Ooh looky - a TH that does from ~42Hz-500Hz +/-3dB

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Stumbled across this while trying to come up with a new TH alignment (POC6) that focuses on minimizing the out of band peaks rather than just concentrating on passband flatness. The driver in question is the cheap Parts Express PA310 12" driver (measured parameters).

The "trick" here seems to make S1-S2 a bit longer than usual, and then stuff it. The peaks can then be brought down substantially in level with only a very minor impact on low frequency performance.

Folding it might be a challenge though. Neither the THAM fold or SS fold seem to be good fits...
 

Attachments

  • 2016-09-14 (4).png
    2016-09-14 (4).png
    11.9 KB · Views: 1,004
  • 2016-09-14 (5).png
    2016-09-14 (5).png
    16.1 KB · Views: 985
  • 2016-09-14 (6).png
    2016-09-14 (6).png
    20.8 KB · Views: 990
I think the problem you'll run into is that the closer you get to the top of the passband, the smaller the wavelengths are going to be compared to the horn size. At those higher frequencies the waves will stop oscillating down the horn and start bouncing around within the horn, creating all kinds of interesting interference. A straight FLH can avoid this, but a TH by nature will require at least one fold.

That's just my guess as to what will happen, anyways. I'd still be interested to see what would happen if it were built.
 
I think the problem you'll run into is that the closer you get to the top of the passband, the smaller the wavelengths are going to be compared to the horn size. At those higher frequencies the waves will stop oscillating down the horn and start bouncing around within the horn, creating all kinds of interesting interference. A straight FLH can avoid this, but a TH by nature will require at least one fold.

That's just my guess as to what will happen, anyways. I'd still be interested to see what would happen if it were built.

Similar to what I've read 're: folded midrange horns. I'd suspect a low pass at 200 - 250hz would be used. That being said, I'd be willing to build a prototype as I have a bunch of sheets of ply just hanging out in the garage.
If that dayton woofer could get repurposed into a two way with a seos waveguide I would consider it.
 
I think the problem you'll run into is that the closer you get to the top of the passband, the smaller the wavelengths are going to be compared to the horn size. At those higher frequencies the waves will stop oscillating down the horn and start bouncing around within the horn, creating all kinds of interesting interference. A straight FLH can avoid this, but a TH by nature will require at least one fold.

Oh, I have no doubt that the output of this TH won't be so nice and smooth all the way up to 500 Hz when folded up in a box :). Based on my POC3 design, I suspect that deviations from the sim will start around 250 Hz, if not a bit earlier, and will manifest as "notches" in the measured response above that frequency. Still, to get a fairly flat passband from just over 40 Hz to just over 200 Hz isn't bad. Perhaps with a better driver and using a similar approach (damping S1-S2 to reduce the out of band peaks), it might even be possible to extend the response at bit lower.

The trick is to come up with a particular fold for this TH. a dual-expansion THAM type fold might work if it can be rejigged to have the drive mounted facing the back rather than the bottom. However that design has a big fold in the S3-S5 section and that will likely result in response aberrations appearing a bit earlier in frequency. A reworked "Keystone" type fold perhaps? There is a bit of unused volume in that fold that can be used for a longer S1-S2. However it's a single-expansion design, so a totally different TH alignment will have to be worked out in HornResp to fit.
 
1)Oh, I have no doubt that the output of this TH won't be so nice and smooth all the way up to 500 Hz when folded up in a box :).
2)A reworked "Keystone" type fold perhaps? There is a bit of unused volume in that fold that can be used for a longer S1-S2. However it's a single-expansion design, so a totally different TH alignment will have to be worked out in HornResp to fit.
Brian,

1) One of the smoother horns from 350 to 6kHz I have owned or measured is the Electro-Voice "Boat Horn" (I unfortunately can't recall the model number), a folded "re-entrant" horn with two (basically) 180 degree turns. Careful attention to folds is needed if smooth response above the transition from laminar flow to ray-tracing frequency is desired. Note the "Boat Horn's" "ripple" does not exceed +/- 1.5 dB above 500 Hz, very simple EQ corrections can make their response virtually flat.

2) The "Keystone fold" is no different than the type of folds used in horns for centuries. My only "claim to fame" (if that ever happens ;) ) is the combination of an easy-to-build fold with an exit shape that smooths out the ragged upper response typical of a "tapped" horn, regardless of fold pattern. Note that the "Keystone Correction" works it's "magic" in the low frequency laminar flow region, fancy corner rounds don't do anything other than reduce valuable cabinet volume in the usual "sub-woofer" range below 160 Hz or so.

The EV/University "Boat Horn" is normally white in color, the pair below have one edge of the horn sawed off for better upper polar response and a smaller package (under one cubic foot) used for yelling at powerboats leaving large wakes in no-wake zones, which flops a sailboat around like a fish out of water. Painted them black when I started using them in parades, they get louder than a dozen bagpipers with only 10 watts :^) .
Using a 12 dB per octave passive crossover, the pair can do over 130 dB at one meter at around 100 watts, well within the safe range of these little aquatic beasties.

I will be adding a small woofer and tweeter to the pair after several other projects are complete, six "B-Low" Keystone subs in the works.

Art ("On the Waterfront") Welter
 

Attachments

  • Pair of EV "Boat Horns".png
    Pair of EV "Boat Horns".png
    153.5 KB · Views: 854
  • Boat Horns.jpg
    Boat Horns.jpg
    79.8 KB · Views: 825
  • Outback Plywood.jpg
    Outback Plywood.jpg
    58.7 KB · Views: 313
Last edited:
2) The "Keystone fold" is no different than the type of folds used in horns for centuries.

Hmm... it looks like I might have referred to the wrong fold. I thought the KS fold had a feature slightly different to the other popular folds - a bit of unused volume at S1 which could be put to the use I'm considering for this TH design. However I went and reviewed your design and that's not the case at all.

Now I'm wondering whose design I was thinking of.... :)

Here's another example of a "smooth" TH - this time using the Eminence Kappa 3012LF driver, which on paper is a bit more capable than the Dayton PA310. The design process is basically the same - extend S1-S2 a bit, then stuff it, and that starts flattening the response above about 120 Hz or so, with little impact at lower frequencies. Then mod S4-S5 and overall horn length to achieve an acceptable bass response curve. In this case the net size works out to around 210 L, which is a bit big for a TH based on a 12" driver. Or is it? :)
 

Attachments

  • 2016-09-22 (2).png
    2016-09-22 (2).png
    20.7 KB · Views: 289
Hmm... it looks like I might have referred to the wrong fold. I thought the KS fold had a feature slightly different to the other popular folds - a bit of unused volume at S1 which could be put to the use I'm considering for this TH design. However I went and reviewed your design and that's not the case at all.

Now I'm wondering whose design I was thinking of.... :)

In this case the net size works out to around 210 L, which is a bit big for a TH based on a 12" driver. Or is it? :)
Brian,

210 L is not all that big for a "horn" cabinet, but "big" is a relative term, all depends where you plan to stick it ;)

You may have been thinking of the "stub" used at the beginning of the horn path featured in many DSL TH designs. I have described the "stub" (Tom's term, a reference to antenna design) with examples in several posts, including those in XOC1's DSL TH-118 clone thread.

I had experimented with a "stub" of sorts in the Keystone, but as it did not improve performance in any respect, was not incorporated in the final design.

Art
 
Last edited:
I had experimented with a "stub" of sorts in the Keystone, but as it did not improve performance in any respect, was not incorporated in the final design.

In the sims it doesn't do anything at all - until you start stuffing it :). Once it starts to get stuffed, the out of band peaks and dips diminish significantly. The end result is a smoother out of band response with little or no impact on in-band performance. Of the course the question would be if it's out of band, why bother? Well, it could mean that simpler and perhaps higher x-overs can be used, and distortion caused by in-band frequencies would be not be amplified by those out of band peaks...

Anyway, it wasn't the DSL designs I was thinking of. This was definitely a fold more like the one in the KS, but with a portion before S1 blocked off and unused. But who knows - a similar technique may have deployed in the DSL designs too. I've never got a chance to look inside one to confirm.
 
Last edited:
Hi Brian,

Were you thinking of Don Snyder's LAB12 TH fold?

http://www.diyaudio.com/forums/subwoofers/143714-lab12-tapped-horn.html

That could be it, though I seem to remember that cutoff stub before S1 being a bit longer. That looks like an MTH fold with the mouth moved to one side instead of facing into the horn.

The stub doesn't actually have to be too long. According to Everest's "Acoustics" book, 4 inches of fiberglass wool 3lb/cu.ft. is enough to absorb down to 125 Hz. So the idea here is to basically make it long enough so that, when you stuff it, all the peaks in the frequencies you don't want (typically above 100 Hz) are damped.
 
Hi Brian,

Here is a quick sketch @ adding a L12 stub to a THAM style enclosure, this could be combined w/ cone compensation if desired:

Regards,

Hmm, that's a pretty interesting way of addressing the requirement, and it should help to reinforce the external panel at S1 and S2 as well - one of the "weaknesses" of this type of fold (large flat panel at a high-pressure location).

I suspect that I can achieve the same result by replacing part of the extended S1-S2 path with a simple volume instead. As it's going to be pretty stuffed, its dimensions should have little impact on FR as the stuffing is going to damp the response at those frequencies anyway. This allows for some flexibility for using this approach in other types of folds like the "SS15" fold (and hmm... maybe that might explain the "stub" used in TD's designs).

HornResp can't model a TH with a volume tacked on at S1, but I'm sure it can come very close if we break S1-S2 into S1-S1'-S2, then incorporate into the horn's layout a volume at S1' that's equivalent to the S1-S1' section in the HornResp model. The length of that section should be short enough so any aberration between the model and the horn's actual response will be at frequencies outside of the bass range.
 
Brian,

210 L is not all that big for a "horn" cabinet, but "big" is a relative term, all depends where you plan to stick it ;)

Any design I come up with has to fit in my car's trunk, and I think one that requires 210 L of net volume is probably going to be a bit too big for that, LOL.

I'm going to measure the available trunk space and work backwards from that. It's likely I might have to go with a somewhat compromised design, but then all designs involve some sort of compromise, don't they....
 
Any design I come up with has to fit in my car's trunk, and I think one that requires 210 L of net volume is probably going to be a bit too big for that, LOL.

I'm going to measure the available trunk space and work backwards from that. It's likely I might have to go with a somewhat compromised design, but then all designs involve some sort of compromise, don't they....
Yes, the pair of subs I will be using for my home theater also are designed to fit in the trunk of my Mustang, so could only be 21.5 inches (67 cm) wide, 18 inches (46 cm) deep and 11.25 inches (28 cm) tall, a bit less than 210 L.

At least they will be lightweight ;^).

The subs would have been done by now had I not spent 20 hours cleaning up after Matthew, hurricanes are a compromise of living in this part of the country. At least it seems to have started cooling off a bit after the hurricane blew through, turned off the air conditioners for the first time in the four months here a few days ago, though have them on again today.

Time to vacuum the leaves out of the house, at least they are easier to clean than the coal dust I dealt with for the 20 years before moving to Florida...

Art
 
Last edited:
What would be a practical material to use for stuffing the stub?
Also what would be the best way to ensure what it stays in place?
I imagine that the material in the stub would be moving around quite energetically.

Polyester fiberfill should do the job.

The area is a high-pressure, not high-velocity area so it shouldn't move around that much. However, a wire mesh or a piece of speaker grille cloth stapled across the stub can be used to keep it in place. Even old t-shirt material should work well for this.
 
I have been looking to find some data on airflow resistivity of lining materials which is the parameter needed by Hornresp and is measured in mks rayls/m.
It seems to be quite hard to come by this data. Has anyone got any links to how to find out this parameter for different acoustical adsorbent materials?
With a denser material it seems possible to dampen the higher frequency ripples with a 50% filling between s1 and s2 avoiding the area where the drive unit is. I have simmed with an airflow resistivity of 4000 to 20000 and that shows significant reduction in the peaks at the higher end of the passband. I note that the default record in Hornresp uses a lining with a resistivity of 40000.
Reducing these peaks could reduce the need for such a high order low pass filter in the system design.
 

Attachments

  • 50 Fill reduction.jpg
    50 Fill reduction.jpg
    80.1 KB · Views: 256
I was also expirmenting with this in sims. I wonder if lining the first section with something like dynamat or bitumen would have impact or not. Guess i need to get my TH18 prototype out again and do some RL testing with differrent materials. Although in the Xoc1 TH18 sim I got more results with filling S4, it flatten out the peaks > 110.

Also note that the xmax is raised a lot.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.