THAM18 - 18" tapped horn
Printed From: Speakerplans.com
Category: Plans
Forum Name: Other plans
Forum Description: Discussion / Questions about all the other plans
URL: https://forum.speakerplans.com/forum_posts.asp?TID=69189
Printed Date: 27 March 2026 at 2:15am
Software Version: Web Wiz Forums 12.08 - https://www.webwizforums.com
Topic: THAM18 - 18" tapped horn
Posted By: martinsson
Subject: THAM18 - 18" tapped horn
Date Posted: 22 July 2012 at 11:24am
|
Since the THAM15 was made public back in the day alot of questions regarding an 18" version has been popping up in the mail and on forums, and even if not acting on this with haste (to put it mildly) the thought has allways been there and more then a few sketches on the matter has seen the light of day. Johannes Rodin, the guy who has been involved in all the THAM designs to date, and myself finally got around to putting the pices thogether for the THAM18 and we now got to the point where we feel it's ok to go public with the design, bare in mind that not all the details have been worked out yet, but the basic design (i.e general folding) is now mature enough to be shown, more detailed documentation can be found on the link below. There is a build planned in here in Sweden that hopefully will show if the correlation between the simulation and folding layout is good enough, and (more importantly) if the end result is good enough, but as far as theory and expirience goes we belive in this (so far) "paper product" enough to go public with it, not making any promises in process. As usual, if you want make your own version based on this basic design then by all means go right ahead, it's free and based on open development, alterations and improvements are encouraged and wellcome, regard this as a platform, build it as is or take it further, it up to you. Also, I'm halfway drowning in "will this driver work in THAMxx" questions, so please note that I will not answer these questions, or I would be doing little else, no disrespect intended, but please give it try yourself, all the data you need is allready available, as is the excellent software hornresp, simulations of other drivers will be shown as time progresses. This is the THAM18 development proposal, more detailed and refined documentation can be found on the link below.  Using the FaitalPro 18HP1060 as default driver :  The standard 2PI spl response based on above input :  More information can be found here (more refined vesrion : http://www.martinsson.cc/blog/index.php?entry=THAM18" rel="nofollow - http://www.martinsson.cc/blog/index.php?entry=THAM18 ------------- Swedish Pro-fi diy-nerd - http://www.martinsson.cc/blog/" rel="nofollow - http://www.martinsson.cc/blog/ |
Replies:
Posted By: Teunos
Date Posted: 22 July 2012 at 12:48pm
|
Looks good, but the sim shows internal volume of 395 liters. I redid the whole process of the sim, this is what i came up with. First of, the sketchup model with underneath it the hornpath folded up and the unfolded hornpath.  The sim input and output with your driver (and i've been kind on the section where the negative expansion is) and that even excludes the volume of the driver in the mouth of the horn.  And a comparison with your first sim.  ------------- Best regards, Teun. |
Posted By: Teunos
Date Posted: 22 July 2012 at 12:54pm
|
And even my sim is very kind on the design, true internal volume is only 285 liters including driver, and my sim is even 311 liters. Btw, i'm not trying to critisize your design, but having experience building multiple THs before, I'd look into it before building the cab. ------------- Best regards, Teun. |
Posted By: martinsson
Date Posted: 22 July 2012 at 3:44pm
|
Hello Teunos Thank you taking an intrest in the design and for investigating it, but I'm struggling to understand a few of the conclusions you come to, not saying they are wrong, just not clear to me. Depending on how you choose to route the L23 path in the corners you will alter the system volume, if you chose the middle path instead the volume will change as well, in your description above the corner volumes are taken away completly and yet the sound will be influenced by these areas, granted I'm not sure exactly how, but since they are there it makes more sence to me to include them. I belive that your method reduces the true internal system volume of any perpendicular folding by default, in this case you show a simulated sum length decreased by 21.5cm (approx 10%) compared to the method I use :  Again being careful saying I'm right about this, I simply choose the model that I think describes reality in the best way, the one that makes most sence to me, and so do you, and I confess to lacking the knowledge to state which way is more right.
There are 6 places in the hornpath where you get localized negative expansions due to the folding, all of which are neigh on inevetable if trying to keep the design fairly simple, to which one do you refer ?
But you do, and you should, critisism is neccesary, and to question is a must, all I can do is to show how I arrived here without claiming anything, the THAM18 is designed and simulated in the same way as the THAM6, 10, 12, and 15, and I hope the THAM18 shows the corresponding step up in preformance compared to the others, and if so - job done ! ------------- Swedish Pro-fi diy-nerd - http://www.martinsson.cc/blog/" rel="nofollow - http://www.martinsson.cc/blog/ |
and i've been kind on the section where the negative expansion isPosted By: malina
Date Posted: 22 July 2012 at 6:54pm
|
Interesting. I would realy like to hear the results and maybe some comparison to THAM15. Size/output, suited for live bands or just playback etc? Keep up the good work. ------------- Malina Audio |
Posted By: Teunos
Date Posted: 22 July 2012 at 7:29pm
|
I noticed when simming it with your input screen that the internal volume of the simulation is 395 liters. Upon modelling the cab in sketchup i couln't believe this was correct. The surface area of the side cut is only 5069 sq. cm which equates to 285 liters of internal volume, so simply put there is a free 100 liters in your sim. Even in mine there is around 30 free liters, and that excludes the driver volume.  The parts of the horn of which i think we can both say there is no question how to measure it's length i indicated with yellow red and blue  All the green parts are under inspection.
Which of our methods is more correct is decided by the theory of fluid dynamics. Here's my attempt at a very crude explanation. Imagine a tunnel, and for ease assume it's perfectly cubical(so disregard the 3% expansion) with a 90 degree bend at the end expanding into an also cubical tunnel. Assuming the flow in the tunnel is not fully turbulent so has a low reynolds number, down at the convex part of the corner there will be a relatively high pressure which creates slowly propagating eddies curling in the bend. On the other hand at the concave end of the bend , where the air pressure is low (due to the corner being there) sound will be deflected towards the curvature of the bend. What this basically means is that the air closest to the concave part of the bend gets sucked into the bend, and the average path length the air ,or sound, ''sees'' is not the average path length as indicated in your graphic in the bend but is actually smaller than that. That is why i always estimate the average length of the corner using a circle as seen in my graphics or as you say disregarding a part of the bend. Although all the above is air pressure and velocity related {and therefore soundpressure related}, (the effect reducing with increasing air pressure), the principle will be valid upto crazy high sound pressure due to the low density of air (apply this theory on water and it fails miserably above relatively low speeds) [Here's an example of an article we studied in class http://www.cfd.com.au/cfd_conf09/PDFs/149MOS.pdf" rel="nofollow - http://www.cfd.com.au/cfd_conf09/PDFs/149MOS.pdf please stop at results, it gets quite boring after that part....]
I think the folds are pretty neat all the way through, the only real trouble area is where there is fast expansion, compression, expansion and then compression again all in parts where reflections back down the horn are most likely to occur.  Don't think it's worth the effort complicating the design putting corner reflectors in anywhere except where the S curve folds back down to the left again in the upper right corner. Having said all that, i'll try to fire up a accurate akabak model simming each individual piece in the horn when i get the xp machine running again (akabak doesnt run on win7). ------------- Best regards, Teun. |
Posted By: martinsson
Date Posted: 22 July 2012 at 10:20pm
|
Thanks for clarifying, you sure put alot of work into this, and for wich I'm very grateful, this is why I like forums like these, I have learned alot here and in other forums, it is by far the best education one can get outside academia, and I still learn alot, like in this case. The main reason for the discrepancy in volume comes with the folding, this is not to be argued, the same is true for all the THAM designs to date, the hornresp simulations I made shows a nice even expansion which is not the case with this folding, there in lies the biggest part of the missing volume, this is true. The smaller parts of the discrepency are in the corners, and seeing as I now lack the knowledge of bernoulli equations for advanced fluid mechanics I was wrong there as well, this is also true. All the designs I have been involved in to date, the THAM6, 10, 12, 15 and now 18, years of work, all follow the same flawed logic, and I can not and will not dismiss your theoretical proof and resoning, it is as far as I can understand nothing short of correct. Designing a folded TH is almost the very definition of compromise, the trick is to balance these compromises in such a way that the end result is usefull and still shows an overall gain over other principles given the same volumetric restrictions and number of drivers, and with this in mind I still belive this is a good design. Without restrictions and constraints there is no need for compromises, and I belive there is no practical use for a uncompromised design out there, infact I don't belive such a design even exisits. ------------- Swedish Pro-fi diy-nerd - http://www.martinsson.cc/blog/" rel="nofollow - http://www.martinsson.cc/blog/ |
Posted By: mykey-
Date Posted: 23 July 2012 at 6:47am
|
I looked at the SPL chart on the other thread that Matt posted and thought that there is no way that shortish fold is going to achieve that output. And to skip from what you have done to what Tuenos has put up is quite a difference in design size If you've been doing it long enough you can just look at a sketch and see what the outcome would be. ------------- BbbBBRAAAAPppBBBBbgushhhhhhhhssshhhhhGrAbRAAAAAAPPPPPp = Dubstep |
Posted By: martinsson
Date Posted: 24 July 2012 at 9:22pm
|
I feel it might be time to try to describe how we got here, and why the folding and proportions are the way they are, what compromises has to be done and why. Here are the set of challanges we needed to sort out if using the THAM folding and my simplified way of simulating : 1. / Kepping the L23 path length in the area of 190~220 is required if trying to stay within the desired passband for PA use, wich usually only needs to reach down to around 40Hz, this is not easily achived when designing with 18" drivers in mind, the box grows, and so does L23. 2. / When using larger drivers the realtive cone stiffnes is most often lower, this calles for a larger the normal S2 in relation to Sd, around or preferably below a Sd/2 was the aim, given the expansion this then drives the box size, wich impacts negativly on point 1. above. 3. The expansion, and this is where we see the differance described above, in THAM15 the expansion is a fixed 3deg up until the coupling chamber in the front, and this is where all the THAM designs have a rapid expansion, hence the term "coupling chamber", and the THAM18 is no exception. Why not simply increase the expansion angle or "step" it up ? the answer again is simply - cabinet size (volume), the seemingly minute differance between 3deg and 4deg rotation would amount to undesired volume increase and/or absurd box proportions given the rather generous S2. These limitations is there due to the fact that the design should be kept simple, the internal structure should simply be rotated, leaving odd cutting angles out of the equation, leaving more potential DIY'ers happy, this is a key element of the THAM designs, they are simple to build. 4. / Now we arrive at the S3, and in the case using of larger drivers such as 18" this area needs to grow significantly, the base value to shoot for here is roughly Sdx2, this also inflicts upon the layout in a negative way in such a way that it (again) drives the L23 length, and thereby tuning, beyond (below) the desired 40Hz and more the that it also reduces overall sensitivity and adds volume, this is the reson for the "cupling chamber" described above. 5. / After this we find the S4, and in order to keep the level up above the passband this needs to be even larger, without adding to much L34 length, and this is why we find the rather rapid end expansion of 60deg. The top front dead volume is there in order to keep L23 down, keeping resonable cabinet porportions and adding S4, it might also make a nice palce for handles and adds to box stiffnes in the upper front end. All of these parts must then be balanced, put together and folded inside a standard box-like volume with easy to handle outer proportions and dimensions and with THAM like folding inside. This is how we arrived here, and as for the theroy/reality ratio, well if the previous designs are something to go by this seems to be a good way to take the problem on, not saying that it could not be done better, but it is not arrived at lightly, the time spent on the balancing of compromises in the case of the THAM18 are well above any other THAM design to date, the result however remains to be seen. The summary is, if these compromises are not made and balanced in a good way in keeping to the "normal sized and easy to handle" box proportions then there is no practical need for the design in the first place. No promises, no predicition, no demands or claims, just good times spent doing what we think and hope will benefit diy'ers out there, that's all. ------------- Swedish Pro-fi diy-nerd - http://www.martinsson.cc/blog/" rel="nofollow - http://www.martinsson.cc/blog/ |
Posted By: mobiele eenheid
Date Posted: 24 July 2012 at 10:24pm
|
Hi Martinsson, I think you have to think more like an air molecule and less like a human.
Imagine that your surrounded by your fellow molecules, way above your boiling point so your trying hard to keep your private space. Behind you is an immens crowd that has a steady and impressive forward and backworth movement, limiting you movement, pushing you in their direction.
I think, that most particles 129.7 mm after S3 just keep on going forward, not suddenly all take a big step upwards (only the air molecules closest to that 60 degree expansion will). In that way only a fraction of the totall, experiences that expansion.
Basically saying that 34.84 cm might be a bit optimistic in (in relation to that S3 and S4 anyway) and a large contributor to the higher internal volume in the sim.
Based upon on ather design I say the internal volume is in the 280 - 290 liter ballpark. Having said that, I do think the frequency response as presented in Hornresp is feasible in that internal volume.
@Teunos: I always was under the impression that bends actually lengthen the air path, not shorten it. How would you go about calculating Reynolds number in this situation? I assumed that the viscosity of air might be a bit to low to be reprersentable for fluid dynamics (and might insert complexities as well).
Toevallig een paar weken geleden een tentamen over kolloidkunde gehad, vandaar de interesse.
Best regards Johan
|
Posted By: Teunos
Date Posted: 02 August 2012 at 11:45am
Disregarding which starting point of measuring corner length is more accurate. Having simulated manycabs I can understand how a sim could indicate slightly larger/smaller volume due to above discussion and the slight differences in volume of the individual horn segments depending on where a certain Sx or path length is chosen. However, there is a large discrepency between the volume in the original simulation of 395 Liters and the true cab volume of only 285 Liters. I also know a TH can have real measurable lower frequency extension then the simulation indicates, having built THs and having been surprised as well, altough i'm confident my simulation was about spot on as perfect as it could have been. Personal opinion regarding tonal character, not frequency response; But I also still think the transparency and level of detail (although impulse responses indicate otherwis) of a TH gives that impression of even more bottom end presence whilst not being there in reality when measuring the frequency response. However i feel the most confident about a build when the sim approaches the real life cabinet at its best, and nobody is going to be able to make a cab that sims as a x liters cab, build it 100 liters smaller and then have the same measured response as simmed. And if it does, you either are very lucky or messed up the original simulation completely. My only advice to Mr andersson is to review his own simulation with a critical point of view, i know I would do so if there was a 100l difference. Back to you Johan, the final expansion near the horn mouth does add sensitivity to the output. In my 18'' TH i also have a expansion similar to this on the THAM18. I also wondered whether it would make a big difference so a single panel was cut to fit in and basically extend the speaker baffle making it a single panel. Although i could not directly hear any difference, Smaart showed almost the same Freq. response but there was a slightly step down in sensitivity of around 1-2db centered around 70Hz with a narrow Q. Please correct me if i'm wrong; About the reynolds number, it will indeed be problematic at least to calculate this for air, especially under fast changing air pressures in loudspeakers and probably it would be better suited to apply the theorem of aerodynamics instead of general fluid dynamics. My reasoning is based on logic and knowledge i have so far and although there are probably more turbulent effects in the flow of air than there are predictable, i've always been under the impression that air is more than likely to behave comparable in closed systems to fluids of low density. Misschien wanneer we volgend jaar vloeistofdynamica en fysische transportverschijnselen krijgen dat ik het hopelijk zelf ook allemaal wat beter snap, tot nu toe is het meer inleidend geweest allemaal. Studeer je Biomedische wetenschappen of iets dergelijks? ------------- Best regards, Teun. |
Posted By: martinsson
Date Posted: 28 November 2012 at 7:30pm
|
I did a simple study of how the folding impacts the simluated result, this new version of the THAM folding came up in the THAM18 designprocess. http://www.martinsson.cc/blog/index.php?entry=THAM18---new-and-traditional-THAM-folding-compared" rel="nofollow - http://www.martinsson.cc/blog/index.php?entry=THAM18---new-and-traditional-THAM-folding-compared ------------- Swedish Pro-fi diy-nerd - http://www.martinsson.cc/blog/" rel="nofollow - http://www.martinsson.cc/blog/ |
Posted By: martinsson
Date Posted: 12 December 2012 at 10:55am
|
More detailed documentation (drawings and sims) can be found here :
http://www.martinsson.cc/blog/index.php?entry=THAM18-details" rel="nofollow -
http://www.martinsson.cc/blog/index.php?entry=THAM18-details ------------- Swedish Pro-fi diy-nerd - http://www.martinsson.cc/blog/" rel="nofollow - http://www.martinsson.cc/blog/ |
Posted By: login4
Date Posted: 12 December 2012 at 10:57am
|
thank you very much, as a tapped horn ill enjoy a read over it ------------- CELTIC SUBSONIC SOUND SYSTEMS |