Can a marginal sub ruin a good soundstage?

[snmhanson]

I have been thinking lately about the sub I am running with my Statement Monitors. It is a DIY sub consisting of a 3cf sealed box, a Dayton 15" Titanic driver and a 500 watt Bash amp. It recently stuffed it with polyfill to try to tighten the bass up a bit. I am also thinking of building a ported 12" to cover the deep bass when viewing HT. All of this work on the subs got me thinking about how a sub could potentially affect a superb speaker like the Statements. Is it possible that a sub that is not on the same level as the front speakers could detract from the listening enjoyment? Or can your ears and brain differentiate between the two? How about surrounds? I am going to try to build some surrounds to go with the Monitors but they are going to be a different, and likely compromised design. Do I run the risk of degrading the quality sound that comes from the front sound stage?

Thanks,

Matt

[ThomasW]

Highly unlikely that a good sub with proper location, proper XO point and EQ would do anything but make the mains sound better.

Let's see a frequency response plot of the sub....

[madmac]

I have been thinking lately about the sub I am running with my Statement Monitors. It is a DIY sub consisting of a 3cf sealed box, a Dayton 15" Titanic driver and a 500 watt Bash amp. It recently stuffed it with polyfill to try to tighten the bass up a bit. I am also thinking of building a ported 12" to cover the deep bass when viewing HT. All of this work on the subs got me thinking about how a sub could potentially affect a superb speaker like the Statements. Is it possible that a sub that is not on the same level as the front speakers could detract from the listening enjoyment? Or can your ears and brain differentiate between the two? How about surrounds? I am going to try to build some surrounds to go with the Monitors but they are going to be a different, and likely compromised design. Do I run the risk of degrading the quality sound that comes from the front sound stage?

Thanks,

Matt

My take is that a sub cannot really hurt the overall sound that much because it functions at such a limited frequency range. However, if it's a bad sub running bad drivers with bad everything.......including the amp that drives it, then I guess it could somewhat worsen the sound somewhat. Your surrounds, center and mains should ideally be the same brand, year, model, and manufacturer however for ideal HT and music playback quality. :W

[Bear]

Too high of a crossover point would be the main way that a sub would hurt the soundstage. As Thomas says, though, if you get the slopes, XO frequency, etc. right, then it should blend relatively seamlessly into your mains.

[Paul W]

IIRC, the old 12" Titanic had a very pronounced breakup around 300-400 Hz. If crossed too high, or with a low-order slope, it would mess up system midrange.

Don't know about the newer 15", but you might want to run sub-only nearfield measurements to see if there are any FR spikes above crossover.

[snmhanson]

Thanks for all of the replies. I am fairly new to DIY so I don't have the measuring and testing equipment that I need yet. I plan on ordering a mic soon though, so I can start taking measurements. In any case, I don't necessarily hear any issues with my system and the sub seems to blend in fairly well with my speakers. However, I definitely don't have a golden ear and probably wouldn't notice anything unless it was a ways out of whack. I posed the question more out of curiosity than anything, and because I would like to build another sub to pair with the system and want to be sure I am not robbing Paul to pay Peter so to speak.

Matt

[brent_s]

Matt, do you mean the BASH500 as in #300-752? If so, the stock HPF of that amp is around 30-31Hz with a 12dB/octave rolloff below that. There's also a 2.5ish dB peak up around 36-37Hz. The spec sheet identifies the two resistors that need to be changed to shift the cutoff frequency and boost along with a table of suggest values. You can model the HPF in WinISD Pro by adding a high pass/SOS filter to a model of your sub...enter the Fc and Q from the P-E tables to see how it changes the frequency response.

-Brent

[snmhanson]

Thanks for the reply Brent, though you have kind of opened a can of worms for me now. I have never worked with filters in WinISD but I can see how they can significantly change the response slope. I see where to add a HPF, but then it asks for the subtype and three variables - order, cutoff and Q. Cutoff is Fc I believe and Q is a given, but what do I use for order (2 seems to be the default)? And what subtype should I use? I also don't see the Fc and Q figures to use for the original HPF setup on the Bash 500w amp (with the 10K and 60K resistors) - any ideas what figures are?

Thanks,

Matt

[brent_s]

Thanks for the reply Brent, though you have kind of opened a can of worms for me now. I have never worked with filters in WinISD but I can see how they can significantly change the response slope. I see where to add a HPF, but then it asks for the subtype and three variables - order, cutoff and Q. Cutoff is Fc I believe and Q is a given, but what do I use for order (2 seems to be the default)? And what subtype should I use? I also don't see the Fc and Q figures to use for the original HPF setup on the Bash 500w amp (with the 10K and 60K resistors) - any ideas what figures are?

Thanks,

Matt

Subtype is "SOS, user specified Fc and Q".

Order is fixed at 2nd order (12dB/octave) even though the dialog box lets you change it. I personally think SOS is supposed to mean Second Order Sallen as in Sallen-Key, which is the type of HPF circuit most plate amps implement. S-Ks can implement other slopes (3rd/4th/5th/etc. order), but WinISD doesn't seem to support them at this time, although it looks like the intent is there.

Use the values for 10/56 from P-E's chart for the stock setup. Given part tolerances, you're probably talking tenths of a dB difference in the real world.

-Brent

[snmhanson]

Thanks Brent. I am having fun now. I have experimented with several HPFs and have been able to get down to the low 20s at -3db, which is quite a bit lower than it is was initially modelling at. I am pretty flat except I can't seem to get rid of a ~1bd hump before it starts rolling off though. Is that typical? Also, I don't have a LPF in WinISD. Do I need to model one to get an accurate curve or do I just set the amp to where I want it to cross over? Right now it is flat above the curve to the edge of the chart. I just assumed I would set the crossover on the amp to where I want it.

Thanks again,

Matt

[BOBinGA]

You should most certainly model the low pass filter. In fact, this can, in a roundabout way, get you a few more hertz in low frequency response at the slight expense of a db or two of efficiency. Try lowering the tuning by 3 or 4 hz and then apply a low pass at 60 hz. The lower tuning will put a small bump in the very low frequency response below the cutoff frequency, but when you add the LPF it will even out the response. Play with the low end tuning until you get it as close to flat as you can, but don't worry too much about a 1 db hump since that will be much less than what the room will induce.

-Bob

[snmhanson]

Thanks Bob. I can't find the parameters to use to model the LPF for the Bash amp. I assume I use my crossover setting on the sub (80Hz) for Fc but I'm not sure what to use for Q. I am learning a lot here, but getting a bit lost at the same time.

What would really help me, if anyone feels motivated and has the time, is if someone could model my sub on WinISD and send me the file. It is a ~3.25 sf sealed box with a Titanic 15" MKIII and (stock) Bash 500 watt amp. I am wondering if the model that I used is giving me inaccurate information, and how low and flat I can make the sub by changing the HPFs. I think that if I could see it properly modeled it would help me understand everything much better. Anyone interested?

Thanks,

Matt

[BOBinGA]

Matt,
I did a thread many months ago about how to model in WinISD. Brent even contributed quite a few good ideas in it. I'll see if I can dig it up when I get home from work tonight. In the meantime, for the high pass filter in WinISD, I usually just use the "peaking second order" and all you need to put in is the number of db that it peaks and the peak frequency and use the default values for everything else.

-Bob

[BOBinGA]

Here's the thread I was thinking about:

https://www.htguide.com/forum/showthread.php?t=36210

I'll try to run a model of the Titanic and Bash amp tonight and see if it comes close to what you were thinking.

-Bob

[snmhanson]

Thanks again for all of the help Bob. I think I could be getting somewhat close to maybe kind of understanding what I am doing in a way. I've got the stock amp modeled right now so it peaks at ~110db at 50Hz and 125 watts signal. I am getting -3db at 32.5Hz and 88Hz but it rolls off quickly after that (-18db at 20Hz). If I use 5db of bass boost at 20-25Hz per the chart, I can get to 20Hz at only -7db and my -3db moves down to right around 30Hz and only give up 1db at the peak. Does this all make sense and sound about right? I would still love to see someone else model this and compare what I came up with - but I know time is valuable.

Thanks,

Matt

[BOBinGA]

Matt,
Here is the Titanic 15 in a 4.0 cu.ft. box. (Sorry, but that's about as small as I could make it even fully stuffed.) The yellow trace is a closed box. The gray and green traces are ported boxes tuned the same to 20 hz. The gray trace is with the PE Bash 500 and the green trace is the with an O Audio 500 watt BASH amp with a 6 db boost at 16 hz:



The PE BASH has a 2.5 db peak at about 36 hz. To me, this doesn't really help the ported box work any better than a closed box. You can get to 30 hz easily, but it rolls off pretty quickly below that.

For your amp, I would recommend the O Audio 500 watt BASH. http://www.oaudio.com/500W_SUBAMP.html
It costs less than the PE amp and it is much more flexible. The extra control is not just marketing in its case. You can see the green trace's vast improvement in performance below 30 hz by using the O Audio adjustable/tunable HPF. Plus it has a parametric EQ circuit that you can use to tame the inevitable peaks caused by your room.

From the picture, you can see how I model the HPF and LPF. All the graphs use the same 50 watt input. This equates to about 200 watts at the 16 hz boost frequency, but the amp can handle that. On the Box tab, under the Advanced link I set Qa to 5 for the closed box (tightly stuffed) and to 10 for the ported boxes (moderately stuffed). I had to add stuffing to get the box size down to 4 cu.ft. Using the O Audio amp, you might not need a separate sub for very low frequencies like you mentioned in your first post.

Put these filters in your model and play with the box size and tuning. You might be able to fine tune it even further. Also watch how it affects cone excursion. I hope this helps.

-Bob

[snmhanson]

Cools - thanks Bob. Looks fairly close to what I came up with, except my curve is shifted slightly to the right - due to my smaller cabinet I assume. I appreciate the time and effort you've put in to helping me. I've still got some things to learn but I think I at least have a better idea of what I need to focus on.

I should clarify that I am modeling this sub after the fact, as I have already build it and just left the Bash amp in stock form. So this is an exercise in trying to accurately model my current sub and do determine how I can tweak it's response curve to fit my needs. I do plan on building a second sub though with the same 3.25 cf box except ported and using a 12" driver, so this new knowledge will come in very useful there as well.

Thanks again,

Matt

[mackintire]

There ARE some boxes that have some acoustical voodoo:

Exhibit A: https://www.htguide.com/forum/showthread.php?t=37249

The tune on the mini tunes to 16hz with proper stuffing, the standard version and the max both tune to around 13hz.

The acoustical line "measured", with a mic, always calculates longer than the physical line for the port (model). I've never figured out why....

From talking to the creator Neo Dan: the acoustical vent is near 30" longer than the expected modeled port response.

Mackintire

[BOBinGA]

Matt,
Sorry I missed the fact that this was an existing box. My ADD impulsivness got the best of me again. Oh, well.

Nonetheless, I think you are getting the idea. Since you are looking at using the same 3.25 cu.ft. box for a 12" sub, Here is something modeling I put together because I was considering pretty much the same thing using a PE RSS315HF-4 and an O Audio amp. The Shiva is a nice woofer too and will obviously make a great sub.

Here is the PE sub in a 3.0 cu.ft. ported box tuned to 16.5 with a 4"X33" vent at 150 watts with no bass boost. This is close to a 3.25 cu.ft. box with the vent entirely inside. Its not too pretty (yet):



It also unloads and exceeds XMax by a substantial margin below the tuning frequency.



Now, let's do a little voodoo of our own and make it pretty. We add in a 4 db boost at 20 hz with the O Audio amp HPF. (You can easily do the same with the PE BASH amp with a resistor change.):



And, here's the magic, it no longer exceeds XMax:



Unloading below tuning is one of the biggest problems with ported subs and this is an elegant way around it. Use a lower than normal tuning and then add in a peaking HPF that both boosts the 20 hz output and at the same time stops the cone from unloading. Plus, we get 100 db from 16 hz up. Pretty good for a 12" sub in a 3 ft box.

-Bob

[dsrviola]

I've had very VERY good subs degrade the soundstage. Turn your sub off and see if the soundstage sounds better. (Not trying to be flippant). In my case (donning the asbestos) I think might have been phase issues, but I think it was actually how the sub was hooked up electronically (running of the speaker cables into the high level inputs). No matter how much fiddling I did with the adjustments, the soundstage was "cleaner" w/o the sub. Next I'm going to try low level inputs.

As for adding stuffing to the box......in my experience, if the box is the right size in the first place stuffing degrades the sound of a sealed sub. I know this also goes against common thinking.

Just passing along my experiences.

[snmhanson]

Thanks for the help Bob. This looks like the sub I want to build next. I was initially thinking I would use the 12" Titanic, but I like what you've shown me with with the RSSHF. I am getting a curve very close to yours when I try to model it myself but it is one db higher at the peak. I may not be entering the driver data correctly because WinISD is giving me different specs on some metrics than Dayton shows. If it's not too much trouble could you e-mail me the driver data and maybe even the project file if you still have it? Also, if you still have the Titanic 15 driver I would love that file as well. Let me know if you can do that and I'll PM you my e-mail address, unless you can attach it to a PM. Thanks again - I really appreciate the help.

Matt

Matt,
Sorry I missed the fact that this was an existing box. My ADD impulsivness got the best of me again. Oh, well.

Nonetheless, I think you are getting the idea. Since you are looking at using the same 3.25 cu.ft. box for a 12" sub, Here is something modeling I put together because I was considering pretty much the same thing using a PE RSS315HF-4 and an O Audio amp. The Shiva is a nice woofer too and will obviously make a great sub.

Here is the PE sub in a 3.0 cu.ft. ported box tuned to 16.5 with a 4"X33" vent at 150 watts with no bass boost. This is close to a 3.25 cu.ft. box with the vent entirely inside. Its not too pretty (yet):



It also unloads and exceeds XMax by a substantial margin below the tuning frequency.



Now, let's do a little voodoo of our own and make it pretty. We add in a 4 db boost at 20 hz with the O Audio amp HPF. (You can easily do the same with the PE BASH amp with a resistor change.):



And, here's the magic, it no longer exceeds XMax:



Unloading below tuning is one of the biggest problems with ported subs and this is an elegant way around it. Use a lower than normal tuning and then add in a peaking HPF that both boosts the 20 hz output and at the same time stops the cone from unloading. Plus, we get 100 db from 16 hz up. Pretty good for a 12" sub in a 3 ft box.

-Bob

[Bear]

Matt,
Sorry I missed the fact that this was an existing box. My ADD impulsivness got the best of me again. Oh, well.

Nonetheless, I think you are getting the idea. Since you are looking at using the same 3.25 cu.ft. box for a 12" sub, Here is something modeling I put together because I was considering pretty much the same thing using a PE RSS315HF-4 and an O Audio amp. The Shiva is a nice woofer too and will obviously make a great sub.

Here is the PE sub in a 3.0 cu.ft. ported box tuned to 16.5 with a 4"X33" vent at 150 watts with no bass boost. This is close to a 3.25 cu.ft. box with the vent entirely inside. Its not too pretty (yet):



It also unloads and exceeds XMax by a substantial margin below the tuning frequency.



Now, let's do a little voodoo of our own and make it pretty. We add in a 4 db boost at 20 hz with the O Audio amp HPF. (You can easily do the same with the PE BASH amp with a resistor change.):



And, here's the magic, it no longer exceeds XMax:



Unloading below tuning is one of the biggest problems with ported subs and this is an elegant way around it. Use a lower than normal tuning and then add in a peaking HPF that both boosts the 20 hz output and at the same time stops the cone from unloading. Plus, we get 100 db from 16 hz up. Pretty good for a 12" sub in a 3 ft box.

-Bob

This is a simulation that I would classify as "highly suspect". It looks like a way around Hoffman's Iron Law (+2dB SPLmax, significant bass extension), so I would guess that there's a problem in the simulation. The same amp driving the same driver in the same box doesn't have enough acoustical changes (e.g., horn loading) to cause a systemic change like this.

One of the most common ways of extending the frequency range of a bass driver is with a Linkwitz Transform (LT). In essence, what it does is lower absolute output for deeper bass from the same driver and cabinet volume. In other words, it uses Hoffman's Iron Law by trading SPL for extension. Jon Marsh and Tony Gee have both posted DIY designs that use this technique in the past year or so. Mark K has a good explanation (including graphs) over at his site, as well:

https://web.archive.org/web/20110808103612/http://www.audioheuristics.org/primers/Linkwitz_transform/Linkwitz_transform.htm

I'm not sure what the BASH amp is doing with its EQ circuit, but I suspect it is a variation on an LT. If so, then by changing the simulation around, I suspect that what you will find is an absence of magic: decreased absolute output in exchange for increased bass extension. At least, this is how it has worked on all of the other EQ systems I've played with (SMS-1, DCX2496, MiniDSP).

[BOBinGA]

Matt,
Sorry, but I didn't save the project files. I've actually run this simulation so many times that I can almost do it in my sleep. I'll see if I can create some project files for you this weekend. I actually have an RSS265HF-4 in a 2 ft box. The sim works almost identically except the 265 doesn't play as loud as the 315. Your one db higher is probably because I used a Qa of 10 on the Box tab Advanced button instead of leaving it at the default of 100. 10 is equivalent to foam lined with moderate stuffing.

Bear,
I would consider this a quasi-LT circuit. Like any LT type circuit, the biggest trade off is the power required for the subsonic bass. Even though you enter 150 watts in the sim, since you are boosting by 4 db at 20, you are actually asking the amp to provide a lot more power at the boost frequency. To get a 4db boost, the amp needs to produce about 400 watts. (+3db would require double the power, so I am guessing at the 400 watt figure, but its not too far off.) That's why you need a 500 watt amp for this. You also can't expect any more than the 106 db peak or the amp runs out of power or you exceed XMax or both. It doesn't violate Hoffman's iron law because it does take a lot of power in exchange for LF extension. But 100 db is loud enough for me.

When I first came across this model, I also asked myself what's wrong. It was one of those "It's too good to be true" kind of things. But I really can't find anything wrong.

The drawbacks as I see them are that you need 500 watts and the vent pipe has to be ridiculously long and large to prevent chuffing. But a 3" vent for a 10" sub and a 4 inch vent for a 12" sub are OK down to the low 20's even at 100 db and no problem at lower volumes. Output also drops off very quickly under about 16 hz. But I'll take that, too.

I have my RSS265HF tuned this way and the in room measurements are down just 2 db at 20 at moderate volumes, so reality seems to match theory.

-Bob

[Bear]

Bear,
I would consider this a quasi-LT circuit. Like any LT type circuit, the biggest trade off is the power required for the subsonic bass. Even though you enter 150 watts in the sim, since you are boosting by 4 db at 20, you are actually asking the amp to provide a lot more power at the boost frequency. To get a 4db boost, the amp needs to produce about 400 watts. (+3db would require double the power, so I am guessing at the 400 watt figure, but its not too far off.) That's why you need a 500 watt amp for this. You also can't expect any more than the 106 db peak or the amp runs out of power or you exceed XMax or both. It doesn't violate Hoffman's iron law because it does take a lot of power in exchange for LF extension. But 100 db is loud enough for me.

When I first came across this model, I also asked myself what's wrong. It was one of those "It's too good to be true" kind of things. But I really can't find anything wrong.

The drawbacks as I see them are that you need 500 watts and the vent pipe has to be ridiculously long and large to prevent chuffing. But a 3" vent for a 10" sub and a 4 inch vent for a 12" sub are OK down to the low 20's even at 100 db and no problem at lower volumes. Output also drops off very quickly under about 16 hz. But I'll take that, too.

I have my RSS265HF tuned this way and the in room measurements are down just 2 db at 20 at moderate volumes, so reality seems to match theory.

-Bob

Bob - No worries. I don't use WinISD, so that filter may be correct as it relates to the BASH amp. However, it may be worth a replacement driver to see if there is a combination of filters one could apply to more explicitly mimic an LT circuit and see if it produces the same results. It also creates a testable hypothesis, of a sort, since the in-room bass response is guaranteed to need a lot more tweaking than just a simple notch + shelf if one wants it "flat" and "deep".

[BOBinGA]

Bear,
That's where the O Audio helps again. In my room, there is a large peak centered at about 32 hz. I have to use the parametric EQ on the amp to knock that down. But after that, its pretty much flat from 22 to 60 or wherever I cross to the mains.

-Bob