try to be less explosive in the future...

But seriously, once you get your head around the basics, it starts to make more sense. As Adhoc pointed out, you need ridiculously soft and very deep springs for ordinary light-weight speakers.... or you need to add a LOT of mass ... or you need to apply some extra "gravity", with mechanical pressure. Or you need specially shaped pads, such as hemispheres or cones... But that makes it even harder to calculate, because then the reaction isn't linear: the more you load it, the more area takes the load... so the form factor changes with loading, and thus the frequency and damping too... But they work...

My favorite material is still Sorbothane. They sell hemispheres for different speaker weights, that do work quite well.

Edited to add: I just went back and re-read what I wrote... Ummmm... I PROMISE it made a lot more sense while I was writing it! Really! I thought it was very clear and simple back then! But now.... Hmmmm....

- Stuart -

That is a pretty awesome explination my friend. Thank you. You certainly earned your gold star today.

Ahah no Stuart, it's just me. So many infos to digest. Need more time hehe. You guys are fantastic (Thomas, Stuart, Jason, Adhoc, Jens!)

Ha! looks like the subject is pretty well covered by now. Thanks Stuart for the post.

It's actually not a complicated subject at all, just one that is misunderstood and often surrounded by pseudo-science from manufacturers selling you snake-oil "decoupling speaker interfaces". Which creates most of the confusion to start with.

The science is over a 150 years old.

The only issue you need to solve with speaker decoupling is the static load vs dynamic load (=rocking modes).

I'm also prone to recommending spending your money on more important things first, like proper LF treatment that will bring a much higher ROI. Speaker decoupling is one of those aspects of a design you zoom into when all the rest is taken care of and solid.

Out of curiosity, what are all of the ill effects of improperly decoupling speakers, and how would they show up in room measurements/how would you measure their effects?

Lets say anywhere from a high natural frequency in a rigid spring, to really squishy sorbothane, to something like auralex mopads. How would these manifest themselves?

Some more on vibration isolators, a floating floor and other stuff to ponder if one wants to DIY an anechoic room. A Master Degree Project: "Design of a Fully Anechoic Chamber". Quite easy and plain language, to get a grip of what matters.
http://www.diva-portal.org/smash/get...FULLTEXT01.pdf

(I'll pass though. My wallet is too thin for what is needed; 18 metric tons of insulation and 368 metric tons of concrete.)

Just to update you guys... I think the system is not worth it (to my room!).

Money and time spent on it are not worth it.

But that would look pretty cool hehe.

I learned a lot, hope this thread will help others!

Many thanks to the community as always

Hi @ all,

If been lurking the forum for quite some years now but never posted anything.
Thanks for all the professional insight that's been given me over the time but so here we go

After I read Stuarts post several times I understand that the key to decoupling the Speakers from the stands you have to get the resonant frequency of the springs as low as possible (5-6hz) because dampening with sylomer in the end flattens the Isolation curve towards the hearable spectrum.
This can either be done by mechanically pulling the top part with a spring as integrated in the northward system (?) , which is not feasible for me and my engineering capabilities, or add additional weight to the system.

In my given case I'm trying to help a friend of mine successfully integrate his ATC scm110 into a soft-soffit and am now looking for a good speaker stand/decoupling solution. My idea was to build a case for the Speakers with a compartment where I can fit in heavy lime stones like these below:
[quote]KS Plansteine
This way I think its possible to "tune" the weight of the whole top part (ATC's 73kg + Stones(max 85KG) + MDF ) by observing the springs static deflection to achieve the desired low resonant frequency.

Is it true that a positive side effect of basically attaching the speakers physically to the added weight is that the speaker will move less in z-axis when the bass woofers move a lot of air at high SPL? resulting in less parasitic oscillation /distortion? As far as I know Thomas from Northward uses Springs not only on y-axis but also other axis to get rid of that problem but does it have an effect although not nearly as slick and professional looking?
Hope you can follow me, as my technical English may not be precise,

All the best, Daniel


I added some sketches drawings of the idea:

As you understand this could you please provide me with a real world calculation of this?
He said something about a 5 mm deflection, how do I calculate the resonance frequency?
I ask this for a friend.

Its stuarts post, ask him.

Not trying to argue against, or be the gotcha guy at all with this... I'm just wondering if I'm gathering correctly that when talking about freestanding monitor deployment, the scenario is drastically different than when set on the table, to the point where these products are really a different animal entirely.

https://ethanwiner.com/speaker_isolation.htm

Haha, I just wondered if anybody understands the really easy, 150 year old physics. And Stuart seems to have one of his sabbaticals.

I could do it if i really wanted to, but i prefer a handy dandy calculator jens showed me a few years ago.

No, the physics is the same. You want the speaker decoupled and stabilized on all axis, but having the speakers on the desk isnt ideal in the first place

I would first take the motto of The Hitchhiker's Guide to the Galaxys serious: Don't Panic!. Than see what Ethan has to offer and when you are totally sure there is something wrong with your speaker placement grasp the decouplers. And don't forget to read on speaker membrane doppler distortion!

It seems that if it is human nature to anticipate (i.e. trying to solve problems that haven't popped up yet), I'll be 20% more inclined to do so than most people. LOL

Lol, theres PLENTY of other problems to solve before you worry too much about monitor decoupling. I have my monitors squished into boxes with sorbothane bumpers to "decouple" them. Are they truly decoupled?? Probably not. Do i care? No... Would i hear a difference if they were? Highly unlikely