Collaborative Tapped horn project

staggering the woofers smooths out the group delay curve and makes the peaks and dips go away for the most part.


I tried all sorts of combinations of differing waveguide lengths in a split waveguide configuration. I have yet to find a combination of freqs that doesn't deliver some sort of discontinuity within the passband. I have finally given up pursuing this, returning to a single waveguide. In doing so I now have a very acceptable fr, group period delay that is virtually horizontal across passband. Why beat myself up needlessly? Perhaps I'm just frustrated, but I honestly don't see how I can improve on what I have.

The idea of push pull is of value to me, as I have seen deflection curves and "it ain't purty". I have not done checked into this except as a mental exercise but, consider the air load on the diaphragms of each speaker. Consider how the pressure on each diaphragm changes as amplitude within the waveguide, especially at resonance. Now, expand this to the push pull arrangement with different length waveguides. What have you done to the linearizing effects? Different forces now on each driver, seems that the concept of linearizing has been nullified at least to some extent by the introduction of unequal loads, worse, modulated by the im frequencies created. Again, this thought has not been run through equations by me but I'd be quite surprised to find that this is not the case.
 
Staggering the woofer will make almost no difference until they are 1/4 wavelength apart. For a sub the upper cutoff should be around 80 hz. So as a starting point the wavelength of 80 hz is 14.125 feet. Divide that by 4 and you get 3.531 feet. Any closer and the staggering at that frequency will not effect anything. For lower frequencies the wavelength will of course get larger. 32 hertz as an example is 35.312 feet long.

Mark
 
Mark, what I am doing is placing 2 speakers on a single baffle side-by-side, but one speaker facing opposite from the other, one blowing in while other blows out sort of thing, but connected out of phase such that the cones are still moving in same direction with applied signal.
Now, the tuned pipe behind the speaker baffle is split up the middle, such that one pipe is tuned to one freq while the other pipe is tuned to another. You might envision two organ pipes in a rank, one pipe shorter than the other. I then adjust the tuning of one pipe, leaving the other pipe alone. Recombine the outputs in front of the speaker (back to the typical TH) and measure the resultant. As we change the tuning of one pipe, the spl curve will subsequently change, as does the the group delay, impedance curve, etc.

Is this what you are referring to? Or, are you splitting the mounting baffle, both speakers rear pipes tuned to the same frequency,and moving the now 2 baffles back and forth such that there occurs a physical displacement on the 2 speakers.
 
Two push-pull mounted 12s.

TH212.gif
 
Last edited:
When you got two drivers like that, not loaded quite the same, is it better
to wire in parallel or series? I'm asking is better for impedance peak or dip
to receive the greater power? A moving cone produces greater back EMF,
so that should be the one less acoustically loaded? We want to dump the
power into the other one, thus a parallel wiring is better? Or maybe I have
it backwards...
 
When you got two drivers like that, not loaded quite the same, is it better
to wire in parallel or series? I'm asking is better for impedance peak or dip
to receive the greater power? A moving cone produces greater back EMF,
so that should be the one less acoustically loaded? We want to dump the
power into the other one, thus a parallel wiring is better? Or maybe I have
it backwards...

At the frequencies of interest, both push pull cones in djk’s 3684 post would see very similar loading, the difference is similar to what a single cone of double the diameter would see on the leading vs. trailing edge.

The decision to wire in parallel or series should just be a question of what works better for amp power.

One issue with a design such as in djk’s 3684 post is that 24 inch distance from one end of the cones to the other can result in some deep upper pass band dips and peaks, keeping that distance short, that is the 2x 12" along side each other, would avoid that problem.

But then it would not be a push pull.

For what it's worth, push pull won't do much to reduce distortion unless using speakers that behave differently going forward than back, that is speakers with suspension problems.

Danley's decision to use push pull on some of his TH designs has been based on fitting more cones in for added displacement, not for distortion reduction.

Art Welter
 
Last edited:
I've experimented with series and parallel wiring. In a horn like the one in djk's post, it doesn't make much difference. If there are significant differences in loading, parallel wiring is better.
(Notice in djk's design, the two drivers have similar path length - the distance between the back and front of each driver. You can also build the horn so that the two drivers have unequal path lengths. You need to use AkAbak or similar modelling to determine how much of a difference there is in driver loading.)
 
Art, may I refer you to :

"Distortion Reduction in Moving Coil Loudspeaker Systems Using Current Drive Technology" , P. G. L. Mills & M. O. L. Hawksford, J.Audio Eng Soc, Vol 37, No3, 1989.

Pay particular attention to the selection of driver for this development and the reason for doing so. Then look at the accompanying diagrams. Its more than just suspension that's at issue here.

There are also a series of references provided.

The rest of article is also interesting, but not pertinent to this discussion.

With all this research, how do they have time to just sit and listen?


Barry
 
An externally hosted image should be here but it was not working when we last tested it.


Doesn't seem to bother Danley, he has several models with similar loading.

"has been based on fitting more cones in for added displacement"

All the above models could have used the same number of drivers without push-pull mounting.

"not for distortion reduction."

Well I guess that's an unintended side effect then?
 
adding in an expansion chamber to AkAbak model

Not real strong on picking proper way to implement solutions within AkAbak, so please allow me to ask:

I want to split a waveguide and add in an expansion chamber. My solution is to use an enclosure, as I need to define cross section and the in/out port sizes. Is there a more appropriate way?
 
Hi sdad,

Aren't you the one who claimed he read through this thread? :) One of the problems with this, and similar threads, is the lack of an index. Oh well, go to around Post #1187, and back and forth a few pages.

Hope this helps,
Regards,
 

Attachments

  • Cordraconis_AkAbak for Dummies_2008_Feb19.txt
    8.8 KB · Views: 74
  • AkAbak script_MaVo_jnb_Post 1177.txt
    3.6 KB · Views: 54
Efficient tapped horn that reaches to 100 Hz

I built a pair of Erik's 38 Hz Tang-Band tapped horns featured on his Volvotretter site. They integrate well with my 100 Hz front-loaded horns, but I'm not happy with their efficiency (only 83 db). I'd like to build another tapped horn (or pair) that will be in the 101 - 104 db range that can cross over at 100 Hz and extend as low as possible (I realize that 20 Hz is probably not possible, but somewhere between 20 and 30 Hz would be ideal). Has anyone built a TH that meets these requirements?

Thanks!

http://sites.google.com/site/diyfirefly/
 
I built a pair of Erik's 38 Hz Tang-Band tapped horns featured on his Volvotretter site. They integrate well with my 100 Hz front-loaded horns, but I'm not happy with their efficiency (only 83 db). I'd like to build another tapped horn (or pair) that will be in the 101 - 104 db range that can cross over at 100 Hz and extend as low as possible (I realize that 20 Hz is probably not possible, but somewhere between 20 and 30 Hz would be ideal). Has anyone built a TH that meets these requirements?

Thanks!

http://sites.google.com/site/diyfirefly/

Post #6 shows a TH loaded with a cheap speaker (under $100) with the response near what you are looking for, over 105 dB sensitivity.
Has not been built yet, you can be the first on your block.

http://www.diyaudio.com/forums/subwoofers/186620-eminence-eminator-drivers.html#post2534140

I have not checked the sim myself, but I assume it hits X max at 87 watts.
125 dB to near 30 Hz on 87 watts should satisfy the home enthusiasts.

Art
 
Efficient tapped horn that reaches to 100 Hz

Thanks for the suggestion Weltersys. It looks like the driver used in that deisn is 91 db. Has anyone tried a design with a driver in the 100 db range like the Eminence Kappalite 3015LF? I'd like to experiment with a passive crossover between my 104 db mid-bass horns and tapped horns and passive between my midrange and tweeters then use a single 2-way active crossover instead of the two I'm using now.

thanks!
 
Thanks for the suggestion Weltersys. It looks like the driver used in that deisn is 91 db. Has anyone tried a design with a driver in the 100 db range like the Eminence Kappalite 3015LF? I'd like to experiment with a passive crossover between my 104 db mid-bass horns and tapped horns and passive between my midrange and tweeters then use a single 2-way active crossover instead of the two I'm using now.

thanks!
Tons of posts using the 3015LF, none of them go very low, and the in box sensitivity is lower than the Emanator loaded box.

After doing the time alignment needed (10ms) between my Keystone TH and the horn mids that go above them at 100 Hz, there is no way I'd want to use passive crossovers or active crossovers without delay.

Art
 
An externally hosted image should be here but it was not working when we last tested it.


Doesn't seem to bother Danley, he has several models with similar loading.

"has been based on fitting more cones in for added displacement"

All the above models could have used the same number of drivers without push-pull mounting.

"not for distortion reduction."

Well I guess that's an unintended side effect then?

That’s what not what Tom said.

I ribbed Tom about the PP arrangement of the TH-812 as he had had a long, ongoing “discussion” with Wayne Parham, who claimed lavish reductions in distortion with push pull in his monster sized dual Lab 12 horn, which Danley basically said would not happen.
In reply to me he said the desision was to fit the drivers in, not to reduce distortion.
The inner speakers on the TH-812 could not be turned over. The outer speakers need the angle to make the whole package be a truck pack integer.

As far as I know, Parham has no distortion measurements on a stock LabSub, so I don’t know if his PP horn has less distortion. At rated power, my front load standard configuration Lab 12 BR cabinet has less distortion than his PP cabinet.

I have found TH have more distortion than BR, PP may be a bigger advantage in TH than BR.

I wish I could have reversed one Lab 12 to test as a PP when I did my distortion tests on the Keystone TH, but that would have required messing up the mounting for the 18" speaker.

Have you ever measured the distortion of your cabinets run “normal” compared to PP ?

Art
 
Tons of posts using the 3015LF, none of them go very low, and the in box sensitivity is lower than the Emanator loaded box.

Note that sim for the Eminator is into 0.5*PI space (to match the intended use). That would explain the difference in efficiency. I think the 2*PI sensitivity works out to around 97.5db, and you can knock 3dB off of that for the 1W efficiency.

Thing is, you're not going to get a high-efficiency TH with a <30 Hz cutoff point unless you're willing to give up a good part of your living room, or be creative with how the large volume is actually achieved (aka TH-SPUD).

I agree wrt passive x-overs. Not a good idea with THs IMO.
 
Note that sim for the Eminator is into 0.5*PI space (to match the intended use). That would explain the difference in efficiency. I think the 2*PI sensitivity works out to around 97.5db, and you can knock 3dB off of that for the 1W efficiency.

Thing is, you're not going to get a high-efficiency TH with a <30 Hz cutoff point unless you're willing to give up a good part of your living room, or be creative with how the large volume is actually achieved (aka TH-SPUD).

I agree wrt passive x-overs. Not a good idea with THs IMO.
Brian,

Thanks for clearing that up, I mistook 0.5*PI space for half space.

Art
 
2 Pi or not 2 Pi?

Hi green heron,

In Post #3692 you describe volvotreter's 38Hz TH as low efficiency "only 83 db". In Hornresp that would mean using 4.0*Pi, and Eg=2V. At a more normal 2.0*Pi and Eg=2V you get 90dB.

As to the 3015LF: a 650 liter TH will get you to 96dB @ 35Hz w/ a -3dB @ 20Hz (2.0*Pi / Eg=2.83V). So it takes only two of those small boxes to get you to your target :), provided you can live w/ 2.0*Pi (and 1300 L in sub enclosures).

Regards,
 
Another AkAbak question

First attempt at working with an op amp in AkAbak. Here's a snippet:


System 'Bandpass2'

|High pass
OpAmp Def=' ' Node=18=19=20
Resistor Node=1=18 R=10kohm
Resistor Node=18=0 R=10kohm
|Resistor Node=17=19 R=10kohm
Resistor Node=20=19 R=10kohm
Resistor Node=19=0 R=10kohm

Node 1 is the voltage source, node 17 is output from previous stage low pass. As you can see, I simplified to what is supposed to be a non-inverting amp. 10k is a test only value. I rem'd out the diffy input.

Inspect-voltage at 1 is a flat line, as expected. I did check 17 and it is a low pass function, though, again, not currently used.

Node 20 has no output. And that's my question, how come?
As I said, the preceding low pass op amp stage is working, seems a non-inverting amp should work, right?. And lastly, no, when I connect node 17, I still get no output.

I feel foolish here, but it has me! I'll take the jeers if I can get the help.