TY

It's proving really hard to drive out the ripple from the models. I'm actually seeing similar amounts of ripple in the bandpass-style systems as I am for short horns. In general, it's because the outlets can't be big enough to match the airload in such a small cabinet. Consequently, whatever acoustic system is in the cabinet, its reactance isn't sufficiently damped. Compromises, again. At least I can improve considerably on the TX.

I've looked at all the 18 sound horns for the HF, and can't find a 2-inch with 90 degree horizontal spread. I'll probably go for the P audio 3223 ( www.p-audio.co.uk/products/db_pr ... h-3223.htm ) which is 90*55 to 500Hz and has no diffraction slot. It looks rather similar to the BMS 2230 ( bmsspeakers.com/cone-speaker-com ... ide-intro/ ) rated to 700Hz. Too similar. So if I believe BMS's figure I should cross above 700Hz but I'm now looking at mid designs that should be OK up to 700Hz. There won't be much overlap though, so careful fiddling with crossover slopes/phasing will be needed.

In terms of mid drivers, it's presently a toss-up between the current 2*Delta 10 or a single P Audio SN 12-B, which has 300Hz EBP, 100dB and MMS=48, all pointing to good midrange capability. Which may be why it's the driver used in the Porn Horn.

Latest on the TY project:

I've updated the models to incorporate baffle step effects and a rough approximation of the dual sources (as seen in the TX cabs). Baffle step is a big deal - it can seriously reduce efficiency in the lower part of the range. Also, the low radiation impedance tends to increase ripple.

I'm now looking at 2*B&C 8PE21 (EBP=450Hz) in a 10th-order bandpass for each cabinet. I won't release plans until they've been built and measured due to the high risk with such a novel design concept. But if it works it'll get 155 to 690 Hz -3dB, with 106.5dB average sensitivity in band and 400 watts. That should be able to keep up with the BMS4590, though it may leave my subs behind!

Here's the result of an Akabak model of the TY. This isn't final, because I have to extend the akabak model to include more mechanical details and make sure everything fits in the box. Design is about 44ish litres at present (including HF), still based on 2*8pe21.

Just in case you thought no-one was paying any attention to this....

My personal opinion is that you've reduced the bandwidth of the low mid section too far. In order to get a clean crossover to the sections above and below it, you need at least half an octave of extension outside of the intended crossover points. I think you'll end up with some 'holes' in the response, audibly even if you manage to smooth them out according to the response plots.

Hi

It's not too bad. My subs get up to 200Hz, so I can overlap them a fair bit (I've got the low end down a little since that plot). The TY won't go any higher at the top, it's inherent to the bandpass design. But I could run my comps down to 600 or even 500 at the expense of a little distortion. I'm also planning to model the midbass, comp and xovers all together with phases and figure out the settings using that model.

The only way to know for sure is to build them and have a listen, and I'm going to do that in March.

By the way, I've switched to the 18 sound 8-inch, because it has a lower inductance thanks to a shorting ring. Compare the impedance graphs. Turns out inductance is as big a factor as cone mass.

Cheers, John

I think I've found my phase plugs for the TY:
www.allianceonline.co.uk/crocker ... fc036.html

I'm going to be off the forum for about a week - when I get back, it'll be build time!

A quick update:

I have all the parts now (cheers Chris) and have finalised the design. No rear-chamber port, and the C chamber is now a little longer so that its second pipe resonance (the 3/4-wave resonance) is coming in at about 850Hz and helping to extend the upper -3dB from 700 to 800Hz. The system is then virtually silent up to 1.8K, allowing me slack in the xover.

I'm now running a system where I use my own software to calculate and generate an Akabak script from overall cabinet dimensions and some key internal dimensions. So it's figuring out lengths and areas and accounting for port end corrections as well as the desires radiation model. I'm in the process of getting the software to generate a cutting list from the same basic dimensions.

Construction phase has begun!

Outer shells, comp/horn mounts and top-hat have been built!

Still tweaking the models though. I've made an 11th-hour decision to drop the mid-band sensitivity from 105.5 to 104 in exchange for more output at the bottom end. No point in being super efficient in the mid if I'm having to EQ up the bottom end and consequently putting all my power into the bottom end. This change will spread the power more evenly.

Remember, I'm doing all my modeling with a harsh baffle-step effect factored in. Without baffle step (ie into infonite baffle) the TY is something like 106 flat down to 150, which is crazy.

Keep it coming! It's certainly going to be a unique design.

yup, I'm very curious, Pics Pics please smiley2