So I have been out of the diy audio realm for over a decade. I do not remember any talk of tapped horns when I was last paying any attention. Transmission line spreadsheets and mathlab worksheets were the esoteric ideas. I moved, retired my (2) 15" subwoofer 16 ft box, bought a forclosure, renovated it, got married, had 4 kids, grew up a TINY bit, but I still have those old 15" subs sitting around. Now just recently I have had a bit more time after kids bedtime, combined with the itch to get some home theatre bass and EDM bass back into my life. I have not had a windows machine in over a decade either, so I had a bit of a learning curve finding something decent to replace the programs I remember (winISD pro, DOSbox, blaubox from memory) that would run on a mac or in linux. I gave that up once I saw how many people use hornresp and just sucked it up and build a windows VM off of the free Microsoft internet explorer testing VM images.
So finally, after a few days of reading forum threads and playing with hornresp, I think I *might* be a novice instead of a 'completely clueless' user of the sim package. With that all said, the drivers I have are Dayton part 295-190 (Dayton Audio DVC385-88 15" DVC Series Subwoofer) and my initial goal was to go as low as I could go. I cloned a few designs from various threads, tweaked them, abandoned them, and in the end I have modeled a ~24 ft tapped horn, 7 ft ported box, 3.25 ft sealed box, and a 9.5 ft TL. I modeled each enclosure up to the Xmechanical limit when given 500w. (paper spec sheet I have has 15.1 Xmax, 22.6mm Xmech). That might be too ambitious? Should I keep it to 15mm even if I am fine with a bit of distortion in the 10-20hz range? I just today discovered the active filter utility which let me adjust the high pass precisely to get the full excursion I could get at low freq while not just limiting total power for higher freq. Before that discovery I was just limiting to 100w across the board. I chose 500w as that is the size of the class D module I have to power this sub. The option is always there to build a second sub in the future.
I have attached all models, in 2pi except the additional plot of the TH which I also included in .5pi. I can take snapshots of the other enclosures if anyone is very curious, but I figured the tapped horn would be the main discussion point, so that is the one that I included plots for:
1 watt, 2 pi, no active filter
500 watt 2 pi active filters
500 watt .5pi active filters
Which gives my end result of 130db at 20 hz from a single 15" driver that was under $100 back when I bought it. This just seems too good to be true, even taking into account that it is insanely massive, and I still have no idea how to begin figuring out how to fold the thing. IS this a simulation anyone can look at and immediately tell something is terribly wrong with? If things seem fairly legit, is there any way to tame the 60-120hz ripple or is it not even worth trying? How do I start figuring out how to fold this?
So finally, after a few days of reading forum threads and playing with hornresp, I think I *might* be a novice instead of a 'completely clueless' user of the sim package. With that all said, the drivers I have are Dayton part 295-190 (Dayton Audio DVC385-88 15" DVC Series Subwoofer) and my initial goal was to go as low as I could go. I cloned a few designs from various threads, tweaked them, abandoned them, and in the end I have modeled a ~24 ft tapped horn, 7 ft ported box, 3.25 ft sealed box, and a 9.5 ft TL. I modeled each enclosure up to the Xmechanical limit when given 500w. (paper spec sheet I have has 15.1 Xmax, 22.6mm Xmech). That might be too ambitious? Should I keep it to 15mm even if I am fine with a bit of distortion in the 10-20hz range? I just today discovered the active filter utility which let me adjust the high pass precisely to get the full excursion I could get at low freq while not just limiting total power for higher freq. Before that discovery I was just limiting to 100w across the board. I chose 500w as that is the size of the class D module I have to power this sub. The option is always there to build a second sub in the future.
I have attached all models, in 2pi except the additional plot of the TH which I also included in .5pi. I can take snapshots of the other enclosures if anyone is very curious, but I figured the tapped horn would be the main discussion point, so that is the one that I included plots for:
1 watt, 2 pi, no active filter
500 watt 2 pi active filters
500 watt .5pi active filters
Which gives my end result of 130db at 20 hz from a single 15" driver that was under $100 back when I bought it. This just seems too good to be true, even taking into account that it is insanely massive, and I still have no idea how to begin figuring out how to fold the thing. IS this a simulation anyone can look at and immediately tell something is terribly wrong with? If things seem fairly legit, is there any way to tame the 60-120hz ripple or is it not even worth trying? How do I start figuring out how to fold this?
Attachments
Jared,
1) Yes, too ambitious. Best to design around hitting Xmax at RMS or AES power rating.
2) There will be enough distortion at Xmax ;^). Past Xmax, Bl drops quickly, and more power just generates heat, which increases impedance, which reduces power (thermal compression).
3)44.72 volts into 2.4 ohms is over 800 watts, which could easily smoke the "350 watt" voice coil if you could get it from a 500 watt amp ;^). For some EDM which has less dynamic range than a sine wave, even 175 watts could be pushing into "char-land" or 3 dB (or more) thermal compression. Other than that, (and forget .5 Pi and driving past Xmax), and use "Par(abolic)" rather than "Con(ical)" expansion, nothing seems obviously wrong.
4) Anything much past 2 octaves for a TH is doing pretty good, your sim looks quite smooth relative to it's bandwidth.
5) You could look at some of the long TH like the DSL DTS10 or Josh Ricci's Gjallerhorn for some folding examples.
The advanced centerline method is pretty accurate when figuring the folds:
Horn Folding - a brief study of the centerline vs advanced centerline method - AVS Forum | Home Theater Discussions And Reviews
You can also print the Hornresp schematic, cut it out and trace sections onto a sheet of paper to get ideas.
Have fun!
Art
Last edited:
A quick note on Xmech vs Xmax...
- Xmech is the amount of cone travel that can occur before damage. Having a design that relies on hitting Xmech is a design that will work to capacity for a couple of cycles before the damage is permanent. Damage can include smashed voice coil, folded cones, and torn suspension/surround.
- Xmax is where things start sounding bad. How bad is down to each individual driver, but driving past Xmax is generally a no-no unless you've done a lot of testing and found the results acceptable.
Chris
- Xmech is the amount of cone travel that can occur before damage. Having a design that relies on hitting Xmech is a design that will work to capacity for a couple of cycles before the damage is permanent. Damage can include smashed voice coil, folded cones, and torn suspension/surround.
- Xmax is where things start sounding bad. How bad is down to each individual driver, but driving past Xmax is generally a no-no unless you've done a lot of testing and found the results acceptable.
Chris
Hi Art, thanks for the response. It is weird in that the paper spec sheet I have does not match the last revision at parts express. I attached the paper I have which I guess should be a better match to the exact units I have? In any event, I had always treated this as a 600w woofer, 300w per voice coil, but I see that should have always been 600w max, 300w rms woofer. I have adjusted my input down in the sim accordingly. Another point I missed was using the dual 8 ohm coils in parallel I just assumed use 4 ohm for the horn resp voltage calc. I did not know to use the dc resistance. Thanks for that tip as well! So long story short, it was too good to be true! I lowered the displacement and power input and am "down" to 115db in 2pi at 20 hz. I think I will build one.
Attachments
Thanks! great to know, I will give it a try and then hopefully save ~10 gigs disk space by tossing out this windows vm.
FWIW, given the options, I'd go with an offset TL for that driver. The box will be much smaller and easier to build, and it should be usable over a wider frequency range with a smoother roll-off at low frequencies. Here's an idea to start off with. It can likely be fine-tuned (by adjusting the taper and perhaps adding a vent (see my POC6 design for how this can be folded into a box) to move the first response null up to 200 Hz, so it's way out of the sub's passband, and best of all there are no out of band response peaks to worry about.
Attachments
Thanks Brian, you are right, that looks great and makes life very simple in comparison. I will sim it up a bit, thanks for providing the inputs. The single TL I mocked up was not offset, and I had not revisited it with active filters to protect it or with the newer correct power inputs. I like this model compared to the one I did as like you said, there is no giant notch out in the 65-120hz range.
Attachments
You're welcome. BTW, before building anything, I suggest measuring the driver's parameters (REW and a simple impedance jig can do it). Or you can take the impedance measurements with DATS, import them into REW and use them to determine the additional parameters required for the "semi-inductance" model in Hornresp, which will more accurately describe the behavior of long-coil drivers like the DVC385.
I've only done 1 TH, the "SPUD" clone so my conclusion can be taken with a grain of salt.
In my experience, TH perform great on simulation and measurement of pure test tone.
Unfortunately, they perform poorly on music for audiophile use.
Stored energy my friend.
Transient response is bad and on some content, I mean the music track that excite just the right resonance, it's unbearable.
Mine ended up in the fire pit.
Had done 3 back loaded horn too, (lowther and fostex based design).
They exhibit similar issue as TH but shifted higher in frequency.
They too, lit up my cold evening cooking marshmallow...
No doubt that TH are very efficient and would work adequatly for Pro-audio.
That where you see them commercially too.
As for bass, I'm using multiple sealed subwoofer now. Must use more driver to get same spl but it's worth it.
Sorry for the negative comment but it's been my experience.
Edit: I second Brian idea of a TL instead of the TH.
In my experience, TH perform great on simulation and measurement of pure test tone.
Unfortunately, they perform poorly on music for audiophile use.
Stored energy my friend.
Transient response is bad and on some content, I mean the music track that excite just the right resonance, it's unbearable.
Mine ended up in the fire pit.
Had done 3 back loaded horn too, (lowther and fostex based design).
They exhibit similar issue as TH but shifted higher in frequency.
They too, lit up my cold evening cooking marshmallow...
No doubt that TH are very efficient and would work adequatly for Pro-audio.
That where you see them commercially too.
As for bass, I'm using multiple sealed subwoofer now. Must use more driver to get same spl but it's worth it.
Sorry for the negative comment but it's been my experience.
Edit: I second Brian idea of a TL instead of the TH.
Last edited:
etalon90: I have somewhat similar experiences. TH can be made to work well, and certainly sound impressive at first. But it takes a bit of processing power and delay+phase corrections to get the most out of it. A classic "Bass reflex done right", TL or equivalent often (but not always) have less phase and group delay disturbances in the frequency range you want it to work.
well since I have two of the drivers, I decided to stick with what I have done in the past since I know it will work extremely well and just do a ported box.
I will likely experiment with the other driver and a TL or TH, but I came to the conclusion ( maybe for this driver but not in general?) that to go lower with a TH or a TL is not possible without a much larger enclosure. IE at 8 cuft, nothing goes lower than than the ported box as excursion in all the designs is the "wall" and to even get to that wall might be a much bigger box for a TH or TL.
In any event, this is a 70X17X13 box with a 2X11.5 port 22" long. The strange shape is so it can sit behind the couch and be 100% WAF compliant.
I will likely experiment with the other driver and a TL or TH, but I came to the conclusion ( maybe for this driver but not in general?) that to go lower with a TH or a TL is not possible without a much larger enclosure. IE at 8 cuft, nothing goes lower than than the ported box as excursion in all the designs is the "wall" and to even get to that wall might be a much bigger box for a TH or TL.
In any event, this is a 70X17X13 box with a 2X11.5 port 22" long. The strange shape is so it can sit behind the couch and be 100% WAF compliant.
Attachments
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.