[Luigi]

I will nevertheless re-try the backwards-facing FT17H horn tweeters, because I recall how it improved things in terms of imaging and staging on my last OB speakers with wings. The reason I removed them was to try to get to the bottom of the head-in-vice imaging, which turned out to be the wings.

I cannot see how Neo3s would work in this situation, unless the B200 is shelved back to around 90dB and similarly the Alphas. Im guessing that the crossovers with resistors and whatnot will achieve that.

I have some Neo 3s, but have not tried them in this setting simply because of the sensitivity differences. And anyway, I like how the B200s sound crossover-free if there's a tube in the system somewheres.

[-Richard-]

Hi Luigi ~

I spoke to Danny of G&R Research today... he sells the B&G Neo 3, and his thoughts confirm your own... if one is using a resistor and inductor to shelve-down the B200's from it "natural" rise, it will also lower its sensitivity from 97db to perhaps 93db (using a 4 to 6 ohm resistor)... so Neo 3 and B200's might be a better "match" in that case... how that would affect the Alpha 15's is not clear.

However there is a post on audiocircle about tweeters, in which Danny does a shoot-out using several different tweeters including the Neo 3 and he felt the Neo 3's performance was quite special. If you are interested you can read about it here: http://www.audiocircle.com/index.php?topic=23488.0

I am beginning to re-think my Open Baffle configuration... I might now be open to trying a resistor with an inductor to "tame" the B200's in order to better integrate the use of a tweeter like the Neo 3. Danny felt that definitely firing backwards... or toward the ceiling, works best... if fired facing toward the front there are "cancelation" issues, according to him.

If I take this route I will inform everyone with a post what I observe.

With Warmest Regards ~ Richard

[fakamada]

Hi Rudolf,

One thing is to maintain dipole pattern and another is to maintain proper off axis response matching between the drivers.
If you don't want lose fullrange seamless sound you have to crossover really high ! I'd say about 8khz. At this point B200 is already beaming so you have to use very narrow firing tweeter. Back of the driver is not beaming so we should use wide dispersion tweeter.

If using tweeters back-to back like you do I'd suggest lower crossover point. Maybe 3-4khz.

There is even more interesting stuff. Direct sound and diffused (reflected) sound is percived differently by our brain. Thats what the new Orion 3 is all about. Linkwitz equalised back tweeter to have different frequency response. But one thing that is beyond designers control is reflective material of the back wall.

   

[Luigi]

Had a spare hour last night so retrieved the old FT17H horn tweeters and hooked them back into the system with a 2uf cap to high pass them at around 10kHz.

Threw on Boatman's call by Nick Cave, which happened to be in the CDP. Had not heard it in a while so not a good perspective to come from but what an amazing album this is, haunting lyrics, stuck-in-your-head melodies.

The imaging seems more corporeal with the rear tweeters back in; the performer just seems to be more real between the speakers which I guess is what reproduced music is trying to achieve.

The speaker as a whole sounds particularly good with a bit of volume on board, but detail is more there at lower level now than without the tweeter. This is readily evident when the source is rubbish, like from the Sky decoder (set-top box).

People reckon the FT17H is a bit spitty and I know what they mean listening to them up close, but as an ambience tweeter at the rear, I think they work well here. Will now play around with positioning them but I dont think it much matters where they are located when rear firing, so long as they are rear firing.

I think there is more than just added "airiness" happening here as the imaging is more real, the positioning of instruments more clearcut.

[Luigi]

I remember also Danny and others saying there would be phase issues using a full ranger like a B200 with a front firing tweeter, which is why I use the rear firing FT17H.

That said, I have never tried a forward facing tweeter with an OB set up. I have another set of 96dB tweeters hanging around so might just see what happens when I have one set forward firing, one set backwards. I guess the way to get around the phase and cancellation issues, as some have suggested, is to roll off the top end of the B200. Might be worth a stab, as Rudolf suggests, since it isn't much of an issue to at least try this. Glad I bought all those tweeters to mess with years ago!

[Rudolf]

2. What is the range where this "effect" begins... and therefore what is the cross-over point we would be looking for when implimenting the tweeter?

Richard,
I have to split the answer into different parts, since there are so many and different aspects. First let me show some front/rear measurements of different drivers in OBs:


This is from the Linkwitz Phoenix - the ScanSpeak 21W/8554.


This is the Vifa 10 BG 120 - a 3" driver


This is the Visaton FRS 8 - a 3" driver too.


This is the Monacor SPH 176 - a 7" driver.

Regardless of the dipole pattern: for every cone/basket construction there is a point, where front and back frequency responses walk apart. The back radiation will not simply be attenuated - it is worse. The back radiation will develop its own character. There is a tendency that big drivers will have that point at lower frequencies than small drivers. But it is very much depending on each individual construction. With some drivers different mounting methods (rear-front) can have a great influence too.
This does not only apply to 0° versus 180°.  The rearward frequency responses to the sides (30°, 45°, 60°) can change in a similar way because of the cavities formed by the basket, the cone and the spider.

If you simply add a rearward tweeter in parallel with the widerange driver, you only get a giant mess in the backward radiation - all kinds of phase issues included. What I am talking about mandatorily requests a real cross-over between widerange driver and tweeter at driver-dependant frequencies from 2-4 kHz.

I don't say that the B200/Alpha15 isn't a great combination in itself, but one can do better. The price is increasing system complexity and need for control of course.

[Rudolf]

3. If you are using 2 tweeters per baffle, as you suggest ("a pair of really small dome tweeters mounted back-to-back"), then the forward facing tweeter is radiating those upper frequencies (where the tweeter plays) directly into your room... in that case what "extra" thing does the back radiating tweeter do that makes it more effective, in the Open Baffle application, than just the one front facing tweeter is doing?

I've simulated two polars with tiny Dayton ND20 tweeters - left for one (monopole) tweeter and right for two in back to back fashion with one tweeter connected in reverse phase (dipole):



The polars are taken from 0°, 30°, 60°, 90° ... values only. That makes them somewhat jagged. Front is top and rear is bottom in the diagrams. You can see that only the right configuration follows the dipole figure 8 pattern. Apart from the red curve (8000 Hz), where the dipole is "blooming", the other patterns keep their position and their shape. They just grow and shrink a bit according to changing SPL.
The left polars are keeping their circular shape, but move somewhat forward with rising frequency.

Note that the right configuration is following the radiation pattern of the B 200 in OB below 2 kHz. With a monopole tweeter, the B 200 would have some trouble in the crossover region, because the radiation pattern would change from dipole to cardioid and then to monopole.

4. Also, is there an issue with a possible "cancellation" of frequencies arising from the back-to-back arrangement of the tweeters?

The sideway cancellation is a wanted part of the dipole radiation pattern. Unwanted cancellations would occur if you combine a monopole tweeter with a dipole B 200.

[Luigi]

Thanks to the postings of Rudolf, I decided to mess with tweeters front and back and glad I did, because now I seem to be getting a degree of smoothness in the midrange that the system had been missing, and some top end sparkle, with wide imaging as well. All good.

Rudolf suggested earlier that mounting two tweeters back to back gave the smoothest most natural top end and staging, and I figured it would also ameliorate the B200 upper mid shout, given I would be crossing them over at around 4kHz.

It was about that time that I remembered I had some Neo3 BG dipole tweeters somewhere. So I removed the FT17H from the rear, and simply wired these in, with an inductor to the B200. The tweeter sits right on top of the baffle just above the B200, so is radiating out in both directions.

And this is sounding better than I imagined it would, though the Neo 3s are very "neo" at present, with only a few hours on them.

These are a relatively inexpensive ($NZ70 each) wide ranging tweeter that seem to do a swell job here; they image smoothly across a very wide area, and seem precise as well. The whole speaker is now much smoother and treble sounds better delineated. In terms of sensitivity they appear to be a reasonable match with the B200s as well. Will now set about hooking them up with something a bit more exotic.

The nice part about this solution is I only needed to buy two tweeters instead of four. Better yet, I already had these lying around unused, so it feels like a free upgrade!

Thanks Rudolf for the tweeter suggestion. It's a good one.

[Rudolf]

Luigi,
thank you very much for trying my suggestion. I'm glad it worked out for you.

My own conclusions regarding loudspeaker design come to a large part from some theoretical knowledge in physics and psychoacoustics. I can only punctually verify singular aspects with my own practical experiments. Everything in between is wild, but hopefully educated, guessing on my behalf. So it is a relieve every time when someone else proofs me right (at least partly ) with his own system and components.

I'd like to add two commemts:

First about the positioning of additional tweeters:
At 4 kHz the wavelength is 8.6 cm. It is very wise to keep the tweeters as close to the midrange driver as possible to avoid phase issues, as you did.

Second about the Neo3:
Looks like the dipole tweeter of choice nowadays. For the price you can't do better. There are people who find it lacking in power. But those people tend to employ four Peerless XXLS drivers as dipole woofers (per channel).

Hope you are well in NZ these days.

Rudolf

[Luigi]

Hi Rudolf
Yes, we are well in NZ, though those at the other end of the country, in the South Island, are recovering from a big earthquake, centred near to major population centre (Christchurch), 7.1 on the Richter scale and quite shallow, 10km. The local geologists did not even realise there was a faultline at the quake epicentre; now there is evidently a crack in the earth 24km long. Some paddocks that were flat now have small hills in them.
The quake happened very early morning and amazingly no-one was killed.

Anyway, back on topic, I will post a pic of the speaker with the Neo3s added tomorrow. Yes, they are sitting very close to the B200, virtually atop them. Imaging is pretty darned good now; it no longer changes with each small head movement as it used to with the old baffle which had wings. And instruments just kind of hang and shimmer in space.

Moreover, the speaker seems more natural somehow, possibly better integrated? With the rear firing Fostex tweeter there was definitely more treble action going on, but somehow it never seemed to gel quite like this Neo3 does.

As you say, the BG units are quite cheap drivers, so for DIY are an excellent, if slightly odd-looking choice. But then with this speaker, appearance is very much number 11 on a list of 10 priorities.

Must say I'm a bit surprised the 90dB sensitivity because the dipole version works nicely in this setting but then MJK reckons the Alphas in H frame are around 90dB, and likewise the B200s with an inductor are probably down around that too.

Wonder how this measures?

[jonners]

It was about that time that I remembered I had some Neo3 BG dipole tweeters somewhere. So I removed the FT17H from the rear, and simply wired these in, with an inductor to the B200. The tweeter sits right on top of the baffle just above the B200, so is radiating out in both directions.

Hi Luigi - I think I might give this a try. Are you just using 1st order crossover? - What value L and C?
Thanks - John

[Rudolf]

Luigi,

the Boxsim simulation of the B 200 in your new baffle reveals a lot:



Black is the B 200 in infinite baffle. Official efficiency values are taken from that diagram at 1 kHz. You see how 94 dB are just reached at 1kHz. At 700 hz it is only 90 dB

Blue is the response in your narrow baffle. The dipole peak around 1 kHz emphasizes the already visible 1.4 kHz hump of the B200.

Red is the influence of the 0.6 mH inductor. Dashed red includes an added parallel 6 Ohm resisitor.

A more detailed simulation shows the response at 90 dB at 500 Hz and again at 2 kHz. In between is a hump of 5 dB:



Dashed line is the 6 Ohm resistor again.
I don't know if the hump exists in reality too, but it could be fixed with a notch filter.

The above would explain, why the B200 fits quite well with the Neo 3 efficiency.

Rudolf

[canzld]

Hi Luigi

The quake happened very early morning and amazingly no-one was killed.

putting it mildly. Thankfully all my rellies in ChCh are fine - although some of their houses not so much.

Interesting reading about your experiments with tweeters. While you've been busy going down that road, I’ve pretty much abandoned the h-frame, small baffle setup with the B200/alpha combo. It does gives a much better bass response, but in my hands anyway, it does no favors to the B200. My initial thought was that the B200 rising response was more apparent than on my previous 40x24inch wingless OB, and further listening confirms this. Mids are thin sounding (relative to previous), with piercing treble at times (and this is playing with a tube in the line). This problem could solved by swinging the h frame around and placing the small b200 baffle right to the front – in essence recreating a larger flat baffle for the B200 (16x40). IMO this improves the tone of the B200 significantly. This is very clear when taking the alphas out and just listening to the B200s.

Since a side firing h-frame isn’t great for imaging, I’ve done a quick convert to open backed box. There is some dB loss in the bass as a result I think, but the b200 is much better. Shame - i liked the h-frame bass.

One thing I noted while playing around was that perception of the rising response was very dependent upon speaker placement. Normally I have them 3 feet off the wall, which accentuated the rise significantly when using the small baffle, with 7 feet clearance it wasn't as bad,- although the opposite wall is also much closer in the 3 foot configuration, so I’m not sure which aspect is more important  .

[Luigi]

Jonners
I just calculated inductor and cap values using 4000Hz as the crossover point, and I got 0.24mH for the inductor to the B200 - and I had a 0.23mH inductor spare so used that - and cap value to tweeter to make it fire from there was around 7.0uf, IIRC.

Canzld
Interesting the different results we get from similar applications. Naturally there are just millions of variables here that affect how they sound in a given application, but you just have to go with your gut feeling sometimes. If it doesn't gel to you, then something must be wrong. And you can obviously exclude things like room and electronics because you're familiar with them.

My experience is similar to yours but different. Im getting what seems like really decent integration between alpha and b200 now. The alpha is low passed at around 300Hz, and there's next to no voice content from that, even with your ear right up to the cone.

I made sure to go by the dimensions MJK recommended for the H frame and am thrilled with the bass results. It is not overwhelmingly strong but after living with B200 bass for some time it is a revelation. Im not at all tempted to double up on the Alphas. And the articulation is something else. I feel occasionally some particular bass notes seem to be MIA - probably a room thing - and then the next song proves me wrong. So Im not messing with the H frames any more, except to get rid of that unsightly particle board.

The Neo3 tweeter solution has generated extremely good results from my perspective; the system is not only sounding reasonably coherent but is now vastly smoother in the mids with the inductor low passing the B200, starting at around 4kHz.

The tops just seem to get better as the tweeter breaks in, not that this takes long, but it is much better now than when I first inserted it a few days ago. And the imaging and staging is almost reach out and touch it real. I like that, though some don't seem to. But we're trying to create illusions here, and if the illusions seems lifelike, so much the better from my standpoint.

But what I really like is the musicality of the system; it just seems to bounce along with the music and if a song is playing and you're listening for, like, whatever regarding the character of the speaker, you kind of get lost and forget and start just actually listening to and enjoying the music. Never thought I'd be able to say that, because I have never had a speaker that does that, not ever. Love the size of the stage too; big cones mean big music, if you get my drift. And yet now the speaker isn't really in your face, except for physically. Here is how it looks now.

































































































 

[jonners]

Jonners
I just calculated inductor and cap values using 4000Hz as the crossover point, and I got 0.24mH for the inductor to the B200 - and I had a 0.23mH inductor spare so used that - and cap value to tweeter to make it fire from there was around 7.0uf, IIRC.

Thanks Luigi.
There is enough room above the B200 in my OBs for me to able to mount the Neo3 within the baffle. But do you think it would be better to lower the baffle height and mount them at the top as you have done? Presumably they come in more steeply in the latter case.

[Luigi]

Jonners. Sorry, cannot answer your question regarding whether or not they come in steeper.

 But Ive seen these mounted in a rectangular hole going right through the baffle, making them dipole. And off-axis output is minimal, so no real reason why you shouldn't be able to mount them within the baffle. They'd certainly look better! And if that doesn't work out so well, you can just saw the top of the baffle off and mount 'em on top anyways. That's probably how I'd go about it. Others might have more experience with Neo3s here.

[versus rider]

Hi Luigi, I'm surprised you can cross your alpha's at 300Hz, mine sound dreadful at 300Hz, I am using a digital crossover and crossing mine at 120Hz and they are sublime. In fact I wouldn't swap my OB woofers for a box speaker ever. I can adjust the crossover point with the digital crossover but at 300Hz they sound awful, which is a shame as I would like to try a midrange spherical horn but that only goes down to 300Hz and I couldn't live with my woofers at that frequency.

[jonners]

But Ive seen these mounted in a rectangular hole going right through the baffle, making them dipole. And off-axis output is minimal, so no real reason why you shouldn't be able to mount them within the baffle. They'd certainly look better! And if that doesn't work out so well, you can just saw the top of the baffle off and mount 'em on top anyways. That's probably how I'd go about it. Others might have more experience with Neo3s here.

I've just been reading Rudolf's paper. I haven't studied it in detail yet, but he does state: "...the effective diameter of a baffle should not exceed double the diameter of the cone". So it looks like it will be best to reduce the baffle height and mount the Neo3s on top.
Rudolf - can you comment please?

[Rudolf]

I've just been reading Rudolf's paper. I haven't studied it in detail yet, but he does state: "...the effective diameter of a baffle should not exceed double the diameter of the cone". So it looks like it will be best to reduce the baffle height and mount the Neo3s on top.

 
Well, I should (and will) extend above statement with the addendum "... double the diameter of the cone, if no compromises are acceptable." Later on in that chapter I do explain how a compromise could look.
 
Why do we need a baffle in the first place? It helps the driver to play low frequencies - that's all. So if there is no baffle, the Neo3 is not supported at low frequencies. It comes in more steeply and at a higher frequency. What we gain in return is the most controlled off axis directivity and the least edge diffraction possible with that driver. 

If we need the low frequency extension of a baffle, we have to compromise above values. The worst compromise would be to position the Neo3 far away from any baffle edge. The most acceptable compromise for the Neo3/B 200 combi would be to put the Neo3 just above the B200 in a shared baffle and to cut that baffle just above the Neo3.  I you can stay away from making saw dust for another day or two, I will try to verify my above predication with a Boxsim simulation of the Neo3/B200 combi.

[jonners]

 
Well, I should (and will) extend above statement with the addendum "... double the diameter of the cone, if no compromises are acceptable." Later on in that chapter I do explain how a compromise could look.
 
Why do we need a baffle in the first place? It helps the driver to play low frequencies - that's all. So if there is no baffle, the Neo3 is not supported at low frequencies. It comes in more steeply and at a higher frequency. What we gain in return is the most controlled off axis directivity and the least edge diffraction possible with that driver. 

If we need the low frequency extension of a baffle, we have to compromise above values. The worst compromise would be to position the Neo3 far away from any baffle edge. The most acceptable compromise for the Neo3/B 200 combi would be to put the Neo3 just above the B200 in a shared baffle and to cut that baffle just above the Neo3.  I you can stay away from making saw dust for another day or two, I will try to verify my above predication with a Boxsim simulation of the Neo3/B200 combi.

Thanks Rudolf, I will study the rest of your paper tonight.
Sawdust-making is not imminent, since I haven't got the Neo3s yet!