DINAS big brother(s...
 
Notifications
Clear all

DINAS big brother(s)

Page 3 / 4

ajc9988
(@ajc9988)
Estimable Member Registered
Joined: 6 months ago
Posts: 126
Topic starter  

@chedwin - Thank you for the input. It is appreciated. Nonetheless, and unfortunately, there are few resources that have taken the insulation, placed it within a cabinet with the rest of the system being controlled, and systematically testing the ability to reduce resonances or standing waves within the enclosure. You can see some on the material's effect in a closed enclosure in the book Loudspeaker Design Cookbook, but there are many more types of insulation out there and very little on testing each type (whereas many do not want to put their whole lineup under testing because it undercuts their marketability of their tested solutions, in part).

If I am understanding you correctly, then, you would be more ready to recommend, for a subwoofer enclosure part of a cabinet, constrained layer damping (placing a viscoelastic polymer or sheet between two solid materials, such as plywood>viscoelastic material>plywood) which would likely help to reduce box resonances due to the nature of converting the mechanical energy into heat for dissipation. That was something I toyed with the idea of doing in addition to insulation, but decided against for this project.

I also, in considering the framing for my box, reviewed this from audioholics: https://www.audioholics.com/loudspeaker-design/detailed-look-proper-loudspeaker-cabinet-bracing/finite-element-analysis-part-i

This used LEAP 4 or 5 to model the placement of braces along a plane and the effect on stiffness and displacement of the plane. It's an interesting read.

Other than that, still a noob to speaker design (although I have progressed a lot since I started). So please take the above as speaking from ignorance and feel free to correct me as needed.

Now, the way in which I used that reference for the insulation is just to estimate possible efficiency to address standing waves in the enclosure. Although I did roughly look at it on the sub part of the cabinet, it also informed which materials would likely deal with the reflections inside the upper compartment with the woofer (I decided on cost, likely just getting Roxul 60 rigid board, which is cheaper than the denim acoustic batts for what I have found so far, mainly at ATS Acoustics, but if you know a better/cheaper source, I'm all ears). I mixed that information with the information found on this cheat sheet of sorts (table of frequencies and wavelengths): https://www.jdbsound.com/art/frequency%20wave%20length%20chart%202013.pdf

That is more to give an idea of, where a standing wave appears, which frequency to look at and what to try to trap for the frequency (and how the box design itself allows for a wave of a size to exist and perpetuate itself). But, once again, a noob.

But once again, thank you for the input. I do understand those values are not perfectly translatable for use in this case, but it does help ballpark for when it makes sense to use what (such as knowing which insulation is even capable of reducing well a specific frequency found in a standing wave).


Chedwin liked
ReplyQuote
Chedwin
(@chedwin)
Eminent Member Registered
Joined: 4 months ago
Posts: 22
 

@ajc9988 i should have worded my previous post better to make clearer that when im talking about the data not translating to speaker cabinets i meant for the overall cabinet/main low frequency chamber or a standalone subwoofer box not smaller internal chambers for other drivers. what you have done and they way you have referenced the data to find the best material for mid and high chambers within the main cabinet is absolutely fine. appologies for the oversight in detail on my part there

 

in terms of the sub chamber you are in fact correct that i would be more likely to reccommend constrained layer damping for a sub chamber or a standalone subwoofer. from a purely scientific point of view contrained layer damping would be much more effective than single layer ply and insulation and in something like a 2 x10" standalone sub or bigger i would strongly reccomend considering it. however the practicality of it has to be determined on a case by case basis for cost, time, effort and other measures already taken.

if you have already done FEA calculations for your bracing placement then it sounds like you are already at a good place for the sub chamber, the bracing alone should be plenty to get the best sound out of the speaker. in this instance i would say constrain layer damping isnt needed on top of current measures taken

 

hope that clears some things up a bit. cant wait to see how these turn out 😀 

Josh Evans
- Professional Live Sound Engineer
- High End Commercial AV Install Technician


ajc9988 liked
ReplyQuote

ajc9988
(@ajc9988)
Estimable Member Registered
Joined: 6 months ago
Posts: 126
Topic starter  

@chedwin - No apologies necessary! Although I found those resources for generally designing the mid-high chamber, I was also trying to figure out whether it could be used efficiently for the sub chamber. But after reviewing the materials, the designs, etc., it just didn't make sense (unless trying to greatly explode the size of the cabinet in order to fit the volume of thicker insulation needed). Part of this exploration was prodded on by some applying insulation to the design known as The Hammer.

image

But considering the box was a kit and others building the kit did not use insulation, there is the chance the design was never meant to have insulation and the dimensions not adjusted to account for and accommodate the insulation.

As discussed though, the materials just are not really able to handle low frequencies very well without going insanely thick, etc. And, as we both agree, constrained layer damping would likely do much better to control box resonance than insulation in any event.

I have not yet done FEA calculations for the bracing placement, but do plan on testing my design. The inclusion or exclusion of insulation for the sub chamber was the primary last factor before finalizing the design. I then would like to move that design over to an FEA software that can import the design from sketchup. From there, it is using the plywood's values in order to roughly test resonances, bracing locations, etc. I was inspired to do FEA from that article, but from the beginning used the findings in that article in order to determine what dimensions should be used for the bracing (like thickness double that of the wall materials, the length away from the wall impacting the influence on the modal frequencies, etc.). It also made me think about what can count and act like a brace within the structure, such as a full shelf to separate the upper chamber, which carries both the chamber for the plate amp and the chamber for the woofer and tweeter (the tweeter is a sealed design, so I didn't need to design a third chamber to support the tweeter).

Now, to add constrained layer damping (hereinafter "CLD"), all it would take is adding a line to my spreadsheet with the thickness of the layer, then adding a second layer of plywood thickness to the outside dimensions of the cabinet, widen the front baffle (trying to keep the product to 4' high and 2' deep, so widening the bottom outside from about 17" to accommodate the extra area would be doable; in fact, then you could design the bottom sub cabinet to slide into the overall outer frame for the speaker cabinet, allowing for further compartmentalizing in construction and testing) in order to come up with the design for that. I have created a worksheet to automatically adjust the dimensions by updating values. So I might create two designs, one with CLD and one without. Then just run the FEA on the two designs.

Either way, you are probably correct, especially since, due to my altitude effecting air pressure effecting the linear excursion point of the sub going in, and thereby the sub being limited to under 108dB for each tower (the target was to make tower speakers which could play up to 105dB, just because, which so far the Scanspeak 30W/4558T00, the SEAS U18RNX/P, and the SEAS Titan 27TAC/GB can all reach, so that should be doable with long term power limits of the speakers, with long term defined by the ISO meaning thereof), which means it isn't playing as loud as it may play in other scenarios (likely remaining around 70W for 105dB).

But, I will have more as the design is about to enter into the sketchup phase for what the math says the dimensions should be, which then will lead to the FEA calculations.

Just wanted to say you were not incorrect in understanding the question. Also thank you for the design input! I really enjoy gaining any and all insight that I can get.

Edit: What materials or products would you use for a viscoelastic layer in a CLD cabinet?


Chedwin liked
ReplyQuote
Chedwin
(@chedwin)
Eminent Member Registered
Joined: 4 months ago
Posts: 22
 

@ajc9988 Im not really able to give you mid layer material suggestion for CLD cabinet as usally when im applying layer damping principles its in the building construction design phase for bars and clubs using comercially purchased subwoofer(s) not self designed cabinets. For example im currently working on a concrete and gypsum board design with a triple 1200W 2x12" subwoofer array for a high end bar installation.

I can share the rough draft drawing for it showing how I normally use layer damping concepts in my work, this one is free layer damping where the gympsum layer is the damping material. Grey is large poured concrete block with a cavity cut out for the sub array to fit in, blue is gypsum board lining top, sides and back, red is wooden battens to isolate subs from floor and yellow is a plywood surround. The cavity in the concrete is about 2m wide and 3m deep.

Untitled 1

I dont think any of the materials I would normally be using in buildings would be suitable to CLD a speaker cabinet. You could try 1/4" rubber gym matting/flooring. Never tested it but thats the best thing coming to mind right now that seems worth attempting

Josh Evans
- Professional Live Sound Engineer
- High End Commercial AV Install Technician


ajc9988 liked
ReplyQuote

ajc9988
(@ajc9988)
Estimable Member Registered
Joined: 6 months ago
Posts: 126
Topic starter  
Posted by: @chedwin

@ajc9988 Im not really able to give you mid layer material suggestion for CLD cabinet as usally when im applying layer damping principles its in the building construction design phase for bars and clubs using comercially purchased subwoofer(s) not self designed cabinets. For example im currently working on a concrete and gypsum board design with a triple 1200W 2x12" subwoofer array for a high end bar installation.

I can share the rough draft drawing for it showing how I normally use layer damping concepts in my work, this one is free layer damping where the gympsum layer is the damping material. Grey is large poured concrete block with a cavity cut out for the sub array to fit in, blue is gypsum board lining top, sides and back, red is wooden battens to isolate subs from floor and yellow is a plywood surround. The cavity in the concrete is about 2m wide and 3m deep.

Untitled 1

I dont think any of the materials I would normally be using in buildings would be suitable to CLD a speaker cabinet. You could try 1/4" rubber gym matting/flooring. Never tested it but thats the best thing coming to mind right now that seems worth attempting

A couple of the materials I have considered to date (some of which agree with your suggestion) include:

1) isolation material for sub-flooring, like Peacemaker underlayment ( https://www.audimute.com/peacemaker-soundproofing-underlayment );
2) mass loaded vinyl;
3) viscoelastic polyurethane (two pail mix type, sometimes used in marine and engine rooms for noise isolation, etc.; not meaning green glue, just am spacing on the name of the company I did think of getting some from); and
4) bitumen rubber.

I would look for cheaper construction materials that have similar properties, but many have VOCs that both limit for inside use and also the chemicals must be checked in comparison to the construction materials of the drivers themselves as the VOCs can harm some speaker drivers.

But I did not think about rubber gym matting rather than underlayment. Might need to price that out to see how well it works.

Edit: Pyrotek's Decidamp DC30. That was the poly barrier I was thinking of. https://www.pyroteknc.com/products/decidamp/decidamp-dc30/
Which I would also take recommendations on market alternatives to decidamp dc30 as well.


ReplyQuote
Chedwin
(@chedwin)
Eminent Member Registered
Joined: 4 months ago
Posts: 22
 

@ajc9988 sub flooring isolation material like that one is maybe a bit on the thin side for a CLD speaker, better suited to stopping vibrations from footfall but could be usefull on the bottom of the cabinet to help isolate the cabinet from standing surface though depening what its sitting on. looks like they have separate thin/dense to block sound waves permeating a surface and thick/less dense to stop vibrations travelling along a surface. thick and dense would probably be more ideal in a speaker

mass loaded vinyl should work, i have 5kg/m2 mass loaded vinyl as an underlay for my LVT floor instead fo regular underlay. although similar to the sub flooring isolation material MLV is often available thicker whilst maintaining density. MLV should also be available cheaper than an underlayment, at least here in the uk it is

paint on solutions like DC30 should work very well acoustically but are very permenant, only use if you know you never need to separate the surfaces for any reason

 

of the mentioned options if say rubber gym mats or suitable density/thickness mass loaded vinyl if you want a less permentant more flexible solution or DC30 (or simlar) if your happy to sacrifice the ability to take apart for some improved performance

 

Josh Evans
- Professional Live Sound Engineer
- High End Commercial AV Install Technician


ajc9988 liked
ReplyQuote

ajc9988
(@ajc9988)
Estimable Member Registered
Joined: 6 months ago
Posts: 126
Topic starter  

@elliottdesigns - Update on the ADAU1701 programming. I made my first attempt to do a mono Left channel programming for the upcoming speakers, with a 3-way double precision Linkwitz-Riley 48dB crossover, an infrasonic filter with a 6th order Butterworth (brought down the top dB of the sub, but the sub can now reach deeper with F3 of around 19.48Hz), an FIR filter with 550 points, a limiter, a gain per driver (sub, woofer, tweeter) potentiometer and overall volume knob, a signal detection circuit to turn it off when nothing is coming, and inversion switches in case you plug the drivers in backwards.

Note, this is before programming in any solutions in the event that time delays or phase are issues.

Now this is a first attempt, so please do not think this is done. It compiled without an error, but that is not the same as a finished, working product. I also plan to make a version that uses SigmaStudio's autoEQ to check results.

This is the output data from the compiler:

################## Summary ########################
(Note: Estimates are based on a 48 kHz sample rate)

Number of instructions used (out of a possible 1024 ) = 1001

Data RAM used (out of a possible 2048 ) = 766

Parameter RAM used (out of a possible 1024 ) = 683

 

So, as you can see, I'm running out of the finite instructions allowed (the ADAU1701 is limited as it can only process so many instructions in a given period). What I might start looking into is adding an ARM developer board to the mix (analog discusses this as using a microcontroller with the DSP, with a recommendation of an arduino like the Teensy 4.0 available at https://www.pjrc.com/store/teensy40.html for $20). It would take learning more depth on the programming to integrate it on the software side, and adding an additional mount to my home-made plate amp, but that would greatly increase the number of instructions allowed, which also means having an even better FIR filter, thereby getting more comparable to a SHARC (which has a version with a 1GHz ARM processor, or the mini with the 500MHz cores). Granted, the Teensy 4.0 only has a 600MHz frequency, thereby being more comparable to the mini, but this is only the start of my search. I need to see what other boards have been paired with the ADAU1701 (or analog chips, generally) and which people have already poured the hours necessary in to make them work together. If I find that, then adding in a board will happen (if not, but I find enough on adding the Teensy 4.0 or 4.1, then I'll do that, because $80-120 for the boards and standoffs is not bad).

 


ReplyQuote
ElliottDesigns
(@elliottdesigns)
Estimable Member Moderator
Joined: 2 years ago
Posts: 210
 

@ajc9988 

Nice, I've been working on moving over to raspberry pi 4s, yes they are expensive and very involved, but once I've figured out, on a per speaker basis I will have more DSP capability than the 2x4 HD, better sound quality, and the capability to add a music server, WiFi capabilities to allow for Spotify streaming, multiroom playback and even a smart assistant. Once I've completed all of this (which will take a long while) I'll be completing a comprehensive guide on how to do it on this forum.

Elliott Dyson - Mechanical Engineering Student and 3D printing & Design Freelancer


ajc9988 liked
ReplyQuote

ajc9988
(@ajc9988)
Estimable Member Registered
Joined: 6 months ago
Posts: 126
Topic starter  

@elliottdesigns - https://digital-audio-labs.jimdofree.com/english/raspberry-pi/adau1701-i2s-driver/

One of the first searches for rpi I did popped that up, which means there is a way to use the i2s to possibly have the rpi as the microcontroller for the 1701 boards, which will be cool to have that with the jab5s I already have.


ReplyQuote
ElliottDesigns
(@elliottdesigns)
Estimable Member Moderator
Joined: 2 years ago
Posts: 210
 

@ajc9988 Yeah, you could do that, the audio systems for processing the audio pipeline and doing the DSP are in their very early days and are next to impossible to get working at the moment, let alone easily. That's why I'm working through it at the moment and then I'll be creating the guide. I've been spending 2-4 or more hours a day for the past two weeks on figuring out what I can get working, so I'm trying to put the time in for you all, just saying before you go in and try getting it to work yourself, trying to save you the effort 😅.

Elliott Dyson - Mechanical Engineering Student and 3D printing & Design Freelancer


ajc9988 liked
ReplyQuote

ajc9988
(@ajc9988)
Estimable Member Registered
Joined: 6 months ago
Posts: 126
Topic starter  

@elliottdesigns - I already have the JAB5s. I have for months. And I moved the signal detection circuit earlier. I stopped the popping sound I was getting with sudden power offs and honored the filters better. The subs have cleaner sound with proper infrasonic filter and crossover applied. But there is definitely more work. Just hoping the price of plywood drops soon...

 

Edit:
Also, since I already have the ADAU1701, I'll be working from that end on posting general sigmastudio builds and learning on using an arm chip tied in potentially. If the RPi by itself or tied with a 1701 work better, then that will be included down the road. But, looks like we all have some really cool projects working!


ReplyQuote
ajc9988
(@ajc9988)
Estimable Member Registered
Joined: 6 months ago
Posts: 126
Topic starter  

@elliottdesigns - Hey, so how many coefficients can you run on the RPi 4? The arduino solution can do 20,000, but in 0.367 seconds. Trying to check on the RPi 4 and how much memory is needed for the number of coefficients processed. Mainly because I can buy a 2GB RPi 4 now without an issue to start toying with development, but need to know if the NEON and FP are that much better that it can spank the coefficient processing of the arduino solution.


ReplyQuote

ElliottDesigns
(@elliottdesigns)
Estimable Member Moderator
Joined: 2 years ago
Posts: 210
 

@ajc9988 ah, not sure about latency since I use linear filters and make sure each of my driver's have the same amount of taps of processing applied to them, so sorry I cant help with that part. I'm using a pi4 4gb, not sure how many the 2gb can handle but I have no where near reached the maximum the pi is capable. I know lesser hardware has been able to run upwards of 1,000,000 taps in total. So 🤷‍♂️. Also what is NEON and FP?

Elliott Dyson - Mechanical Engineering Student and 3D printing & Design Freelancer


ReplyQuote
ElliottDesigns
(@elliottdesigns)
Estimable Member Moderator
Joined: 2 years ago
Posts: 210
 

Oh, just found out that this figure took place in 400ms. But latency varies widely depending on what audio pipeline you use. For example pipewire is the best by far but it's still in early-ish development, although perfectly usable for the most part

Elliott Dyson - Mechanical Engineering Student and 3D printing & Design Freelancer


ajc9988 liked
ReplyQuote

ajc9988
(@ajc9988)
Estimable Member Registered
Joined: 6 months ago
Posts: 126
Topic starter  

@elliottdesigns - FP = Floating Point, which the ability to do floating point calculations can help expedite the execution of the coefficients. Neon is an Arm Neon is an advanced single instruction multiple data (SIMD) architecture extension for ARM processors they started including.

Neon instructions allow up to:
  • 16x8-bit, 8x16-bit, 4x32-bit, 2x64-bit integer operations 
  • 8x16-bit*, 4x32-bit, 2x64-bit** floating-point operations 

 

The extra compute is where the A series, at a certain point, overtakes the inherent abilities of the DSP module within the Cortex-M series microcontrollers. It takes more footprint, but the additional compute really can give it run away advantages.

And that is why I was trying to figure out the coefficients per time period on the A72 architecture (IIRC that is the Arm version for the RPi4, that or A73, but I think it was 72). The main reason is for this to work with a TV as these speakers will not solely have music. The delay doesn't matter for music so long as all channels are synced, which the same number of taps helps to do. Instead, you can only put a delay on the TV for so long, so being able to finish the FIR filter in a reasonable time is something I'm also considering in the design.


ReplyQuote
Page 3 / 4

Newegg

Share:
%d bloggers like this: