Suggested XO design sequence & tools – aiming high - long *

[rick57]

Hi

I’m outlining my intended project, for feedback and suggestions.

I haven’t previously built a speaker - I’ve been told this is too ambitious, but if I’m going to make something – I want it to really give me a kick!
I liked the Orion I heard a lot, and started to gets bits to build the SL Phoenix (rather than the Orion, for more headroom) and have several of the parts, but I haven’t got around to it - if I’m going to do a lot of work, I want something not only really natural sounding, but also with effortless dynamics for rock. (The Phoenix can wait)
After learning more recently about the NaO, from a dipole ‘base’ I want to gain dynamics, while minimising loss of coherence etc. Perhaps a nod towards the qualities of horn.

So I’ve done a lot of reading here & elsewhere, and accumulated these drivers:
Tw: Aurum Cantus G1
(http://www.e-speakers.com/products/ac-ribbons.html 100 dB ribbons I got for a good price
Mid: PHL 3451 10 inch (a subtle variant of the 3450, http://www.zalytron.com/pictures/phl3451.pdf 100 dB)
Bass: Peerless XLS 12 (many of you know)

The system will be active, with the same 30 w amps on Tw & M, and 200 w for each XLS driver (I have the amps too). I also have a Behringer 3 way adjustable analogue electronic XO. Tw & M have a lot of headroom.

The idea is:
The G1, on a pivot rather than a baffle, so I can fine-tune its angle up & down, reducing somewhat the lack of vertical dispersion at my typical listening position height. Although much listening is right off axis. This will go down to c. 1200 Hz with a 4th order, but per AR2 I’ll try it 1.2 – 2.2 k.
* Does the tweeter need to be on a baffle?

The 10 inch mid in a dipole.
More of the benefits of a dipole are <500 Hz, but I’ve read favourable comments about this driver in a dipole. If it doesn’t work for some reason, I will put in a box. In fact I might build a crude & quick one of each (sealed box & dipole), run mono, and see which I prefer.
I’ve read that a U baffle can make speakers sound honky or phasey. I’ll maybe do a baffle with fold out wings on hinges, to try them U shaped (maybe just for SAF when not lessening) or with a flat baffle.
Modelled in Unibox, from the max level at 600Hz, whether sealed or vented, the 3451 is down 6 dB at the intended lower XO of 200 Hz – and that’s before dipole roll off. I could push the bass up to 250, but not more.
* So - use a shelving filter/ EQ/ Linkwitz transform to balance this and/ or just use a lower slope on the bass driver’s roll off, to ‘reinforce’ the mid’s output?

Bass: Two NaO http://www.geocities.com/kreskovs/NaO.html U frame dipole woofers a side, using Peerless XLS 12 with EQ.
Learning not long ago of the higher efficiency of U frames, and there tuneablity, encouraged me to combine this type of quasi dipole bass (the cleanest bass I ever heard was the Orion’s XLS 10 bass) with the high output G1 & 3451. I’ve been told by the Peerless maker’s technical people that with a steeper slope (although few apparently do) the XLS 12 can be crossed around 200 Hz.

To ease the load at the low end on these dipoles trying to keep up with the super efficient Tw & M, I’ll probably do a sealed sub a side, XO about 50 Hz. (making it a 4 way!)

The hardest thing will be getting a smooth response - the XO and in particular EQ.
JohninCR tuned his dipole systems by ear with pink noise & music. With the rolloffs I expect in both lower mids and low bass, plus driver anomalies and frame resonances, it may be a lot easier to use an auto EQ (eg the Behringer 1024/ DQX2496) on the mids and/ or bass.
I envisage that the sequence would then be: CD – pre – XO
XO - Tw
XO - Linkwitz transform – EQ – Mids
XO - XO - Linkwitz transform – EQ – U frame dipole woofers
XO - Linkwitz transform – EQ – sealed sub

* I think that none of those affordable Behringer type devices do a Linkwitz transform (LT) or shelving filters – I’d probably need either them as well??
* How to feed the correct parameters into the LT spreadsheet – measure with a Behringer mike? Will Speaker Workshop be capable of the rest??

** Suggestions to get a smooth response without an overlong/ dead end series of trials are welcome!

Thank you!~

[rick57]

Any suggestions at all are welcome ~

[ThorstenL]

Konnichiwa,

Originally posted by rick57
I’m outlining my intended project, for feedback and suggestions.

Allow me to keep my suggestions general.

Given that you have a load of different drivers and amp's and already have decided on an active system I'd suggest make the whole shooting match modular, with the ability to slot in place various Mid/Hi Modules and various LF modules; possibly with only a baffle adapter removed and with a larger part of the tructure finished nicely and from good materials.

Then fit coarse (3db or more per setp) attenuators to the Amplifiers with enough range to offset around +/-18db driver sensitivity and adjust the attenuators to work well with a Behringer DCX2496 digital X-Over (needs modding of course).

Make sure that you can measure the system response (PC) and that you can effecively programm the X-Over.

Then you can spend the next year or two lsitenening to all sorts of different driver combos in a level playing field and the ability to get a given frequency response out of the lot. You may be able to buy more drivers and if you do not like them to re-sell them only with modest loss and you should be eventually able to develop a system that meets your requirements to a "T".

Then you start transfering the Digital X-Over curves into analogue circuitry and optimise the Amplifiers. The end will likely be a system that ius highly optimised to your taste and quite astonishing with a lot less effort and money most of us have invested. Anyway, my comments.

Sayonara

[AR2]

I like Kuei Yang Wang's comments. With finished system three or four way, there are so many variables that you will never stop tweeking. Here are few findings after a year of listening and tweeking of my system:

If you are planning to cross ribbon so low as 1.2 get some ribbon replacements. They are very fragile, and they are easily destroyed.

I reached nirvana by replacing small tube amp ( ???) I had in place for ribbons with GC. PHL is still driven by Cary audio 2A3. It is absolutely perfect match. There is clarity, but not anoying one typical with regular tweerers. There is no harshnes just beautiful clear pinpoint accuracy.

PHL and Aurum are dream match. I am not completely happy with Lambda and PHL. It is hard to seamlesly stich these two guys. Normally - it is 87dB and 101 dB. Lambda could be safely crossed even at 500Hz. I find that my original crossover point of 225 Hz doesnt work as well as 290 or 300 Hz. Lambda needs a lot of power to match PHL and even 500 W is not overkill. I have 200W mosfet for it, but my 300 W tube amp does much better job.

I would be very interested to hear your PHL in dipol configuration.
I am glad that you are inspired by my project, just like I was with projects find on DIY Audio.

Good luck, and post some pictures.
AR2

[rick57]

Thanks guys, you've clarified a few things.

My guess is that these may be steps involved:

1 Build 3 way boxes & stuffing
2 Mount drivers
3 Install amps & adjustable XO
4 Listen - tone CD + music
5 Compare mid sealed/ open
6 Adjust XO Hz
7 Compare to a reference Spendors/ large bookshelf/ Phoenix?
8 Measure RT60
9 Try different measuremnts (on axis, power, transient response, ?)
10 Measure - from 4?m - with SW??
11 Loop steps 4 - 7 & 10
12 Build LT for bass, or JK's NaO EQ
13 Listen
14 Build LT for mids
15 Listen
16 Assess & buy Behringer/ driverack EQs
17 Try different amps
18 Tweak as 3 way
19 Build subs
20 Listen in different posiitions
21 Tweak as 4 way
22 Compare to reference Spendors/ large bookshelf?
23 Tweak again
24 transfer digital XO curves into analogue circuits

Mmmm, I' thirsty

[rick57]

Can Speaker Workshop do all of the software-measurement parts above - or would it be easier with a commercial package like SoundEasy?

[AR2]

I never been abble to use fully and easily Speaker Workshop. It was always tricky and not very cosistant. I got Sound Easy and it is smood. Little bit of learning curve...but prety good and consistant. Even more, it will tranfer some settings (not all) and communicate with Behringer DCX 2496. My biggest disandvantage with SW was not having good manual.
This could be just my experience. Many other people claiming great succes with Speaker Workshop.
AR2

[JMB]

Go to www.audiodiycentral.com for a Manual. With the help of several folks much more adept than I, it may provide many of those wishing to use a very accurate and feature packed piece of freeware the shortcut from the steep learning curve that has been it's thorn for years.

Good luck and I hope that it helps,

Jay

p.s. I don't believe that it will do an RT60 but you could download one of several other pieces of freeware such as TrueRTA that I believe will do what you are trying to do.

[rick57]

Any comments on the completeness or sequence of the steps at Post 5 (above)?

[ThorstenL]

Konnichiwa

Originally posted by rick57
Any comments on the completeness or sequence of the steps at Post 5 (above)?

Yes. I am not sure you are going about this the right way.

To repeat my suggestions:

1) Build the basic mechanical foundation that will be acceptable in the living room, look good and where you can add in several possible baffles with varied drivers, making sure the largest drivers you want to use can be fitted.

2) Get the Amplifiers and Digital EQ set up to work well and make sure you can reliably measure the various possible responses of the ssytem and that you can set up the used digital X-Over/EQ correctly to achieve the desired response (which may not be a flat semi-anechonical one at all BTW).

3) Fit the various Driver combos and screw around withe the X-Over settings untill you have tested all combinations and drivers you have / want to test and all the possible X-Over settings or untill things are "good enough",

4) Transfer the digital X-overs transfer function to a really high quality analog X-Over (eg not one using a ton of Op-Amp's).

5) Experiment with better amplifiers.

The point is, unless the "experimental rig" can reside in your main listening room and you can do a lot of listening with it, long term, wide range of music you will not be able to really ***** well what you are listening to. The Digital X-Over & EQ give you a pretty much instant ability to adjust.

Ciao T