AVS Forum banner

1 - 20 of 32 Posts

·
Registered
Joined
·
50 Posts
Discussion Starter #1
I will be building a couple sonotube enclosures for my UM18-22's. I would like it tuned to 17(ish) hz in a 20" diameter tube.

I see the recommended enclosure size for infrasonic is like 12cu.ft. Not really possible for my space constraints.

Can I adjust the port size of an 8cu.ft enclosure to achieve an 17hz tune? Possibly lower? I vaguely understand that air velocity through the ports can be an issue?

What are your thoughts.
 

·
Registered
Joined
·
316 Posts
I am hardly a tuning reflex expert, but I can say that there are potential issues. You can tune it lower by going longer, and the the potential issues are length of the port, turbulence or obstruction of airflow if you have to bend the port. You can tune it lower by using a smaller diameter, and you have potential issues due to air velocity and turbulence/chuffing, constriction of air flow at high output (making it act more sealed, losing extension at high output). You also have the possibility that as you tune lower with a smaller than recommended box volume, the output of the driver and port in the area between where either dominate will combine with a sag in the response. Whether or not that's a problem will depend on the degree.

I am sure somebody will come along with a more correct complete answer, but that is my better than nothing (hopefully) understanding.

The other thing is that if you're dealing with bass reflex, it might be a good idea to do modeling based on the measurements of the actual drivers that you receive, not the published specifications. In the case of the UM-18's, I got 4 driver samples, all of them very close to each other (which is good), but none of them at all near the published specifications. In my application, they worked out fine for what I needed, in a sealed application, that is relatively forgiving if you apply your driver correction carefully in DSP, but if you built a reflex system blindly based on the modeling or generic suggestions provided by manufacturer data, you might not be getting what you thought you're getting, and if you don't have a means to make ground plane measurements of the response outside, or some other method that separates what the subwoofer is doing, versus what the room is doing, you might assume your room has problems that it doesn't, and chase fixing one, or giving up, when in fact the problem is earlier up the chain.

Since P.E. has a no hassle return policy (only if you return it immediately), I would suggest ordering however many drivers you need, and a DATSV3 if you don't already have one, and testing the drivers immediately, and doing your modeling based on those results, as well as running them on high excursion near their limits free-air with some very low frequency tones, like 18 Hz or so, dan decide if they will do what you want from a modeling perspective, and how much of the 'linear' excursion is linear enough for you to use it. Results vary apparently. I don't know what the currently sold batch is like, and tech support probably doesn't know either, and if they tell you to expect the drivers to settle into published specifications with break in, I wouldn't count on that.

That being said, I am only one data point, and maybe my measurements are somehow screwed up, so don't take my experience as a necessary indication of the one you're likely to have. My suggestions is just to ensure that you don't. Trust but verify, as they say. :p
 

·
Registered
Joined
·
4,657 Posts
In the case of the UM-18's, I got 4 driver samples, all of them very close to each other (which is good), but none of them at all near the published specifications.

While very few people have/use DATS on their drivers, I don't think I've ever seen anyone have results that were way off from the sims using the quoted specs. If you have, can you point me to the thread/posts?

Also, did you posts the measurements from your drivers? I would like to see your results. Thanks.
 

·
Registered
Joined
·
316 Posts
While very few people have/use DATS on their drivers, I don't think I've ever seen anyone have results that were way off from the sims using the quoted specs. If you have, can you point me to the thread/posts?

Also, did you posts the measurements from your drivers? I would like to see your results. Thanks.
With the exception of the original databass measurements, which were the reason I bought the drivers in the first place, I also haven't seen anybody post actual measurements of either free-air parameters, or outdoor ground plane measurements that shows what the finished product is doing itself, without the inteference/help of the room effects.

Most everything I've seen anybody share is in-room response, and mostly on the emphasis of room tuning/'correction'. There's nothing wrong with that, but I wonder if my batch samples were temporary anomalies, and I just got unlucky, or if PE's Chinese supplier made a convenient running change, didn't tell anybody, and nobody noticed because most all of the people using them wouldn't notice a difference if they weren't actually measuring them, or didn't have older versions to compare. I really don't know, I can only wonder and speculate, but I'm not going to keep buying more drivers to see if production has shifted back to original parameters. P.E. tech support gave me some baloney about them being within the range of the 'Golden Rule' of expected break-in. I know drivers break in, but if they break in that much with a 50% shift after more than 60 hours of serious pounding, only shifting less than 5%... that's not breaking in, that's breaking. But anyway...

In my application, they're sealed, so while they definitely weren't the drivers as represented by the manufacturer, or measured by Josh, for my application where I'm not squeezing everything possible out of them, even if they're likely less efficient overall in the motor, and need more low frequency boost than predicted, once Chris P introduced and explained what a Linkwitz transform actually was, I applied at least a rough approximation and I've got to say that they sound pretty darn good, and in my living/kitchen area, the low frequency output capabilities are really more than adequate. I'm glad Rob Mosher talked me out of a pair of 12's. :)

3119880


FWIW, this is free-air measurements of one driver after an accumulated 64 hours of high-excursion full-time infrasonic 'break-in'.

If you figure out how often people watch movies, and how much time in those movies actually has high output low frequency content. It will probably depend a lot on viewing habits and content, but if somebody watched 16 movies a month, which is more than every other day, and the movies on 'average' at a full five minutes (I'm pulling that number out of my butt, I've never done an extensive sample of movies holding a stop watch while watching my woofers) of balls to the wall high excursion output (not just a handful of thuds spread across a long time), that 'normal use' equivalent would be 4 years. If it takes more than this to 'break-in' to published parameters, I'm guessing that most users who aren't fanatic enthusiasts never actually get their drivers broken in. But, I think it more likely that they were just making stuff up or repeating something that somebody else told them because it sounded cool and accounted for the driver's measured differences, and since they don't make their drivers, if a whole batch is out of spec and they don't want to return them and be out of stock for a whole long while and lose money to other products, there's nothing else they can do about it anyway other than just to sell the defects as 'normal'.

HoursTotal MinMov/MnthMin/MovMin/MnthMonths
64​
3840​
16​
5​
80​
48​


I didn't have enough mass around that would stay in one place and stick to the cone to get a full set of measurements, but in terms of Fs, Qts, Qes, they told me enough.

Put in a 4 cubic foot box with 4 pounds, 7 ounces of polyfill and a light piece of material between the fill and the driver, it measured like this...

3119881


Combined with an aggressive shelving filter, it got me under 20 Hz, in room, no problem, within a +/- 5 dB window, down to 16 Hz, and that's with an infrasonic filter implemented.

I love DSP. :p

I could have flattened it out more in that location, but I opted to just let it do what it was going to do beyond that. Splice response with main speakers from 10-180 Hz below before the playback curve was implemented.

3119882
 

·
Registered
Joined
·
50 Posts
Discussion Starter #5 (Edited)
With the exception of the original databass measurements, which were the reason I bought the drivers in the first place, I also haven't seen anybody post actual measurements of either free-air parameters, or outdoor ground plane measurements that shows what the finished product is doing itself, without the inteference/help of the room effects.

Most everything I've seen anybody share is in-room response, and mostly on the emphasis of room tuning/'correction'. There's nothing wrong with that, but I wonder if my batch samples were temporary anomalies, and I just got unlucky, or if PE's Chinese supplier made a convenient running change, didn't tell anybody, and nobody noticed because most all of the people using them wouldn't notice a difference if they weren't actually measuring them, or didn't have older versions to compare. I really don't know, I can only wonder and speculate, but I'm not going to keep buying more drivers to see if production has shifted back to original parameters. P.E. tech support gave me some baloney about them being within the range of the 'Golden Rule' of expected break-in. I know drivers break in, but if they break in that much with a 50% shift after more than 60 hours of serious pounding, only shifting less than 5%... that's not breaking in, that's breaking. But anyway...

In my application, they're sealed, so while they definitely weren't the drivers as represented by the manufacturer, or measured by Josh, for my application where I'm not squeezing everything possible out of them, even if they're likely less efficient overall in the motor, and need more low frequency boost than predicted, once Chris P introduced and explained what a Linkwitz transform actually was, I applied at least a rough approximation and I've got to say that they sound pretty darn good, and in my living/kitchen area, the low frequency output capabilities are really more than adequate. I'm glad Rob Mosher talked me out of a pair of 12's. :)

View attachment 3119880

FWIW, this is free-air measurements of one driver after an accumulated 64 hours of high-excursion full-time infrasonic 'break-in'.

If you figure out how often people watch movies, and how much time in those movies actually has high output low frequency content. It will probably depend a lot on viewing habits and content, but if somebody watched 16 movies a month, which is more than every other day, and the movies on 'average' at a full five minutes (I'm pulling that number out of my butt, I've never done an extensive sample of movies holding a stop watch while watching my woofers) of balls to the wall high excursion output (not just a handful of thuds spread across a long time), that 'normal use' equivalent would be 4 years. If it takes more than this to 'break-in' to published parameters, I'm guessing that most users who aren't fanatic enthusiasts never actually get their drivers broken in. But, I think it more likely that they were just making stuff up or repeating something that somebody else told them because it sounded cool and accounted for the driver's measured differences, and since they don't make their drivers, if a whole batch is out of spec and they don't want to return them and be out of stock for a whole long while and lose money to other products, there's nothing else they can do about it anyway other than just to sell the defects as 'normal'.

HoursTotal MinMov/MnthMin/MovMin/MnthMonths
64​
3840​
16​
5​
80​
48​


I didn't have enough mass around that would stay in one place and stick to the cone to get a full set of measurements, but in terms of Fs, Qts, Qes, they told me enough.

Put in a 4 cubic foot box with 4 pounds, 7 ounces of polyfill and a light piece of material between the fill and the driver, it measured like this...

View attachment 3119881

Combined with an aggressive shelving filter, it got me under 20 Hz, in room, no problem, within a +/- 5 dB window, down to 16 Hz, and that's with an infrasonic filter implemented.

I love DSP. :p

I could have flattened it out more in that location, but I opted to just let it do what it was going to do beyond that. Splice response with main speakers from 10-180 Hz below before the playback curve was implemented.

View attachment 3119882
Most of that was a foreign language to me! But I do see at the end you got these subs to play nearly flat to 16hz in a 4cu.ft sealed box!? Is that mainly because of the DSP? With those results, it seems very doable for me to get a 17hz tune with an 8cu.ft ported box correct? (I do have a miniDSP ordered for my project)

From different port calculations Ive found online say I can achieve this tune with 3 ports 4" in diameter measuring 41" long. Should I be concerned with port air velocity/noise? I"m concerned because the manufacturers recommended port lengths for both 8cuft and 12cuft boxes don't exceed 32".

What about a single 6" port 38" long and tuned to 15hz? I just dont understand why the general consensus seems to be to build a massive, 20+ cu.ft box for these woofers.

I appreciate all the help. Very new to this and trying to do it correctly but the more I read the more overwhelmed I get!
 

·
Registered
Joined
·
316 Posts
Most of that was a foreign language to me! But I do see at the end you got these subs to play nearly flat to 16hz in a 4cu.ft sealed box!? Is that mainly because of the DSP?
I didn't measure it outside, so I think a combination of DSP and I just got lucky with the room response in terms of where the subwoofers happened to go, and where I listen when I care.

With those results, it seems very doable for me to get a 17hz tune with an 8cu.ft ported box correct? (I do have a miniDSP ordered for my project)

From different port calculations Ive found online say I can achieve this tune with 3 ports 4" in diameter measuring 41" long. Should I be concerned with port air velocity/noise?
I am really not the one to ask about the art of getting the most out of a reflex design. I like sealed systems in part because they're easy to not screw up. If you have roughly enough air behind them, or more, the driver can do what the driver can do, and if you need more than that, you add more drivers, or find a bigger/better one. Am I leaving output performance/driver on the table? Sure. No argument. If I have the enclosure volume to pull off a ported design with an UM-18, I can also just put in another driver into that larger box. But compared to many, I just don't play it that loud anyway, so it isn't necessary.

Reflex systems are balancing out a bunch of variables, and while I've spent a little time playing with modeling, I've also noticed that some situations are pretty stable, where changing variables doesn't change the response much and I've also had modeling experiences where you can get this beautiful extended uber flat prediction if all things are just right, but you change something and then you get a mess. Because loudspeakers change behavior depending on how you use them, and recover over time, you want a design that is not only going to balance out all of those variables (and based on the variables as they are, not the variables as they're supposed to be, which was kind of my original point), and also be applicable to your actual use. I.e., if you are going to optimize everything for playing LOUD for significant periods, maybe you want to have the design optimized to incorporate how the drivers behave when they've been cranked up for quite some time, and you don't want their behavior to change a whole lot between moderate background levels and cranked to the teeth.

Im concerned because the manufacturers recommended port lengths for both 8cuft and 12cuft boxes don't exceed 32".

I appreciate all the help. Very new to this and trying to do it correctly. The more I read the more over whelmed I get!
I'd hit up Chris Popovich or somebody similarly qualified for design expertise, as he seems to have not only a far more nuanced understanding of reflex systems, but a wide variety of real-world experience.

The best I can suggest is to measure the drivers you're actually using. If you follow generic recommendations, it might work out great, or fine, or at least work out well enough that if you don't have a direct comparison you'll never know differently. Or, it might not. Spending $130 on a measurement widget you think you'll probably only use once, on a relatively cheap woofer might seem kind of ridiculous, but I paid more than that shipping the woofers back for replacement for woofers that were almost exactly the same, but if I had bothered to measure them within the no hassle return period, unit would have paid for itself with shipping fees, and ultimately allowed me to make a decision to run them as they are (which isn't the worst thing ever) or just get a refund and try something else. If you're putting your time into it, consider what that's actually worth too. If you spend time modeling based on what the woofers are supposed to be, you build it all up, and then you happen to realize that the box needed to be a different size, or you needed a different tuning frequency, or whatever, was saving $130 worth it?

You can use my measurements, but I don't know if what you'll get will be like mine, or like the published specs, or something different. You can pay somebody to measure yours, but unless they're local, you'll probably spend more shipping than simply buying your own tool.

That's just my perspective. As was pointed out, lots of people build based on models using published expectations, or generic build parameters, and never notice a problem, and maybe given the level of DSP applied to dial in the room response the actual subwoofer response is secondary, and if you lose a few dB you'll never miss it, so for most people, in most situations, it really doesn't matter. that could be too.

That being said, it was that DATS device that provided some help that got the integration with the mains dialed in. Before I measured both the behavior of the subs and the mains, I was having trouble getting the combination spliced, and the issue seemed to be a qualitative problem with the subwoofers, and I hated them, like I wanted to fill them up with tannerite, modify their motors with a rifle round, and send the upgraded mess back to Dayton Ohio. I'm glad I didn't do that. They're actually pretty good considering the cost, even out of spec.
 

·
Registered
Joined
·
50 Posts
Discussion Starter #7
I didn't measure it outside, so I think a combination of DSP and I just got lucky with the room response in terms of where the subwoofers happened to go, and where I listen when I care.



I am really not the one to ask about the art of getting the most out of a reflex design. I like sealed systems in part because they're easy to not screw up. If you have roughly enough air behind them, or more, the driver can do what the driver can do, and if you need more than that, you add more drivers, or find a bigger/better one. Am I leaving output performance/driver on the table? Sure. No argument. If I have the enclosure volume to pull off a ported design with an UM-18, I can also just put in another driver into that larger box. But compared to many, I just don't play it that loud anyway, so it isn't necessary.

Reflex systems are balancing out a bunch of variables, and while I've spent a little time playing with modeling, I've also noticed that some situations are pretty stable, where changing variables doesn't change the response much and I've also had modeling experiences where you can get this beautiful extended uber flat prediction if all things are just right, but you change something and then you get a mess. Because loudspeakers change behavior depending on how you use them, and recover over time, you want a design that is not only going to balance out all of those variables (and based on the variables as they are, not the variables as they're supposed to be, which was kind of my original point), and also be applicable to your actual use. I.e., if you are going to optimize everything for playing LOUD for significant periods, maybe you want to have the design optimized to incorporate how the drivers behave when they've been cranked up for quite some time, and you don't want their behavior to change a whole lot between moderate background levels and cranked to the teeth.



I'd hit up Chris Popovich or somebody similarly qualified for design expertise, as he seems to have not only a far more nuanced understanding of reflex systems, but a wide variety of real-world experience.

The best I can suggest is to measure the drivers you're actually using. If you follow generic recommendations, it might work out great, or fine, or at least work out well enough that if you don't have a direct comparison you'll never know differently. Or, it might not. Spending $130 on a measurement widget you think you'll probably only use once, on a relatively cheap woofer might seem kind of ridiculous, but I paid more than that shipping the woofers back for replacement for woofers that were almost exactly the same, but if I had bothered to measure them within the no hassle return period, unit would have paid for itself with shipping fees, and ultimately allowed me to make a decision to run them as they are (which isn't the worst thing ever) or just get a refund and try something else. If you're putting your time into it, consider what that's actually worth too. If you spend time modeling based on what the woofers are supposed to be, you build it all up, and then you happen to realize that the box needed to be a different size, or you needed a different tuning frequency, or whatever, was saving $130 worth it?

You can use my measurements, but I don't know if what you'll get will be like mine, or like the published specs, or something different. You can pay somebody to measure yours, but unless they're local, you'll probably spend more shipping than simply buying your own tool.

That's just my perspective. As was pointed out, lots of people build based on models using published expectations, or generic build parameters, and never notice a problem, and maybe given the level of DSP applied to dial in the room response the actual subwoofer response is secondary, and if you lose a few dB you'll never miss it, so for most people, in most situations, it really doesn't matter. that could be too.

That being said, it was that DATS device that provided some help that got the integration with the mains dialed in. Before I measured both the behavior of the subs and the mains, I was having trouble getting the combination spliced, and the issue seemed to be a qualitative problem with the subwoofers, and I hated them, like I wanted to fill them up with tannerite, modify their motors with a rifle round, and send the upgraded mess back to Dayton Ohio. I'm glad I didn't do that. They're actually pretty good considering the cost, even out of spec.

I may just make another generic post asking for port recommendations to get the lowest possible tune with my enclosure limitations. I'll risk it with a generic recommendation and factory specs :LOL:. I just dont have the time, knowledge or money to do all this testing. haha. Fingers crossed i dont have an issue like you had!
 

·
Registered
Joined
·
4,303 Posts
Why wouldn't you ask your question here?
 

·
Registered
Joined
·
50 Posts
Discussion Starter #9
Why wouldn't you ask your question here?
I feel like I have but I'm not getting an answer. Though maybe if I title my post differently ide get an answer. But here it is:

I'm building a Sonosub that after subtracting speaker displacement and approx port area will be 8 cu.ft net volume.

What is the lowest possible tune I can get, and port size(s) with an 8 cu³ box.

Driver will be a Dayton Audio UM18-22 in a 20" sonotube approximately 48" tall. (52" with the feet). I would build bigger but can't go wider due to space constraints and cant go taller due to WAF.

Would a single 6" port 38" long tuned to 15hz work?

Thanks.
 

·
Registered
Joined
·
4,303 Posts
I feel like I have but I'm not getting an answer. Though maybe if I title my post differently ide get an answer. But here it is:

I'm building a Sonosub that after subtracting speaker displacement and approx port area will be 8 cu.ft net volume.

What is the lowest possible tune I can get, and port size(s) with an 8 cu³ box.

Driver will be a Dayton Audio UM18-22 in a 20" sonotube approximately 48" tall. (52" with the feet). I would build bigger but can't go wider due to space constraints and cant go taller due to WAF.

Would a single 6" port 38" long tuned to 15hz work?

Thanks.
Yes it will work. People build huge ported enclosures for the UM18-22 in order to maximize the low end. Also, you may want to download and familiarize yourself with Winisd.
 

·
Registered
Joined
·
4,657 Posts
I feel like I have but I'm not getting an answer......

As johnson mentioned, download WinISD and load the specs for the UM-18. It will tell you everything you need to know. You can tune it to anything you want. However, port size and velocity will become a problem at some point.
 

·
Registered
Joined
·
50 Posts
Discussion Starter #14
How much power and what high pass filter?

Edit: with 1200 watts and a 13hz 4th order high pass, your port velocity will be 28.6 m/s. That's too high. It will chuff.
2 of them will be powered by a NX6000 with a miniDSP. When you say "4th order highpass" makes me realise I have lots to learn cuz I have no idea what that is! :LOL:

Can you suggest a highpass filter setting? In the mean time I'll do some more research as to what that is.

Its sounding like in order for me to get a low tune ill need to increase enclosure size? This may take some serious convincing to my wife. I will play around with WinISD I'm just really limited with time right now from work and baby at home. No time to use the computer.
 

·
Registered
Joined
·
4,303 Posts
So how will you find time to build two enclosures :unsure:
 

·
Registered
Joined
·
50 Posts
Discussion Starter #16
So how will you find time to build two enclosures :unsure:
No need to be condescending, that doesn't help anybody....

Right now I'm busy. Fortunately my current job gives me tones of free time at work to research, I just can't download WinISD on the work computer.

As for building, I'm a journeyman carpenter, mostly in the cabinet/millwork department, so I have the tools and skills to build them.

Right now I'm trying to come up with enclosure plans, so when I have some more free time at home, I can build them. And by then, hopefully I'll be well read on the tuning process.

It's simply using the rendering apps right now to come up with plans is what's holding me back.
 

·
Registered
Joined
·
217 Posts
You should be able to use 3ea., 4" psp flared ports without any chuffing issues; but for a 18hz port tuning they need to be 37.25" long each. They each will take up a little more than 1/4 cu.ft. if air space each. If you mount the ports so the inside port flare's are close to inside Sono tube wall the actual port tune will be lower than 18 hz.
 

·
Registered
Joined
·
50 Posts
Discussion Starter #18
You should be able to use 3ea., 4" psp flared ports without any chuffing issues; but for a 18hz port tuning they need to be 37.25" long each. They each will take up a little more than 1/4 cu.ft. if air space each. If you mount the ports so the inside port flare's are close to inside Sono tube wall the actual port tune will be lower than 18 hz.
Thanks! had no Idea the position inside the tube made a difference. The plan is to have a 20" sonotube 48"-50" tall. So im getting a gross volume of approx 8.75 - 9 cu.ft. So after subtracting port and driver should be very close to 8 cu.ft. With the height I should be able to just fit a 37" port in. No issues with it being within a couple inches of the driver?

Something else I dont fully understand is how my SVS PC12+ ihas a 16hz tune option yet the 3 ports are only 3" and cant be more then 30". Wouldnt this produce a tremendous amount of chuffing? or perhaps these stats are just skewed, and or Ive never fully pushed it to hear the port noise.
 

·
Registered
Joined
·
316 Posts
I am not certain, but I believe that the lower tuning options involved with the SVS subwoofers involve plugging ports. I.e., the lowest tune option has 2 out of 3 plugged, highest tune has all open.
 

·
Registered
Joined
·
4,303 Posts
No need to be condescending, that doesn't help anybody....
Not being condescending at all. Pointing out that "Winisd" is a very useful and important tool in DIY "speaker and subwoofer" design/building is indeed helpful to everyone;)
I look forward to seeing your progress.
 
1 - 20 of 32 Posts
Top