A while back I built a couple of sealed boxes for 12" Rythmik kits. At the time I followed advice from folks at the HomeTheaterShack.com and built a 56 liter net internal volume. Net meaning net of bracing and of volume displaced by the driver, so real 56 liters of air inside the box. These are stiff, heavy, nicely finished boxes I want to keep. However, as I learned more I modeled the subs in Unibox and Qtc turns out to be below 0.5.
How should I go about adjusting this to get to 0.707? I have a measurement mic and use REW. I could take the subs somewhere outside on an open area to measure response without the room. But what should I measure and what should I work on?
Thanks!
How should I go about adjusting this to get to 0.707? I have a measurement mic and use REW. I could take the subs somewhere outside on an open area to measure response without the room. But what should I measure and what should I work on?
Thanks!
Hi, just because others have said Q should be 0.707 doesn't mean it's automaically what someone wants. It's only a guide. 0.707 "can" give a smooth response down to fb, & then a gentle 12dB rolloff afterwards. But, it all depends on the room/space the box is in ! In a car etc, this would most likely not be the case, as that smallish enviroment leads a Cabin Gain. So a Q of 0.5 or less would help to compensate for that.
In a room it's more likely to be a lot different, it all depends on the size & box placement. In this case a Q of 0.707 could be beneficial.
After you measure the frequency response, & if you decide to go for a higher Q, you would need to make the internal volume smaller. Experimenting with Unibox will show you the vol you require for a certain Q. However, making the vol smaller will raise the fb, as Unibox will reveal.
In a room it's more likely to be a lot different, it all depends on the size & box placement. In this case a Q of 0.707 could be beneficial.
After you measure the frequency response, & if you decide to go for a higher Q, you would need to make the internal volume smaller. Experimenting with Unibox will show you the vol you require for a certain Q. However, making the vol smaller will raise the fb, as Unibox will reveal.
Reduce the internal volume of the box. Do check that a higher Q is actually what you want, though.
Thanks for chiming in.
OK, I need to reduce volume. So Unibox tells me I need to get to 17 liters to achieve Qtc=0.71. On paper, I need to reduce 39 liters. That's a lot!
Anyway, I could fill several bags with sand until I reach 39 liters and throw them inside the box, or place 39 liters of fiber glass panel into the box, for example. Either way would be fine?
But then I want to measure something so I confirm I achieved 0.707 - or whatever Q I finally choose to go for. What should I measure?
BTW, the current design has an Fb of 41Hz, whereas at 0.71 Fb would be 63Hz. What's the implication of this? Side effects?
OK, I need to reduce volume. So Unibox tells me I need to get to 17 liters to achieve Qtc=0.71. On paper, I need to reduce 39 liters. That's a lot!
Anyway, I could fill several bags with sand until I reach 39 liters and throw them inside the box, or place 39 liters of fiber glass panel into the box, for example. Either way would be fine?
But then I want to measure something so I confirm I achieved 0.707 - or whatever Q I finally choose to go for. What should I measure?
BTW, the current design has an Fb of 41Hz, whereas at 0.71 Fb would be 63Hz. What's the implication of this? Side effects?
The room itself will change the q, so you have to measure it outside if you want a BOX q of .707, and the SYSTEM q will only be .707 if you use it outside where you measured the .707 box q. The response curve will tell you the q.
The implication is everything will change. Response curve, impedance curve, displacement for a given power level/frequency, it will be a different design in general.
What is the reason for wanting to change the q? You can change the q with eq too without physically altering the box. It's not exactly the same thing but it is a way to change q.
There is no perfect q. Perfection is flat response at the listening position. The only reason .707 q is popular is because it provided a decent chance of getting close to flat response back before it was easy to simulate and measure and eq response.
Last edited:
Hi,
only 19L seems quite low for a 12" driver .... 56L looks more reasonable.
The calculation of the tuning of a CB volume is easy, You could do it by hand.
You just need Fs, Vas and Qts from the TSPs of the driver and the desired Qtb of the box.
Then:
Vb= Vas/((Qtb/Qts)e2-1) and Fb= Fs*(Qtb/Qts)
Example: A driver with an FS of 25Hz, a Vas of 90L and a Qts of 0.35 shall be tuned to a Qtb of 0.7
Qtb/Qts=2 --> Vb=Vas/3=30L and Fb=Fs*2=50Hz.
Just rearrange the Volume-formula to calculate Qtb and Fb for a given Volume Vb.
jauu
Calvin
Btw: what are the TSPs of Your chosen driver?
only 19L seems quite low for a 12" driver .... 56L looks more reasonable.
The calculation of the tuning of a CB volume is easy, You could do it by hand.
You just need Fs, Vas and Qts from the TSPs of the driver and the desired Qtb of the box.
Then:
Vb= Vas/((Qtb/Qts)e2-1) and Fb= Fs*(Qtb/Qts)
Example: A driver with an FS of 25Hz, a Vas of 90L and a Qts of 0.35 shall be tuned to a Qtb of 0.7
Qtb/Qts=2 --> Vb=Vas/3=30L and Fb=Fs*2=50Hz.
Just rearrange the Volume-formula to calculate Qtb and Fb for a given Volume Vb.
jauu
Calvin
Btw: what are the TSPs of Your chosen driver?
Don't put ANY fiber glass or similar material in the box, as this will have the opposite result. To lower the Vol, the material must be solid. Dry sand bags would do.
17 liters does sound small !
Yes, post the TSP specs.
17 liters does sound small !
Yes, post the TSP specs.
Here are the TSPs:
Rythmik DS1200
Fs: 26.00Hz
Re: 2.80Ohm
Qms: 7.42
Qes: 0.33
Sd: 530.0cm2
Vas: 88.0l
Xmax peak: 17.78mm
(Le): 1.00mH
Power: 300.0W
OK, fiberglass is out.
Sure, 17 liters sounds very small. It's what UniBox is showing, though...
BTW, in my future setup with active XO there will also be digital room correction so I can force the driver to behave as expected. However, the less eq that is needed, the better. Hence I'm trying to "correct".
BTW you now have me wondering if I should focus on the Q in the open or in the room?
Rythmik DS1200
Fs: 26.00Hz
Re: 2.80Ohm
Qms: 7.42
Qes: 0.33
Sd: 530.0cm2
Vas: 88.0l
Xmax peak: 17.78mm
(Le): 1.00mH
Power: 300.0W
OK, fiberglass is out.
Sure, 17 liters sounds very small. It's what UniBox is showing, though...
BTW, in my future setup with active XO there will also be digital room correction so I can force the driver to behave as expected. However, the less eq that is needed, the better. Hence I'm trying to "correct".
BTW you now have me wondering if I should focus on the Q in the open or in the room?
Are these measured? Qes must be much larger for a published 0.472 Qts, 11.24 N/A BL.
Factor in that the published specs are questionable when I let WinISD calculate Vas based on Sd, Fs, Qms, it's ~107.25 L Vs 88 L published, so for a 0.707 Qtc it's more like ~95.6 L.
Even if 88 L is used, then Qms drops to ~2.04, Qes rises and still needs ~76.68 L, so something's amiss...........
Correct, the room dominates, so if the sub is anywhere near a wall, and especially a corner, then depending on how much room gain there is can require a somewhat lower Qtc and since a 0.5 Qtc is critically damped, a good choice if cab size and power handling is acceptable; otherwise, a 0.577 Qtc is a good compromise.
GM
Factor in that the published specs are questionable when I let WinISD calculate Vas based on Sd, Fs, Qms, it's ~107.25 L Vs 88 L published, so for a 0.707 Qtc it's more like ~95.6 L.
Even if 88 L is used, then Qms drops to ~2.04, Qes rises and still needs ~76.68 L, so something's amiss...........
Correct, the room dominates, so if the sub is anywhere near a wall, and especially a corner, then depending on how much room gain there is can require a somewhat lower Qtc and since a 0.5 Qtc is critically damped, a good choice if cab size and power handling is acceptable; otherwise, a 0.577 Qtc is a good compromise.
GM
Last edited:
Servo!
Hello all,
Just want to point out that this particular driver is part of a feedback loop with the amplifier. As such the amplifier will put out whatever voltage (current?) necessary to trace the waveform within the bandwidth of the servo correction loop.
So: too small a box means the amplifier won't be able to overcome the strong air spring of the box compliance; however too LARGE a box can be damaging as the servo can then overshoot as it's expecting a stiffer air spring (smaller box).
From this page (Rythmik Audio 12" servo subwoofer " DS1200 Custom Installation subwoofer) for the DS1200 kit, they suggest 2 cubic feet, which is right around 56L.
Executive summary: too small a box will simply reduce maximum output as the amplifier won't be able to overcome the air spring; too large and the amplifier can mechanically damage the speaker through overshoot. 56 liters seems right on the money, so I'd leave as is!
Cheers,
Tal
Hello all,
Just want to point out that this particular driver is part of a feedback loop with the amplifier. As such the amplifier will put out whatever voltage (current?) necessary to trace the waveform within the bandwidth of the servo correction loop.
So: too small a box means the amplifier won't be able to overcome the strong air spring of the box compliance; however too LARGE a box can be damaging as the servo can then overshoot as it's expecting a stiffer air spring (smaller box).
From this page (Rythmik Audio 12" servo subwoofer " DS1200 Custom Installation subwoofer) for the DS1200 kit, they suggest 2 cubic feet, which is right around 56L.
Executive summary: too small a box will simply reduce maximum output as the amplifier won't be able to overcome the air spring; too large and the amplifier can mechanically damage the speaker through overshoot. 56 liters seems right on the money, so I'd leave as is!
Cheers,
Tal
Good catch!!!
Geeez!! So there was a good reason after all why I was told to use 56 liters. I was beginning to go in circles. Glad you caught this.
Thanks all for helping out!!
Now, onto midbass box design. Fun project.
Geeez!! So there was a good reason after all why I was told to use 56 liters. I was beginning to go in circles. Glad you caught this.
Thanks all for helping out!!
Now, onto midbass box design. Fun project.
Agreed, after I posted I got to wondering about how the electronics might impact cab design, but didn't have time to get educated before you saved me the effort. 🙂
Still, I've noticed quite a few folks over the years have used just the driver in DIY projects, so wonder what its actual specs are.
GM
Another instance of the cart (models) pulling the horse (good sense). The ideal Q for a speaker is tiny. That is something you can get with motional feedback (and something approached by electrostats).
Models like Qs of .7 because they are computationally nifty. No other good reason what so ever.
Or maybe somebody has to explain it to me better.
Ben
Models like Qs of .7 because they are computationally nifty. No other good reason what so ever.
Or maybe somebody has to explain it to me better.
Ben
In audio Q means a lot of different things but in this case Q refers to the shape of the response curve. Tiny Q = no bass (rising response as frequency rises). Before it was easy to simulate, measure and eq, the only thing that could reliably produce a response curve that provided decent response in a variety of different environments was to use math to design for a goal Q. I already explained this. A Q of .7 will generally give a "max flat" response, not rising or falling but flat response down to tuning, and if f3 isn't too low this can blend in with room gain nicely allowing flat extension to lower frequencies. Lower q is usually described as "tight", "fast", "accurate" and higher q is usually described as "boomy" by people that can't measure and have to use subjective terms. Low q is usually described with those subjective terms because of it's lack of low bass.
Now the simulation software does this job for us and there's no need to do the math by hand, and we can now use the simulator and a bit of common sense to design a response curve that will work well with the room's response to provide flat response at the listening position (or whatever other goal the designer has in mind).
Where did you get the idea that there is an ideal q?
Last edited:
Hi,
Instead I'd worry about my common sense, if I needed a computer for these 1x1 calculations.
Praxis tells us for example that with the typical tolerances of the drivers involved a calculation to the second after-koma-digit is not required.
It also tells us that with driver parameters changing with power-/excursion level too precise calculation is meaningless.
So as Kant said ... have the guts to use your common sense .... what You get as response in Your listening room is not the response 'painted' on Your monitor anyway.
jauu
Calvin
btw: how to design a BR in less time than the computer requires for booting: 😉
Step1, calculate Fb and Vb for a Qtb of 0.58 (or 0.6 if easier)
Step2, divide the value of Fb by sqrt2 to get the tuning frequency Ft
Step3, calculate the bassreflex port (or choose a 4" diam, 6-8" long tube if easier).
Step4, build the cabinet using a variable length BR port, set to the calculated length
Step5, tune for best sound by listening 😉
As listed in #6 the formulas are soo easy, You don't need no stinkin sim 🙄 especially if the common sense is missing 😛
Instead I'd worry about my common sense, if I needed a computer for these 1x1 calculations.
Praxis tells us for example that with the typical tolerances of the drivers involved a calculation to the second after-koma-digit is not required.
It also tells us that with driver parameters changing with power-/excursion level too precise calculation is meaningless.
So as Kant said ... have the guts to use your common sense .... what You get as response in Your listening room is not the response 'painted' on Your monitor anyway.
jauu
Calvin
btw: how to design a BR in less time than the computer requires for booting: 😉
Step1, calculate Fb and Vb for a Qtb of 0.58 (or 0.6 if easier)
Step2, divide the value of Fb by sqrt2 to get the tuning frequency Ft
Step3, calculate the bassreflex port (or choose a 4" diam, 6-8" long tube if easier).
Step4, build the cabinet using a variable length BR port, set to the calculated length
Step5, tune for best sound by listening 😉
Last edited:
This post is excellent advice for everyone living in 1932. And for people that don't mind doing a bit of math and whose ONLY design goal is box Q (which is a really odd goal in modern times, since room interaction and eq change system q).
For people living in 2015 and people that realize that there's more to design than just Q, simulators are an excellent tool.
Here's a few things simulators do better than the formulas you posted.
Sims are faster. I can enter t/s specs and have a sim in under 60 seconds, faster than you can calculate your formula. With an additional 10 seconds I can run through every box size from infinitely small to infinitely large AND see the response curve and a boatload of other information at the same time. How long would it take you to run through your formula an infinite amount of times? Other than box size, what else does your formula tell you? My computer is always on, boot time is not a factor.
A simulator helps to visualize the response curve. Good luck with that with a formula.
A simulator shows excursion AND power level required to reach xmax AND potential max spl. This is very useful in selecting a box size that works with a particular amp, selecting the amount of drivers necessary to reach a desired spl, and a bunch of other useful information. Your simple formula does none of that.
A simulator can output .frd and .zma files to be used with other simulators to create crossover networks to maximize phase and response integration with other drivers in the system. Tell me how your formula deals with that.
Drivers with high inductance (Le:Re ratio more than 1) don't follow the rules, your formula will absolutely not give the correct result for these drivers - check measurements of high inductance drivers at data-bass.com to see how much these drivers deviate from a simple model. A simple tweak in the simulator will give much more accurate results for high inductance drivers; results that are actually usable.
A simulator will show velocity, which is absolutely required for designing ported boxes. Your simple ported box formula is a recipe for disaster.
A simulator will allow you to design any ported box response curve you desire, your formula appears to be about as flexible as a brick.
Simulators cannot predict in room response but this is a moot point since your formula can't do it either.
This is just a VERY short list of advantages that simulators have over your antique formula and design method, so I'm not worried about my common sense but I do question the common sense of anyone using this type of inflexible and uninformative longhand math design routine instead of taking advantage of modern tools.
In other words, your formulas are completely useless unless you have a low inductance driver and the ONLY thing you care about is box Q, which is an extremely odd design goal in itself.
Last edited:
Hi,
Yes You can .... do this and that.
In the end You can sim for hours and ages, and if You don´t know what is simmed how and what for, the end result can still suck badly.
What if Your ears tell You Q=0.72 and 30L is perfect, but Your simulator tells You it should be 0.633254 andd 35.56874L?? 🙄
I´m alot more comfortable using a simulator if I have a rough idea in advance whereabout my results should be.
Cause I may perfectly be a Dinosaur I´d like to give You some food for thought: 😉
The towers of York Minster in northern England, built over a time span of 250 years till 1472, weigh about 16,000tons!
Now the area they built this brilliant piece of architecture in was a swamp.
How do You think did the architects manage to foundate the enormous weight properly and standing upright till today, instead of the whole shebang simply sinking in or tilting in a short period of time?
Do You know?
I also wonder which simulator or design software they could have used.
Do You know?
If not .. its worth a visit.
jauu
Calvin
Yes You can .... do this and that.
In the end You can sim for hours and ages, and if You don´t know what is simmed how and what for, the end result can still suck badly.
What if Your ears tell You Q=0.72 and 30L is perfect, but Your simulator tells You it should be 0.633254 andd 35.56874L?? 🙄
I´m alot more comfortable using a simulator if I have a rough idea in advance whereabout my results should be.
Cause I may perfectly be a Dinosaur I´d like to give You some food for thought: 😉
The towers of York Minster in northern England, built over a time span of 250 years till 1472, weigh about 16,000tons!
Now the area they built this brilliant piece of architecture in was a swamp.
How do You think did the architects manage to foundate the enormous weight properly and standing upright till today, instead of the whole shebang simply sinking in or tilting in a short period of time?
Do You know?
I also wonder which simulator or design software they could have used.
Do You know?
If not .. its worth a visit.
jauu
Calvin
Last edited:
How can a simulator tell you what Q should be? I don't know of any simulator that even attempts to tell you what sounds good. If you can't figure that out on your own with all the data that a simulator provides I certainly would advise against trying to get there with a simple formula that gives ONLY box size for a user selected Q value.
The fact is that a simulator will give a metric ton of info including q. If you don't think anything other than q is even worth looking at... I don't even know what to say about that. It sounds like you are encouraging people to design with the absolute least amount of information possible, with the only goal being a randomly selected box Q value, this doesn't make any sense to me at all.
I don't know anything about architecture, certainly nothing about the structure you are referring to. I'm not sure how that relates to this conversation but the one thing that I can say with a fair degree of certainty is that no modern large scale building is designed without help from computers. Not even sure why you would want to try when there are so many tools available. Computers help with everything from structural analysis to hvac vent sizing.
Last edited:
Hi,
I´m not against using simulators.
All I wanted to express that they are by no means the help and cure for all.
You still have to know at least roughly what You´re doing and what Your outcome will be.
But then there will probabely be people around who need a GPS to find their way to the loo. 😛
If after the formulas had been given in #6 You still need a simulator, I´d assume You need to learn at least the basicst basics of Maths and Box design. 🙄
The answer to the TOs question certainly required no computer, but a pocket calculator at most.
jauu
Calvin
I´m not against using simulators.
All I wanted to express that they are by no means the help and cure for all.
You still have to know at least roughly what You´re doing and what Your outcome will be.
Well, the thread headline reads "how to tune my subs to a Q=0.707" and not "I want tons of information".
But then there will probabely be people around who need a GPS to find their way to the loo. 😛
If after the formulas had been given in #6 You still need a simulator, I´d assume You need to learn at least the basicst basics of Maths and Box design. 🙄
No, I´m just suggesting, that one should use his or her brains and common sense before and whilst simulating.
The answer to the TOs question certainly required no computer, but a pocket calculator at most.
jauu
Calvin
Yes, because it has many advantages and it´s used by trained people knowing their field, but still, we´ll wait and see when those marvels of modern architecture collaps, or if they remain standing and functional for >800years also. 😉
Absolutely, I agree.
Designing a box with the only goal being the box Q and having no other goals or information about what you are doing is an extremely bad idea. Sure, you can do it with only your simple formula (as long as the driver isn't a high inductance driver), but why would anyone want to?
The goal of modern architecture is almost never longevity. Things change so fast it's no longer a desirable goal to have a building last for thousands of years. The goal has shifted to inexpensive and decent quality. This was not the case when my brother's Victorian house was built, no expense was spared, quality was phenomenal and the original roof lasted over 100 years.
But just to be really really clear here, are you suggesting that they knew something over 800 years ago that we don't know now? Are you suggesting that modern tools and simulators couldn't design and construct a building that would last indefinitely? It's been pretty well proven that stacking a bunch of big rocks can last thousands of years. I think we have the technology to do a bit better than that now, but budget is an issue.
Anyway, this isn't an architecture forum so it doesn't really matter.
- Status
- Not open for further replies.