New sub design? Constricted Transflex, simple build (series tuned 6th order)

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Founder of XSA-Labs
Joined 2012
Paid Member
XRK971
I tinkered around with HR a little more to try to approximate the proper shape of the first section and of course it puts the driver too close to the constriction , so even though there really isnt any way that i know of to fully sim this box in Hornresponse i do see that lo-and-behold this shape does shift the FB up by a few hertz ... So good call .. I didn't notice that before ..
I did however come up with a variation that we can fully simulate using the throat chamber function .... Looks like it only loses a db or so, and is still as simple as the other design while still maintaining similar outer dimensions (or even a little smaller) .... It is a little more flexible too and could possibly accommodate a wider variety of drivers since changing the shape and length of "Lpt" and "Ap1" seems to have significant effects upon the response curve..
Extension seems slightly better and the low compression ratio (almost 1:1) should reduce the likelihood of physical damage with the drivers such as cone failure ..

Tell me what you think of this ..

I don't see the constrictions shown in your drawing implemented in the HR input page. What you have is similar to a push-pull (PP) tapped pipe, here is a model I made for Legis in this thread: http://www.diyaudio.com/forums/subwoofers/258706-study-dipole-cardioid-bass-horn-14.html#post3994808

If you get AkAbak it would be easy to modify geometry of script to simulate your transflex PP tapped pipe.
 
XRK971,
That PPTP looks like it will produce some serious output with those big drivers and big power! ..... I am looking forward to reading your friend's reports once he gets them built .... The bit about adding an additional rapidly expanding horn to the front was interesting too ...

Ok , I will explain about this smaller box ,I refined my inputs on my Constricted-PP transflex design so now the area is the same on both pipe sections like you would expect with a tapped pipe ..... Take a look at the figures in the lower fields ... I am using the throat chamber function to create the first half of the Tapped Pipe ...
VTC= defines volume of first chamber section (before the constriction)
ATC=The area of the first chamber section
Ap1=Defines the area of the constriction (port)
Lpt=Length of constriction/port

I used a VTC of 48 liters and an ATC of 600 sq cm which creates a pipe shaped chamber of about 80cm in length .... The other pipe section is defined with the typical S1 through S5 inputs which i made to match the first section's length at 80cm also, giving a total of 160+20 (constriction length) for a total of 180cm path length which tunes the box to 40hz .....
:)
 

Attachments

  • ConstrictedTransflexV2.JPG
    ConstrictedTransflexV2.JPG
    61.5 KB · Views: 1,307
Last edited:
Using the constriction in the design above can help alleviate the "saddle" or smiley face response curve that you run into in some situations .... The "saddle" can become even more of a problem when you start spanning multiple drivers down the path due to having to move your "L12" and "L45" points further away from the ends of the pipe (IE 30cm instead of 15cm) In order to accommodate the extra driver or drivers being spanned down the path (when spanning you are supposed to use the center points between the two drivers for simulations) ... So in this case (multiple drivers spanned) a constriction can (under the right conditions) help compensate for the change in response and help prevent the saddle effect .......
 
Of course one could say: "forget about these complications like saddles and constrictions, and lets just keep things simple by putting each 10" woofer into it's own 50 liter 40hz tapped pipe!" which actually looks really good in simulations but unfortunately as it turns out the slender physical dimensions of a standard 50 liter tapped pipe tuned to 40hz is a bit too small to shoehorn one of these Alpine 10s into, not ideal, magnet would probably stick out.... :( ... hmmph ...

SO LETS EXPLORE ANOTHER OPTION ;) ..... Something I came up with while experimenting with the throat chamber function in hornresponse some more last night .... It is possible to fully simulate A mass loaded tapped pipe, a sort of a mixture of an ML-TL (a widely used and trusted design) and a Tapped Pipe but the "mass loading" (a constriction) is placed in a spot that allows the upper bandpass limit to be raised giving this design more bandwidth than any typical 40hz tapped pipe! The mass loading also allows the physical line to be shorter changing the 50 liters worth of dimensions to something thats a little less toweresque and moreso cube-ish and fitting the 10" driver is now much easier :D
Upper bandwidth cutoff is now 180hz instead of 140hz ... A little bit of stuffing/lining in the first 30% of the line would smooth out the small slightly under-damped response bump at 40hz (if it is even noticeable in the real world).. Response above 50hz is only down by roughly 1db compared to these other designs I have posted here, and it is a worthwhile sacrifice for the extra bandwidth ... Design is still very simple, and no need for an access panel! Cool eh? :cool:
 

Attachments

  • 50 liter -ML-TRANSFLEX-40hz.JPG
    50 liter -ML-TRANSFLEX-40hz.JPG
    26.2 KB · Views: 2,157
  • 50 liter -ML-TRANSFLEX-40hz-4cabs-halfspace.JPG
    50 liter -ML-TRANSFLEX-40hz-4cabs-halfspace.JPG
    46.4 KB · Views: 1,314
  • 50 liter -ML-TRANSFLEX-40hz-HR_INPUTS.JPG
    50 liter -ML-TRANSFLEX-40hz-HR_INPUTS.JPG
    60 KB · Views: 1,309
Last edited:
Using the constriction in the design above can help alleviate the "saddle" or smiley face response curve that you run into in some situations .... The "saddle" can become even more of a problem when you start spanning multiple drivers down the path due to having to move your "L12" and "L45" points further away from the ends of the pipe (IE 30cm instead of 15cm) In order to accommodate the extra driver or drivers being spanned down the path (when spanning you are supposed to use the center points between the two drivers for simulations) ... So in this case (multiple drivers spanned) a constriction can (under the right conditions) help compensate for the change in response and help prevent the saddle effect .......

I noticed the same thing regarding "tapped pipe" that X already mentioned in post #21. It has around 2:1 compression ration in the pipe section, that is quite long. It loads the cone very good and linearizes the response.

Besides keeping the saddle effect away, it also seems to allow front horn type of loading in front of the tapped pipe while the response still stays almost as linear. This can reap double benefit, +6dB sensitivity from tapped pipe/horn and +6dB from front horn!:eek: At the same diaphragm excursion.

Using a front horn on a design that already has saddle effect or otherwise unlinear response (too little damping/compression ratio along the way) usually just makes those phenomenoms worse by volumizing the resonances.

Right amount of "damping" (in form of compression ratio) loads the cone the way that it cannot resonate as much in resonance peaks, even if they are made greater with front horn loading.

Also, if the amount of damping is correct, the overall sensitivity (of the tapped pipe/horn section) stays the same as that of a similar tapped horn with saddle effect (not taking into account the bonus from front horn).

I see huge potential of the pipe/constricted loading because of the resonance-taming effect that allows front horn loading as a bonus.

So, is there downsides, can someone think of any (based on sims or experiences) ?
 
Last edited:
Downside is perhaps not as compact as traditional expanding path TH. I also don't think you can ever get as much compression ratio loading as a thin chambered expanding horn. Certainly easier to build if you have he volume.

Yes, maybe intelligent mixing of both pipe and positive taper horn is the way to go. I also made slight changes to the plan, I increased the lenght of the positive taper section. It is even better now in higher freqs. I just have make slight angles to the last couple of walls along the path, footprint stays the same.

See the pictures here: http://www.diyaudio.com/forums/subwoofers/258706-study-dipole-cardioid-bass-horn-17.html#post4003722
 
A mass loaded tapped pipe, a sort of a mixture of an ML-TL (a widely used and trusted design) and a Tapped Pipe but the "mass loading" (a constriction) is placed in a spot that allows the upper bandpass limit to be raised giving this design more bandwidth than any typical 40hz tapped pipe!

The phrases 'What was once old is yet new again' and 'those who don't learn from history are doomed to repeat it' come mind. ;)

Harry Olson's 1937 BLH [the reference for designing wide BW BLHs and now its tapped variants] used a series of 'stepped' rectangular tubes to approximate its HF expansion and fold a long path-length into a [then] reasonably compact box: Patent US2224919 - Loud-speaker - Google Patents

Note that he also did a front & back [compound] horn loaded studio monitor where the expansion is all tubes on the back side, but don't have a picture of its cutaway. A pity as its complexity appears to have been laid out by someone versed in 'maze' designs.

GM
 
Founder of XSA-Labs
Joined 2012
Paid Member
XRK , YEP! Thats right, it is essentially very similar to the Slot Loaded Bandpass(SLBP) which you posted the link for ...They are 6th order series tuned bandpass boxes, the ML-TP above just has a little more path length so it uses a resonance mode that is somewhere between "quarter wave" and Helmholtz (like an ML-TL) instead of just strictly Helmholtz .... :)
 
GM ,
Awesome old patent reference! I saw a couple of compound horns in his sketches , neat stuff , can be simulated using the "CH" mode in Hornresponse .... I played with CH mode a bit and found out that if you tune one resonant pipe or horn section to fill in the 3rd harmonic dip of the other you can actually get some decent output and wide-ish bandwidth ... It makes for a big box though ... The old Bose Wave Cannon was set up this way , as a compound pipe ...
I also saw a scoop (BLH) cabinet set up with a small bandpass section attached to the front of the driver to accomplish the same sort of thing (fills in the 3rd harmonic "suckout") but of course it wont extend the response as high as Olson's Short Horn section would..... So i suppose you sacrifice bandwidth for compactness in this case ..... I will attach a couple of modern examples that i have found out there on the interwebz ..
 

Attachments

  • usefullscoopyf1.jpg
    usefullscoopyf1.jpg
    31.3 KB · Views: 1,135
  • colossusmk1buildpic.png
    colossusmk1buildpic.png
    318.9 KB · Views: 1,171
  • newscoop11pz2 (1).jpg
    newscoop11pz2 (1).jpg
    38 KB · Views: 1,136
Thanks for the reply but I got that far :)

Default record number 1 cannot be edited. Click the Add button to create a new record.

My issue is I don't see 4 CON boxes but rather CON L23 L34 L45, which doesn't match up with the images provided? How do I fill this out because I cannot get it to simulate because it keeps not liking that L23 (should be non-zero) is 0 and well there seems to be nothing i can type there to have it work... I've read the help sections a little but am still having trouble.
 
My issue is I don't see 4 CON boxes but rather CON L23 L34 L45, which doesn't match up with the images provided? How do I fill this out because I cannot get it to simulate because it keeps not liking that L23 (should be non-zero) is 0 and well there seems to be nothing i can type there to have it work... I've read the help sections a little but am still having trouble.

To specify segment 2 as having a conical flare with an axial length of 100 cm, highlight the L23 input box, type C100 or 100C, and then press the Enter key.
 
Of course one could say: "forget about these complications like saddles and constrictions, and lets just keep things simple by putting each 10" woofer into it's own 50 liter 40hz tapped pipe!" which actually looks really good in simulations but unfortunately as it turns out the slender physical dimensions of a standard 50 liter tapped pipe tuned to 40hz is a bit too small to shoehorn one of these Alpine 10s into, not ideal, magnet would probably stick out.... :( ... hmmph ...

SO LETS EXPLORE ANOTHER OPTION ;) ..... Something I came up with while experimenting with the throat chamber function in hornresponse some more last night .... It is possible to fully simulate A mass loaded tapped pipe, a sort of a mixture of an ML-TL (a widely used and trusted design) and a Tapped Pipe but the "mass loading" (a constriction) is placed in a spot that allows the upper bandpass limit to be raised giving this design more bandwidth than any typical 40hz tapped pipe! The mass loading also allows the physical line to be shorter changing the 50 liters worth of dimensions to something thats a little less toweresque and moreso cube-ish and fitting the 10" driver is now much easier :D
Upper bandwidth cutoff is now 180hz instead of 140hz ... A little bit of stuffing/lining in the first 30% of the line would smooth out the small slightly under-damped response bump at 40hz (if it is even noticeable in the real world).. Response above 50hz is only down by roughly 1db compared to these other designs I have posted here, and it is a worthwhile sacrifice for the extra bandwidth ... Design is still very simple, and no need for an access panel! Cool eh? :cool:
I just looked up that driver and it appears that it is 1.7 mH inductance, it seems to be single voice coil as well, correct me if I'm wrong...
 
Bigger Badder Lower

Here is an alternate "Grown" version using 12 inch wide panels with connector joints such that the internal width and depth are 12 inches. I wanted something that would do 35 Hz because for EDM that is basically a must. This cabinet is usable down to ~33 Hz in groups of 4 and 34-35 in singles and the simulation uses the provided 1.7 mH value (from manual), It will be interesting to prototype this and see how close it is to the sim. If anyone else has actually built this design let me know, I'd love to hear about real world measurements.
 

Attachments

  • parameters.jpg
    parameters.jpg
    99.6 KB · Views: 1,018
  • output.jpg
    output.jpg
    79.4 KB · Views: 929
Cad completed and wood cut

Just cut the wood for the sub today, takes almost exactly half a sheet of plywood. I'll be using 3/4 inch because I have 2 4x4's of it that I got for free, half inch could be used with more extensive bracing. Drivers arrive Wednesday.

I will keep posted my build details/completion/testing because I believe I'm the first to test build this design (at least in this thread).
 
Sabaspeed ,
VERY COOL ! looking at the inputs im guessing this is a scaled up version of the ML-TP (ML-TRANSFLEX) hybrid that i posted earlier , right?
Be sure to show us build pics and if possible impedance measurements and response measurements. :)
Where did you end up sourcing the SWE-10S4s from?

When it comes to designs like these I have often wondered if the small section at the end of the pipe (after the constriction) would create it's own high resonance due to the energy from the back of the cone and the small section acting as a small bandpass box (sort of like the "plenum" resonance in a PPSL) , it would define the upper response of the cabinet .... If this resonance's phase interacts well with the 3rd harmonic of the mass loaded pipe then response could be pretty loud and smooth up to 200hz and beyond if we are lucky ... Measurements should be able to tell us a lot about what is going on up there in the upper end of the boxes range ...

If we can verify that a second quasi-plenum resonance is developed up high then i suppose that would make this a type of series tuned compound pipe, perhaps David could end up working that secondary resonance detail into hornresponse's calculations in future HR versions when it comes to "constricted" transflex designs like these?

My guess is that the frequency of that high resonance would actually end up shifted higher in frequency than a standard plenum because the back of that chamber has the small opening (the 116 sq cm port) ... The area and length of that small constriction or port would have to be taken into account when trying to determine where the quasi-plenum's upper resonance and the box's upper cutoff will end up ...
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.