THAM15 - a compact 15" tapped horn

Your sim is missing cone correction. Vtc=5000 and Atc=855 should be in the ballpark.

I can put that in sims, thanks for info. I practice, there is no audible peak, and woofer can go little lower than predicted.
I have stopped using this drivers, now i use drivers with 4 inch coils which i can drive with more power. Also recommend digital crossover, i use DCX2496, upgrade to this from analog crossover is very noticeable.
 
With (i hope) correct parameter simulation is even better (goes tad lower and flatter).
 

Attachments

  • Tham ps761a.PNG
    Tham ps761a.PNG
    40.5 KB · Views: 840
  • Tham ps762a.PNG
    Tham ps762a.PNG
    54.3 KB · Views: 839
Other drivers?

So I know I'm trying to revive an old thread, but I'm wondering if these cabs would work well with the Jbl 2226h low frequency transducers? I have quite a few laying around and I want to put them to use. If anyone could give some insight onto what might be a good cabinet to use these in,ant info would help! Thanks!
 
The standard version THAM15, the only one with propper drawings published is the one, but feel free to modify it as you see fit, there has been numreous proposals on how to improve their prefomrance, such as aperture chamfer, cone compensators, and various reflector proposals.

I have not looked into all of these myself, but scan the forums and ask around if you feel up to it, I'm sure you will find alot of inspiration out there.

Most people seem very happy leving it as it is, the most common modification is only adding a bit of depth to the box in order to fit a full front cover, this does not impact on or alter the design.

Hi Martinsson! Many thanks for all your work. I was all set to build a standard version but then noticed a few customisations people had made with handles etc. In particular I like the look of MagicJohansson's build seen earlier in this thread but I just wanted to ask, since it seems the basic measurements they used are the same as the original plans, should I be worried about the impact of the recessed handles, braces etc on the dimensions of the horn and the resultant sound?

Thanks
 
So I know I'm trying to revive an old thread, but I'm wondering if these cabs would work well with the Jbl 2226h low frequency transducers? I have quite a few laying around and I want to put them to use. If anyone could give some insight onto what might be a good cabinet to use these in,ant info would help! Thanks!
Hi a bit late, I modeled 2226 a while ago and it seemed to fit quite well in the tham15, I have built 4 now with the jbl2226h drivers and love them. they really pack a punch and sound good. Greg
 
Nice to know they are compatible with what I would choose to call a very nicely balanced and common driver.

I did not want to design a cabinet that puts extreme demands on the driver, I never have, the main reason being that if it was to take off as a world wide popular DIY design there is little to no control over what drivers where going to be used with, and it should work OK regardless.

This can be seen in the rather generous S2 area (the amount of air cone side) an idea that came from Johannes Rodin, this puts less stress on the driver in general and it seems to have payed of since I have yet to hear of any design related driver failures (or any other failures for that matter) in the any of the THAM designs.

Today the 15TBX100 is a good average performance driver and there are far more extreme drivers out there, the trend seems to be going towards lower Q and higher BL, and it will only improve the performance further.
 
The 4 segment simulation, the top one, on which all the resulting graphs are based is a more correct way of simulating the folding since it takes the rather steep area increase between S3 and S4 into account.

The drawback then becomes that due to the limited amounts of segments allowed in hornresp, the last expansion is not simulated correctly (between S4 to S5) where it does not see the step and regards it as an even expansion between S4 and S5, it's not a big error and I had to choose and that seemed to be the least inaccurate way of doing it.

The 3 segment simulation at the bottom should really be taken away since it shows a overly optimistic result, the reason being that it does not see the S3-S4 expansion, instead it sees the S2 to S3 as one evenly expanding segment, which is false, this also becomes apparent in the systems volume window.

I hope this clears thins up a bit, and thanks for pointing this out, I will correct the post, emphasizing that the 3 segment simulation is not the correct way to simulate this folding.
 
Last edited: