Designing Dual 18" vented boxes: ED19O.2 vs. SS RL-p18

[nibix kfrundi]

Having read a ton for several weeks and learned a lot, I am desiring more directed knowledge.

I am building a 3300 cu foot (22' x 15' x 10') dedicated theater room, starting with the audio. I am planning my subs and wanted it loud and deep (who doesn't???).

I have also finished reading the excellent book by Floyd Toole (

Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms

Amazon

) and in the interest of good bass for as many as possible, have elected to position two subwoofer enclosures in the middle of the front and the middle of the back wall.

This places some constraints on my box size, notably the depth, which I want to keep at 18-20". I am planning to place 2 drivers in each vented enclosure and using Box Notes have arrived at a box that is 42" tall, 66" wide, and 20" deep with a working volume of 21.7 cu feet (615 liters). Here are the gory details (by the way the port is 8" diameter; the cursor makes it look like 18"):



So, I am getting to the heart of my question here. I have modeled this box tuned to 15 Hz with two ED 19Ov.2 and two SoundSplinter RL-p18 drivers using WinISD alpha.

Here are the SPL graphs without EQ assuming 125 W of system input power. The 19Ov.2 is in blue and the RL-p18 is in red:



Not exactly the flat response from 15-80 Hz that I am looking for.

If I apply parametric Eq (fc=15, Q=1, gain=4.5dB) to each system, then the curves look like this:



I am struck by the flatter curve of the RL-p18, but slightly deeper extension of the 19Ov.2, but let's face it, at this point I am over my head and really don't know too much about what I am doing. Can I flatten the 19Ov.2 curve with a lowpass filter or ??? I am hoping that some of you smart folks can help me optimize the situation and solidify my build plans.

Do note that the 190v.2 costs $245 shipped, while the RL-p18 is $440 shipped. I am cost conscious, so if I can get the 190v.2 to model flatter, then I will be convinced to go that direction, recognizing that this modeling gets you only so far what with the marked room influences that I see in some of the in room FR analyses that some of you all are posting...

Thanks in advance for your advice and engagement in my project and apologies for the loooooong post,

Geoff

[Mike P.]

Be aware that the RL-p18 is currently unavailable and is estimated to be in stock 1st quarter 2010. Your SPL graph of the 19Ov.2 shows a 4 ohm load for both subs. What amp are you using?

[nibix kfrundi]

Mike P. wrote:

the RL-p18 is currently unavailable and is estimated to be in stock 1st quarter 2010.

I knew they were out of stock, but I didn't know that their availability would be that delayed. Those guys at SoundSplinter have some serious supply chain issues. This may make the issue academic, but I'd like to understand the nature of the comparison anyway.

Mike P. wrote:

Your SPL graph of the 19Ov.2 shows a 4 ohm load for both subs. What amp are you using?

I was planning to use the Behringer EP2500, but I am not committed to anything at the moment. Where does the curve indicate that I am showing a 4 ohm load to both subs?

Actually, I see that I left the "series resistance" field at 0.01 ohms, which is probably not correct. When considering two drivers with dual voice coils, what would be the recommended wiring for the two driver configuration with an EP2500 and what should I enter into this field into the "series resistance" field in WinISD?

[Mike P.]

Where does the curve indicate that I am showing a 4 ohm load to both subs?

Your driver file must have a Re value listed as 3.5 ohms which would produce the SPL curve with 125 watts. That would be the D2 version of the sub wired in series. Do you have the D2 or the D4 sub?

As for the "series resistance" in the Signal tab , that is not for the final ohm load of the subs. see "Designing your box" in the help files.

[nibix kfrundi]

Thanks for your help. Mike. I do not own drivers yet. I suspect that I would get the D4 version, which as I understand allows me to create either a 1 ohm or a 4 ohm load depending upon parallel or series wiring of the voice coils, respectively. Wiring the voice coils of each speaker in series and then the two speakers in parallel gets me 4 ohms, correct? I think that will give me the best flexibility for amplification.

In WinISD, Re is set at 3.5. I used the values from the (ED Website) to derive this value. I do note that the Miscellaneous Parameter section of WinISP shows 2 voice coils in parallel with Re=3.5, when I select series for the two voice coils, then Re is increased to 14 ohm automatically. Also Bl reduces to 11.05. Are these values correct for series wiring of the voice coils or should I re-adjust B1 to 22.1?

How do I then adjust the parameters to accommodate for the two drivers wired in series or does the program assume this with the num of driver field indicating two drivers? I so not see where to tell the program that two drivers are wired in series or parallel.

Finally, I am better off un-checking the auto calculate feature and entering all of the discrete values from the ED specs or should I let WinISD adjust the values?

Apologies in advance fir extreme newbie questions.

[Mike P.]

Are these values correct for series wiring of the voice coils

Yes, changing the final load also changes the BL.

How do I then adjust the parameters to accommodate for the two drivers wired in series or does the program assume this with the num of driver field indicating two drivers?

You figure out what the final load is and change the Re in the driver file. When you use the "auto calculate unknowns" feature the program automatically adjusts the BL.

Apologies in advance for extreme newbie questions.

No apologies necessary, we'll make a "Sub Wizard" out of you yet!

Regarding your sub design, will you be using one EP2500 for both cabinets or a EP2500 for each cabinet?

[nibix kfrundi]

Mike P. wrote:

we'll make a "Sub Wizard" out of you yet!

I am looking forward to that!!!

So, here is where I am at:

Baseline data from Elemental Designs web site on the 19Ov.2 driver:



Entering Re and BL into WinISD provides the following:



Changing parallel selection for the dual voice coils to series gives the following Re and BL values, automatically:



If I attempt to halve the Re value to accommodate for two drivers (each with 4 Ohm voice coils wired in series), the BL does not automatically update. So, I uncheck the calculate automatically selection, delete the entered value of BL and allow the automatic calculation of BL:



As one can see the BL is now slightly different. Now I divide Re by 2 to accommodate for the parallel wiring of the two drivers (each with 4 Ohm voice coils wired in series). The resulting calculation gives:



With these driver parameters entered please note the box, port, and filter parameters indicated on the following three plots of SPL, port air velocity, and group delay:







Do I have all of this correct?

Do my ports seem reasonably sized?
I was planning to run them out of the right and left sides of the 66" wide box. I am a bit over the 17 m/s threshold at 13 Hz (calculated velocity = 19 m/s) that may cause chuffing, but any increases in port diameter or number, while reducing air velocity, are associated with an unacceptable increase in port length. I am rationalizing this velocity on the assumption that little source material will have much information at 13 Hz and with two of these bad boys in the room, I am unlikely to get near Xmax.

Are the filter parameters reasonably achievable?

Will this sub kick butt, ie can I expect the observed flat response from sub 15 to 80 Hz???

[nibix kfrundi]

Mike P. wrote:

Regarding your sub design, will you be using one EP2500 for both cabinets or a EP2500 for each cabinet?

I almost forgot about this. I am thinking that I will have one EP2500 for each cab. Does this seem to be the most logical plan rather than use one amp for both cabinets?

[dradius]

you'll need two ep2500s, one for each cab. if you're going ported you want the d2 so the behringer will give them ~650w each. if you go with the d4 they'll get like 1,000w each which i think is too much for them ported.

[nibix kfrundi]

dradius wrote:

you'll need two ep2500s, one for each cab. if you're going ported you want the d2 so the behringer will give them ~650w each. if you go with the d4 they'll get like 1,000w each which i think is too much for them ported.

To confirm, with the D2, my cabs will be 2 ohm. With the D4 my cabs will be 4 ohm. I'll have 650 W into the 2 ohm cabs and 1000 W into a 4 ohm cab?
Posted via Mobile Device

[Mike P.]

A couple of thoughts, just my opinion. You want loud and deep but your modeling does not reflect the subs and amps capabilities. With a pair of the subs in 21.7 cu.ft tuned to 15 hz and with 4.5 db of boost applied, the subs will handle 650 watts each. This assumes the D2 sub wired for 4 ohms and one sub to each channel of the amp. With a Hi-Pass filter applied at 16 hz the subs never exceed Xmax.

Which brings us to porting. There is no issue with an air speed of 26 m/s, many a sub has been designed and built with this speed with no problems. Two six inch ports would have an air speed of 36 m/s, too high. A 10 inch diameter port 35.75 inches long would be just under 26 m/s. You could expect 124 db down to 15 hz with one cabinet. Now that's loud and deep! Should you go this route, I suggest you build one cabinet first, it may well be enough for your needs.

[nibix kfrundi]

Wow Mike. I am clearly missing something in my WinISD file of this build. As you can see from my prior posting, my specs are completely different from what you have determined. I could learn a lot by examining the file that you have made for this driver and this proposed cabinet. Is it possible for you to send them to me so that I can open them and examine them on my computer?

Thanks,

Geoff

[nibix kfrundi]

After thinking a bit about it, I understand the principle of keeping the two drivers independent in the cabinet and driving them from each of the channels in the EP2500. So with the D2 wired in series, I am at 4 Ohm and then driving each driver with 650 W. This is exactly what dradius had in mind. I would still like to examine the driver and project files for this build on WinISD, as I cannot generate the same graph that Mike sent in his last post, but at least a bit of the mystery is receding. Again, many thanks.

[Mike P.]

The file is a non auto calculate version with the parameters you listed from the E.D. website. Note that the Re stays at 3.5 ohms since each sub has its own channel at 4 ohms and the input power for both is 1300 watts. here is the file:

Elemental Designs 19Ov.2 Non AC.wdr

Here is the project file:

E.D.19O.2 - 21.7 Cu.Ft..wpr

[SteveCallas]

nibix wrote:

I have also finished reading the excellent book by Floyd Toole (
Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms
Amazon ) and in the interest of good bass for as many as possible, have elected to position two subwoofer enclosures in the middle of the front and the middle of the back wall.

Don't assume any particular placement will be the best in your room or you may be setting yourself up for failure. Every room is different and you really have to try it out and take a few measurements to be sure.

Also, while some manufacturers may do it, the last thing I would ever want to do is apply EQ boosting around the tuning region of a ported sub. Excursion is already at its greatest just above tuning, impedance is at its lowest at tuning (and cone movement is next to nothing, so the voice coil is getting pretty hot), and then the driver becomes susceptible to unloading below tuning. Not really the greatest place to apply some boosting.


Mike is leading you on the right path - get increased low end output through an increase in sensitivity. You do this by using a larger enclosure. The larger enclosure also allows you to use more adequate porting.

[nibix kfrundi]

Mike,

Those are very helpful. I have a few questions scattered through my ramblings below:

Why do you only specify one voice coil in the driver file?

I am glad that you have confirmed that the system input power on the signal tab is the power into the speaker from the amp. I found the following statement from the help file on this subject to be very confusing:

Term "power" should more correctly be voltage. This term "power" comes from definition by Richard Small, who defined the input power to be P=Eg²/Re (this is to simplify things, because impedance varies greatly depending on the frequency and the box), where Eg is RMS output voltage of your amplifier, and Re is DC resistance of voice coil. Actually, driver's impedance will vary greatly (see impedance graph), depending on the frequency, so this is generally not the power which is taken from the amplifier.

There are also absolutes in the help file that are good to know can be exceeded, such as:

Very important here is to make your ports large enough, so that peak air velocity won't exceed 5% of sound velocity in air. That is, about 17 m/s assuming normal environmental conditions. You can check it by using "Rear port - Air velocity" graph, by first setting simulation power to anticipated maximum power, see "signal"-tab. In this case, Pe is 125 Watts.

I wonder how many people base their understanding of WinISD and this subwoofer design primarily on this help file? It certainly lead me down some wrong pathways. I just found the excellent tutorial by Andrew Vaserfirer here. His recommended upper limit of air velocity of 20-25 m/s is closer to our value of 26 m/s.

Not specifically in the help file, but accessed from here after a google search states:

A high groupdelay will result in a boomy/muddy sound. The groupdelay is created by the basreflexport and is thus one of the downsides of using basreflex cabs.
Most people however are used to this sound since basrelfex is used widespread in daily live.
Not only the highed of the groupdelay is to be considerd. The shape is also important. It's better to have a slowly forming round peak then a sudden needle-like peak that's there for just a few frequencies..Generally the groupdelay will reach it's maximum around the tuning frequency

A rule of thumb coming from Hifi is: frequency x groupdelay = 400 (max). Less being preferable. At 40 Hz that would be 10 miliseconds. A little bit higher (up to 600) around the tuning frequency can be overcome.
From this rule of thumb you can clearly see that the highed of the groupdelay becomes more important at higher frequencies and less important at lower frequencies. This is because the sensitivity of your ears increases as the frequency rises.
A high groupdelay in the 80 Hz and up area for instance will ruine the "kick'. Making it sloppy having less impact. Making the volume bigger or tuning the port lower, will increase the groupdelay. As you lower the tuning frequency the groupdelay will shift down along side with it.

Andrew doesn't talk about group delay in his tutorial. The group delay of the cabinet design that you attached previously reaches 80 ms at 12 Hz. This is partly influenced by the high pass filter, which is certainly important for keeping cone excursion under Xmax. Perhaps at this frequency the group delay is barely audible and will not muddy the sound. How do you look at group delay and manage it for a "tight" sounding bass response?

With the parameters that you have worked out for the cabinet and driver, it is interesting to see how little difference appears between a sealed design and a ported design. For a sealed design, a slightly smaller box (19.6 cu ft) reduces the cone excursion to be similar to the vented design. Everything else equal the group delay falls to 40 ms at 12 Hz and the difference in SPL is a maximum of 6 dB at 16 Hz and only 2 dB at 30 Hz. I applied a 4th order low pass filter at 60 Hz to both designs as well.

They look like this:



What do you think is the audible nature of this difference?

Again, many thanks for all of the insights.

[nibix kfrundi]

Steve, I have read many of your posts and learned a lot from them. Thanks for weighing in.

SteveCallas wrote:

Don't assume any particular placement will be the best in your room or you may be setting yourself up for failure. Every room is different and you really have to try it out and take a few measurements to be sure.

I do understand that the room will have a substantial influence. I will try to maintain flexibility of placement. What do you think about having two cabinets versus one. It seems like two offers more flexibility in tuning to the room, and it should enable a more even bass distribution across the listeners.

Also, while some manufacturers may do it, the last thing I would ever want to do is apply EQ boosting around the tuning region of a ported sub. Excursion is already at its greatest just above tuning, impedance is at its lowest at tuning (and cone movement is next to nothing, so the voice coil is getting pretty hot), and then the driver becomes susceptible to unloading below tuning. Not really the greatest place to apply some boosting.

In Mike's design, the cone excursion maxs (22 mm) at 11 Hz, has a local nadir (6 mm) at 15 Hz and then climbs to 18 mm at 22 Hz before falling at higher frequencies. Impedance is 4 Ohm at 15 Hz. I'm afraid I don't know what "unloading" is, but it doesn't sound good. However our flat FR curve doesn't look nearly so flat without the 4.5 dB parametric boost. I hate to give that up. Now with a 40 cu foot cabinet, I regain all of the low end boost and my port shortens up nicely.

Let's see, 40 cu ft would be 72" x 60" x 20". That might stretch the limits of practicality. Is there a blend of cabinet volume and amplifier boost that makes you more comfortable near the tuning frequency?

Incidentally, the impedance and the cone excursion are the same when comparing the larger cab with no eq to the smaller cab with eq boost. Am I still risking an unloading event with the larger cab?

[SteveCallas]

What do you think about having two cabinets versus one. It seems like two offers more flexibility in tuning to the room, and it should enable a more even bass distribution across the listeners.

Absolutely, I'm all for multiple subwoofers.

How large an enclosure can you live with? How about 400 liters?

[Mike P.]

Why do you only specify one voice coil in the driver file?

The file uses the parameters for 2 coils wired in parallel, I just didn't enter that it had 2 coils. It doesn't affect the modeling.

Group delay in speaker design is one thing, in subwoofers it's another story. It's not something that you would notice.

I'm not able to see the graph you posted above due to a bad internet connection.

As for boosting the low end, with a Hi-Pass filter you're within Xmax, cone excursion and impedance doesn't drop below 4 ohms. I don't see this being a issue with program material. Just don't play a 15 hz sine wave with 1300 watts and your sub will be fine.

[nibix kfrundi]

SteveCallas wrote:

Absolutely, I'm all for multiple subwoofers.

How large an enclosure can you live with? How about 400 liters?

I should be well over that. The 21.7 cu ft design that we have been kicking around is 615 liters and the monster 40 cu ft design that I mentioned in my last post would be 1133 liters.

[dradius]

sorry for the delayed response, it sounds like you figured it out on your own though yeah i'd give each sub its own channel from the behringer. and just to reassure you, i have the same drivers and give them 650w each ported and they are ridiculous. 4 separate boxes would allow you to put one in each corner etc. but of course it'd be more wood and work. either way i think you'll be happy. 4 of these will be awesome either way.

[nibix kfrundi]

Hey Dradius, I see that you built sonotubes. I was thinking of doing that, but rejected it so I could have a lower profile design from a depth perspective. The downside of sonotubes is that their circular cross-section means that they are equally deep and wide. I am curious about your consideration to cut your tubes down and go sealed. I recently noted that mktheater did the same thing.

I do not want to go through the exercise of building vented cabinets only to decide that I want sealed. While Mike's help, I have modeled the vented and sealed cabs (see post #16 above), I have no intuition as to how they would perform/sound. Please discuss your thoughts and experience on this subject.

[dradius]

i'm far from an expert on this stuff, so take everything i say with a grain of salt
my sonos aren't equally deep and wide, they are 24" wide and 6' long. if i cut them down and sealed them, then yeah they'd be 24x24. i talked to eD and determined the d2 version is the way to go if you want ported, since 650w is about right for them to shine. however, they need much more power if you're going sealed (~1000w.) so you'd need the d4 version, which is what mktheater has done. so i'm definitely not going to seal them. the only downside to mine that i was worried about was bottoming them out, but they are tuned pretty low and i also learned my receiver (and most mid-level ones i think) won't even send a super low signal. i've had them running pretty hard and never had any issues though.
i'm really not the guy to talk to as far as sealed vs. ported, i just don't know enough about it. i thought about sealing them because i was considering putting one in each corner and didn't want four 6' monsters in the room not because i prefer sealed to ported.
i'd definitely determine which way you want to go so you can buy the d2 or d4 version accordingly. i'd also consider just 2 drivers and one amp to start. it might be plenty for you. and last but not least, when i bought my drivers they were $50 cheaper than they are now. i really like them, but am annoyed they keep raising prices on all their stuff.

[nibix kfrundi]

dradius wrote:

my sonos aren't equally deep and wide, they are 24" wide and 6' long.

If they are standing upright, then these are 24" deep and 24" wide - that is what I meant in my comments. I am looking for 18" deep with a big cab and the possibility of mounting an 18" driver, this eliminates the sonotube concept and leads to a traditional box.

i talked to eD and determined the d2 version is the way to go if you want ported, since 650w is about right for them to shine. however, they need much more power if you're going sealed (~1000w.) so you'd need the d4 version, which is what mktheater has done. so i'm definitely not going to seal them. the only downside to mine that i was worried about was bottoming them out, but they are tuned pretty low and i also learned my receiver (and most mid-level ones i think) won't even send a super low signal. i've had them running pretty hard and never had any issues though.

Thanks. That is very helpful

i'd also consider just 2 drivers and one amp to start. it might be plenty for you.

Mike suggested that as well. I will give it some serious consideration- saving costs versus efficiency of building two at once. It is easy to get a skewed sense of what is desirable when reading all of the giant "overkill" builds in the forum.

last but not least, when i bought my drivers they were $50 cheaper than they are now. i really like them, but am annoyed they keep raising prices on all their stuff.

I can be philosophical about that. A 20% increase is pretty high, but we are in a tough economy and as best as I can tell, they are offering a great product at a price that when considering the free shipping is still more reasonable than the competition.

Look at Soundsplinter and the RL-p18. How screwed up is that company that they cannot deliver a core product for at least 4 months???

[Ricci]

My 2 cents here. The RLP-18 is a better driver but it's also a lot more expensive and unavailable. Scratch it from the list. Each enclosure at 2 ED 190v2's per cab should get an EP2500. I'd get the D4 drivers and wire each cab to 4ohm mono to get the most power from the amp. Those drivers have quite a bit left after 22mm xmax is met and won't bottom out until after 30mm one way. Also with thermal losses and suspension tightening it'll take more power than the sims indicate to bottom the out.

The other idea I have is the Mal-X 21" driver that just came out. Have you thought about using one of these per cab instead of 2 18's? It'd be a bit more expensive than 2 190v2's, but it should be impressive in a big ported cab.