When tuning a TL subwoofer well below (more than an 8va) the driver's fs it becomes essentially impossible to keep the response flat without EQ. Normally this is not a desirable thing. However, in the case of a subsonic system for HT in which we are essentially only interested in frequencies below 35Hz or so, it might not be such a bad thing.
Here is my thesis. When frequencies drop below 15 or 20 Hz our sensitivity to SPL decreases rapidly and thus a rising response below this level would be useful to compensate. An unstuffed TL has very ragged response above a certain frequency determined by the pipe tuning frequency thus we want to use a lowpass filter on the input so that these wild fluctuations would not be audible. A rising response as we move upward from the center of the passband would allow us to start the low pass filtering at a lower frequncy than would be possible with a response that is flat in the passband. This would make it possible to use a lower order filter than would otherwise be necessary.
Now it just so happens that a TL tuned "too low" gives us just such a response - rising on either side of a midband frequency. Actually it is a bump on the low end but with the tuning carefully chosen and matched with the right highpass we could produce a nearly ideal response characteristic without undue stress on the driver. The driver is further protected by the additional loading at ultra low frequencies provided by the extremely low tuning frequency.
Such a cabinet will of course be rather large but otherwise what would be wrong with such an approach?
mike
Here is my thesis. When frequencies drop below 15 or 20 Hz our sensitivity to SPL decreases rapidly and thus a rising response below this level would be useful to compensate. An unstuffed TL has very ragged response above a certain frequency determined by the pipe tuning frequency thus we want to use a lowpass filter on the input so that these wild fluctuations would not be audible. A rising response as we move upward from the center of the passband would allow us to start the low pass filtering at a lower frequncy than would be possible with a response that is flat in the passband. This would make it possible to use a lower order filter than would otherwise be necessary.
Now it just so happens that a TL tuned "too low" gives us just such a response - rising on either side of a midband frequency. Actually it is a bump on the low end but with the tuning carefully chosen and matched with the right highpass we could produce a nearly ideal response characteristic without undue stress on the driver. The driver is further protected by the additional loading at ultra low frequencies provided by the extremely low tuning frequency.
Such a cabinet will of course be rather large but otherwise what would be wrong with such an approach?
mike
Mike,
I've been developing similar plans for my HT setup. The sub will take the form of a false stage, so size is relatively unimportant. I want to try Danley's Tower of Power tapped horn, but larger and tuned lower with a pair of Tempests inside the 16ft line with a push/pull alignment and a 700L+ cab 8ft long.
Due to the limited info about Danley's design other than exterior cab dimensions, what I've done is use MJK's ML-TQWT worksheet to model the cab focusing only on the port output. I fiddled with it long enough to get a response flat from 40hz down to 20hz with a slight hump below that which doesn't start to droop until about 12hz.
My thought was that my well sealed room should give me plenty of room gain to offset the hearing sensitivity issue (it's all about feel down below 20 anyway), but the bump below 20hz should help offset the probable amplifier roll-off. My goal is to design in more output below 20hz than I will ever need.
If the tapped horn alignment doesn't work out, then I just cover the driver holes in the line and mount them on the exterior. I'd just have to EQ down the response above 20hz. In either case, with 1/4 space loading (before any room gain), I'm looking at 95db+ using less than 2.5mm of my 16mm excursion at 10hz. so 110db at 10hz in-room should be a piece of cake.
I've been developing similar plans for my HT setup. The sub will take the form of a false stage, so size is relatively unimportant. I want to try Danley's Tower of Power tapped horn, but larger and tuned lower with a pair of Tempests inside the 16ft line with a push/pull alignment and a 700L+ cab 8ft long.
Due to the limited info about Danley's design other than exterior cab dimensions, what I've done is use MJK's ML-TQWT worksheet to model the cab focusing only on the port output. I fiddled with it long enough to get a response flat from 40hz down to 20hz with a slight hump below that which doesn't start to droop until about 12hz.
My thought was that my well sealed room should give me plenty of room gain to offset the hearing sensitivity issue (it's all about feel down below 20 anyway), but the bump below 20hz should help offset the probable amplifier roll-off. My goal is to design in more output below 20hz than I will ever need.
If the tapped horn alignment doesn't work out, then I just cover the driver holes in the line and mount them on the exterior. I'd just have to EQ down the response above 20hz. In either case, with 1/4 space loading (before any room gain), I'm looking at 95db+ using less than 2.5mm of my 16mm excursion at 10hz. so 110db at 10hz in-room should be a piece of cake.
Hi,
our ears/brain hearing system has had all it's life to adjust to whatever response it "hears" with.
A sound impulse that includes some bass will sound natural at that level. It's what your hearing has got used to after all.
Now try to reproduce that noise you previously heard outside. Add some EQ because Fletch said so.
The hearing system will recognise the extra bass you have added as being unatural.
My conclusion for all this is leave well alone.
My ears tell my brain they have heard something. My brain does the calculations and tells me it sounds quite like what I have heard before.
It looks up it's ROM and says that' a car door slamming, Oh a heavy door at that.
I have no intention of adding any response tailoring because Munson started talking to Fletcher.
our ears/brain hearing system has had all it's life to adjust to whatever response it "hears" with.
A sound impulse that includes some bass will sound natural at that level. It's what your hearing has got used to after all.
Now try to reproduce that noise you previously heard outside. Add some EQ because Fletch said so.
The hearing system will recognise the extra bass you have added as being unatural.
My conclusion for all this is leave well alone.
My ears tell my brain they have heard something. My brain does the calculations and tells me it sounds quite like what I have heard before.
It looks up it's ROM and says that' a car door slamming, Oh a heavy door at that.
I have no intention of adding any response tailoring because Munson started talking to Fletcher.
I will agree with the previous poster... Flat response is flat response, and no amount of EQ is going to make things sound more natural than that....
Now I will agree that you dont need flat anechoic response on a sub down to infrasonic frequencys. Room gain will help out, and our ears are not sensitive to these frequencys anyhow, so judging their loudness can only be done by how much vibration we feel and rattling we hear.
I really think a TL is a horrible way to approach infrasonic bass. You're BEST off going multi-driver sealed/IB, PR comes second, with Ports in last. TLs and horns are out... the size/efficiency tradeoff is NOT worth it
Now I will agree that you dont need flat anechoic response on a sub down to infrasonic frequencys. Room gain will help out, and our ears are not sensitive to these frequencys anyhow, so judging their loudness can only be done by how much vibration we feel and rattling we hear.
I really think a TL is a horrible way to approach infrasonic bass. You're BEST off going multi-driver sealed/IB, PR comes second, with Ports in last. TLs and horns are out... the size/efficiency tradeoff is NOT worth it
I wouldn't use a TL for infrasound. TL's have a 24db/oct rolloff and are way too big for what you get. For thh same or less volume, you could use 4 sealed or 4th order bandpass subs, which have 12db/oct rolloff slopes. Then you can tailor the response with actives after the fact (only if needed) rather than building it into the sub, which I feel is a mistake.
While the fletcher munson (BTW, for headphones only - you should use robinson-dadson for loudspeakers) curves show a loss of sensitivity at low frequencies, note that the curves are bunched closer together at low frequencies - a perceived 10db increase is acheived with only 5-6dB at low frequencies.
The correct way to interpret those curves in order to make a loudness contour is to take the curve which represents the volume of the recorded sound, then subtract from the curve which represents the volume you will play back at and add the difference between the curves. Lets say the recording was made at 120dB and we play back at 90. So the difference is 30dB at 1kHz, if the curves are only 24dB different at a lower frequency, you need to boost that frequency by 6dB to get the same perceived frequency balance as in the original recording. But then if you play back at 100dB, it will sound too bassy - IMO, the best is to just leave it flat.
I once saw an article in Speaker Builder where a gentleman made a 4th order BP using a chemical barrel as the vented enclosure, the woofer being run IB into a crawlspace with the vent going from the barrel to his listening room. He got an infrasonic F3 and claimed good performanc/sound - all with a cheap ratshack 15" driver.
While the fletcher munson (BTW, for headphones only - you should use robinson-dadson for loudspeakers) curves show a loss of sensitivity at low frequencies, note that the curves are bunched closer together at low frequencies - a perceived 10db increase is acheived with only 5-6dB at low frequencies.
The correct way to interpret those curves in order to make a loudness contour is to take the curve which represents the volume of the recorded sound, then subtract from the curve which represents the volume you will play back at and add the difference between the curves. Lets say the recording was made at 120dB and we play back at 90. So the difference is 30dB at 1kHz, if the curves are only 24dB different at a lower frequency, you need to boost that frequency by 6dB to get the same perceived frequency balance as in the original recording. But then if you play back at 100dB, it will sound too bassy - IMO, the best is to just leave it flat.
I once saw an article in Speaker Builder where a gentleman made a 4th order BP using a chemical barrel as the vented enclosure, the woofer being run IB into a crawlspace with the vent going from the barrel to his listening room. He got an infrasonic F3 and claimed good performanc/sound - all with a cheap ratshack 15" driver.
Hmmm... some interesting thoughts. Certainly you guys have way more experience than I but some things suggested do surprise me somewhat. All the sealed systems that I have simulated have needed a whole boatload of EQ to reach decent output at subsonic frequencies even when relatively high Q drivers were used. I would rather have to cut lows to a box that has too much low output than the other way around I think.
BP looks interesting but it seems like with most drivers the efficiency gets royaly flushed when tuning low enough to get reasonable subsonic output. One exception was the Dayton 15" IB sub in a huge box. Seems like relatively high Q and very low fs are needed to make BP work well.
Is the objection to EBS just the size requirement and steep rolloff? I suppose that if cutoff is sufficiently low the rollof is not a concern.
The TLs that I sim'ed ran about 19-20' long with starting cross section of about 2'^2 tapering down to half that. Folded into four sections it makes a large but managable peice of furniture or a possible under floor system. The output is about 90dB/W down to around 10Hz. Using the Dayton DVC 15" with one coil driven and the other open we are not xmax limited at 10Hz until about 102db (40W). Above 13Hz (specifically at 20Hz) we reach xmax at 110dB (180W). Now we may reach voice coil dissipation limits first for steady state signals but I doubt that normal signals would be of sufficient duration to cause problems here.
I would like to see your proposals on how to provide this level of performance with sealed or BP. I am not being argumentative here. I really am interested in the merits of these other approaches and it helps to see examples. 90dB/W down to 10Hz would seem to be a challenging design target.
mike
BP looks interesting but it seems like with most drivers the efficiency gets royaly flushed when tuning low enough to get reasonable subsonic output. One exception was the Dayton 15" IB sub in a huge box. Seems like relatively high Q and very low fs are needed to make BP work well.
Is the objection to EBS just the size requirement and steep rolloff? I suppose that if cutoff is sufficiently low the rollof is not a concern.
The TLs that I sim'ed ran about 19-20' long with starting cross section of about 2'^2 tapering down to half that. Folded into four sections it makes a large but managable peice of furniture or a possible under floor system. The output is about 90dB/W down to around 10Hz. Using the Dayton DVC 15" with one coil driven and the other open we are not xmax limited at 10Hz until about 102db (40W). Above 13Hz (specifically at 20Hz) we reach xmax at 110dB (180W). Now we may reach voice coil dissipation limits first for steady state signals but I doubt that normal signals would be of sufficient duration to cause problems here.
I would like to see your proposals on how to provide this level of performance with sealed or BP. I am not being argumentative here. I really am interested in the merits of these other approaches and it helps to see examples. 90dB/W down to 10Hz would seem to be a challenging design target.
mike
Sealed takes kilowatts of power and multiple drivers, so you save some on size but spend on power and drivers. BP takes just as much enclosure as a TL to get to 10hz, or multiple drivers so you can go isobarik, but even then you're still looking at 300L plus and still nowhere close to 90db even with floor loading. Danley's getting 95db down to 20hz, so I want to use the same concept with a bigger box and more driver Sd to get down to 10hz.
For me the big box isn't a concern, but I can't go bigg enough for IB. I need it to keep the kids away from my screen anyway, 700L+ saves me money in terms of amplifier power. With my low pass at 30hz, it's really just for HT and I want to design mine with excess capacity at the bottom. I prefer to boost the subsonics for HT to a point the sounds very unnatural for music, but fine is for HT. If I end up needing EQ, I want it to be EQ'g down below 20hz, not up.
For years now I have been reading and hearing people talk about "subsonic" frequencies and how we don't hear them, we just have tactile experiences. Until recently I never had a chance to experiment because I never had a sub that could get lower than 20 hz.
Now I have had my time with a "subsonic" sub and this is what I learned. I learned that I can hear 20 hz just fine. I can also hear 15 hz, 10 hz and 2 hz. Now at 2 hz, is it mostly distortion that I hear? Who cares what it is, I can hear it just fine.
So it occurred to me that since most people cling to the word and definition of "subsonic" when talking about low tuned subs they must be having problems, and instead of fixing them, have decided it is impossible to hear much below 20 hz.
These are the problems I have encountered, and I assume everyone with low tuned subs will face as well.
1. Seating position. If you choose your seating position based on how the mains sound there is little possiblility that the system will sound balanced down to 15 hz. I chose my seating position based on where the tuning frequency of the sub (16 hz) sounded best and then moved the room around that seating position.
2. Source rolloff. No matter what source you play your music with, I doubt response is flat to single digit frequencies. How can you listen to a flat response if the signal doesn't even get to the amp flat?
3. Amp rolloff. See above, and then add the source rolloff and the amp rolloff together. In my situation 25 hz was getting ten times more power than 15 hz. That's an extreme example, but that's what happens when you use cheap consumer electronics outside of their range.
4. Hearing rolloff. In my opinion it must be accounted for. It is my opinion that a sub should be AUDIBLY flat in it's operating range. I used to think that would be impossible because of misleading words like "subsonic" but it is very possible and I have come VERY close, even using my cheap consumer electronics. And in my opinion it sounds much more natural than a system that is flat down to about 25 hz and then gets quieter until you can't hear anything (but feel a little bit) at the tuning frequency.
I admit I like my bass and lots of people would not like an audibly flat response below 20 hz but now that I know it is not only possible, but relatively easy, I will not design any other way.
By the way, this is an HT sub first and foremost, but music sounds very good as well, without changing any settings and with no eq at all.
Now I have had my time with a "subsonic" sub and this is what I learned. I learned that I can hear 20 hz just fine. I can also hear 15 hz, 10 hz and 2 hz. Now at 2 hz, is it mostly distortion that I hear? Who cares what it is, I can hear it just fine.
So it occurred to me that since most people cling to the word and definition of "subsonic" when talking about low tuned subs they must be having problems, and instead of fixing them, have decided it is impossible to hear much below 20 hz.
These are the problems I have encountered, and I assume everyone with low tuned subs will face as well.
1. Seating position. If you choose your seating position based on how the mains sound there is little possiblility that the system will sound balanced down to 15 hz. I chose my seating position based on where the tuning frequency of the sub (16 hz) sounded best and then moved the room around that seating position.
2. Source rolloff. No matter what source you play your music with, I doubt response is flat to single digit frequencies. How can you listen to a flat response if the signal doesn't even get to the amp flat?
3. Amp rolloff. See above, and then add the source rolloff and the amp rolloff together. In my situation 25 hz was getting ten times more power than 15 hz. That's an extreme example, but that's what happens when you use cheap consumer electronics outside of their range.
4. Hearing rolloff. In my opinion it must be accounted for. It is my opinion that a sub should be AUDIBLY flat in it's operating range. I used to think that would be impossible because of misleading words like "subsonic" but it is very possible and I have come VERY close, even using my cheap consumer electronics. And in my opinion it sounds much more natural than a system that is flat down to about 25 hz and then gets quieter until you can't hear anything (but feel a little bit) at the tuning frequency.
I admit I like my bass and lots of people would not like an audibly flat response below 20 hz but now that I know it is not only possible, but relatively easy, I will not design any other way.
By the way, this is an HT sub first and foremost, but music sounds very good as well, without changing any settings and with no eq at all.
It is a tl tuned to 16 hz, and you are right, it does not play 2 hz well. But if I give it all the power it can take at 2 hz (about 1 watt) I CAN hear an audible pulsing. It is not loud at all, but it is loud enough that I can clearly hear it. I can sweep down and it keeps getting quieter and quieter, but I DO have audible output at 2 hz. Distortion or not, I can hear it clearly (if not loudly).
If this sub was built for maximally flat response, I agree I would hear nothing at 10 hz, but as it is I can hear 10 hz just as well as any frequency above 20, just a whole lot quieter because of the power rolloff due to my cheap amp and source.
I invite anyone who does not believe this to come and listen for themselves, but it would probably be cheaper to build one for yourself and find out.
If this sub was built for maximally flat response, I agree I would hear nothing at 10 hz, but as it is I can hear 10 hz just as well as any frequency above 20, just a whole lot quieter because of the power rolloff due to my cheap amp and source.
I invite anyone who does not believe this to come and listen for themselves, but it would probably be cheaper to build one for yourself and find out.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.