| Geek |
| Very handy. Thanks! :) |
|
|
| Curmudgeon |
I've had good luck with the cube root of 2;
.794/1/1.26; and variants such as .794/1/2.52. |
|
|
| Vikash |
| Updated with new ratios. |
|
|
| Ron E |
The square root of 2 is ~1.414, not 1.141.
Usual quoted enclosure ratios are based on:
phi = golden ratio = 1.6180339887499 why do I remember this number to so many decimal places? ;)
1.2
1.25
IMO, none of these is better than any other or any number you could come up with. It makes a lot more difference how you mount your drivers on the baffle (for diffraction) than what the box dimension ratios are. Lining or stuffing in the box is used to damp out the resonances.....
Your app calculates axial resonances, but there are also tangential and oblique resonances, and all are excited differently depending on where you mount the drivers... |
|
|
| Curmudgeon |
Tapered boxes are good too; but these ratios are certainly better than integers, 1/2/2 for example. AND you should also line &/or stuff. Nothin' wrong with belt and braces. And positiong drivers non-equidistant from box edges, (which should have radii > 4 inches) also helps.
But one thing at a time, this is a nifty little calculator. |
|
|
| Cloth Ears |
| quote: | Originally posted by Ron E
Usual quoted enclosure ratios are based on:
phi = golden ratio = 1.6180339887499 why do I remember this number to so many decimal places? ;)
1.2
1.25
IMO, none of these is better than any other or any number you could come up with... |
Ron,
Agreed. But I tend to use the square root of phi (c. 1.27202) for the front of my woofer and subwoofer boxes as it looks better, rather than for any sound reasons. That way the ratio W/H is 1.618/1 (and when you're looking at it from a standing position, perspective also fools you into thinking it's the same ratio W/D).
Internal standing waves I try to damp out (or confuse) by:
1) damping materials;
2) internal bracing with small (eg 2" or 3") holes in it;
3) internal bracing at odd angles and shapes (eg. dowels of different sizes, braces at 15, 22.5, 30 and 45 degrees).
UNLESS I'm doing a TL...
But, Vikash, I think you've got 1.28 in your calculator anyway - and that's close enough to 1.27202 |
|
|
| planet10 |
Nice tool Vikash....
dave |
|
|
| Vikash |
| Any input on the default threshold value? How close is too close? |
|
|
| Vikash |
Added ability to factor ratios.
What's the best way to render a pic on the web these days. Is it somehow possible to output code into renderer that output to a gif or something... |
|
|
| Vikash |
| Hi, someting like that yeah. If there was a simple 3D one, that would be quite cool. But this plotting one might come handy with one of my other projects. |
|
|
| planet10 |
| quote: | Originally posted by Vikash
Any input on the default threshold value? How close is too close? |
A number based on a percentage of the size of the wavelength is probably better than a fixed number....
I see you removed the very useful 1.28 ratio (when you put it back it might as well be 1.272.
dave |
|
|
| Vikash |
Dave the square root of Phi (1.272) can be used by selecting 0.5 from the factor dropdown. Try some others as well and you might end up with dimensions that look half appealing / have a realistic baffle width ;)
http://www.vikash.info/audio/standi...h=&threshold=50 |
|
|
| planet10 |
| quote: | Originally posted by Vikash
the factor dropdown. |
Actually, that is totally confusing... a bit of an explain of what it does would be useful...
dave |
|
|
| Vikash |
| I'm really bad at explaining things in a concise manner. :xeye: I've given it a shot |
|
|
| planet10 |
OK, now you need to add a feedback mechanism with the actual currently selected ratio.
dave |
|
|
| matjans |
[slightly OT]
vikash, can't you just use php/GD ? quite some work i'd imagine but it'll let you just 'draw' from inside your php scripts
http://www.php.net/GD
[/OT] |
|
|
| Vikash |
I use perl ;) I thought about using SVG as I have done before, but I'm there must be a way of rendering simple 3D objects which already have the lower routines implemented.
If there was an equivelant of php/GD for perl then that would work. I'll have to rummage through cpan again. Couldn't see anything obvious last time... |
|
|
| Geek |
| quote: | Originally posted by Vikash
I use perl ;) |
print "There can be only one\n";
:cool:
| quote: | | If there was an equivelant of php/GD for perl then that would work. I'll have to rummage through cpan again. Couldn't see anything obvious last time... |
There is. If you look through Freshmeat for perl-based image galleries, it'll list the equivalent module. I forget what it is, but that's how I found it ;) |
|
|
| Andy Graddon |
There are an infinite no. of irrational numbers between 1 and 2, have fun searching ;-)
Another method of designing is to use 3 numbers that are relative primes. Again, can be fun !!
Go for it !! |
|
|
| simon5 |
Using Andy Graddon method with prime numbers, I managed to get a perfect result ie no highlighted red results.
307 mm
479 mm
631 mm
Vikash, as a suggestion, you could add a logarithmic highlighter instead of a linear one.
I mean 1.00 kHz and 1.02 kHz are not higlighted but if you have 24 Hz and 34 Hz it will be highlighted. IMHO 1.00 kHz and 1.02 kHz are worse I think? Maybe you could make the highlighting threshold customizable too!!!
Nice calculator, to be able to customize everything is perfect. |
|
|
| planet10 |
| quote: | Originally posted by simon5
307 mm
479 mm
631 mm |
0.641:1:1.317
dave |
|
|
| johngalt47 |
Doesn't all this presuppose that you can cut wood to that accuracy? (assuming you use wood)
I can see where it might work with a CNC machine.... |
|
|
| Vikash |
The idea is to play with the dimensions to make sure there aren't any fundamental flaws. Cutting to mm perfection is certainly not realistic. The ratios are added just to suggest some basic ideas. The intention was to work with the suggested numbers by rounding to your requirements and having a quick check that you're not making any major mistakes from the offset.
As Ron has said, it more complicated then this. But I think it is a good starting point. ;) |
|
|
| Ron E |
Here's a little pascal program I wrote back in ~1992 and never finished - as you can see from the descriptively named commented out procedure calls at the end. I wrote this for room dimensions, but it could be modified for box dimensions.
In my old fortran days I had a program that calculated all modes, sorted them and did a bunch of statistics and rated the rooms based on mode spacing. I only wrote teh good rooms to a file and I came up with a list of room sizes that worked best for room acoustics. This type of analysis is often interesting to young engineers who are trained that everything is sort of deterministic in nature.
You could likewise generate effective ratios for box design. You should remember that foam or fiberfill are very nearly perfect absorbers of more than 1-2kHz or so, so those frequencies wouldn't matter much.
Have fun.
----------------------------------------------------
program modes;
var length, width, height : real;
procedure getroomsize;
begin
writeln('Enter the room dimensions in inches');
writeln;
write('Length of room ?');
readln (length);
write('Width of room ?');
readln (width);
write('Height of room ?');
readln (height);
end;
procedure modestofile;
type freqarray = array[0..10,0..10,0..10] of real;
filenametype = string[255];
var i, j, k, lmax, wmax, hmax : integer;
f : freqarray;
f1 : text;
filename : filenametype;
begin
filename := 'output.dat';
assign(f1, filename);
rewrite(f1);
lmax := trunc(300*length/6786);
if (lmax>10)then lmax:=10;
wmax := trunc(300*width/6786);
if(wmax>10)then wmax:=10;
hmax := trunc(300*height/6786);
if(hmax>10)then hmax:=10;
for i := 0 to lmax do
for j := 0 to wmax do
for k := 0 to hmax do
begin
f[i,j,k] := 6786*sqrt((i*i/(length*length))+(j*j/(width*width))+(k*k/(height*height)));
if ((f[i,j,k]<300.0) and ((i+j+k)>0))then
begin
writeln(f1, i:3, j:3, k:3, f[i,j,k]:8:2);
end;
end;
close(f1);
end;
begin
getroomsize;
modestofile;
{plotmodes;}
(*sortmodes;*)
(*printaxial;*)
end. |
|
|
| planet10 |
| quote: | Originally posted by Ron E
Here's a little pascal program I wrote back in ~1992 |
I had one i wrote in System/360 assembler (late 70s), and later one i wrote in Forth... not quite as elaborate as yours thou...
dave |
|
|
|
