Build Thread -- Newbie modeling an hframe - open back mid

[Rudolf]

Rudolf, I'm not sure I buy that definition of Leffective for defining the TL peak, at least not for very shallow wings. If the baffle were 24" wide and the wings were 1", Leffective would be a bit over 12", but we know it should be much closer to 1" than 12".

The definition of Leffective is not mine. It´s part of MJKs work. And I don´t know if he would change it for very shallow wings. But I believe he would not have to do so much. My explanation why:

T-Line is a word for any acoustic design, where the radiation impedance changes sharply from a defined constant value to the impedance of full space (free air). If it would change smoothly, it would be called a horn.

Even without wings the driver on an open baffle "sees" a radiation impedance of "half-space" up to the edge of the baffle, where it abruptly changes to full space. My conclusion: Even if there is no "length" of the TL, the correction factor "+ 0.6 x reffective" would still apply - with maybe 0.6 changed to some smaller factor.

In the light of this there will be no much difference between the frequency of the quarter-wavelength t-line and the dipole peak frequency of the corresponding H frame. It would certainly NOT be possible to have a H frame that could "help in killing that (dipole) null".

My prior statement "You can easily imagine how both length differ with different H frame geometries" was plainly wrong. ops:

[Rudolf]

Brian,

I did a comparative simulation for 1" domes in your version2 baffle compared to a much smaller baffle:

Image not available

The left diagram is showing the response of two Visaton SC 4 ND mounted back to back on a baffle of 56 x 14 x 3 cm with the drivers centered at 8 cm below the baffle top. This hopefully is close to the dimensions of your latest baffle design.
The right diagram is for the same drivers on a very narrow baffle of 56 x 5 x 3 cm with the drivers centered 2.5 cm from the top.
In both cases a 12 dB HP at 2.5 kHz has been added.

It is up to you to decide which response pattern you would like better.

[brianpowers27]

Rudolf, Thanks for the reply. It looks as if the wider baffle provides a relatively flat power response between 2k-6k, wheras the narrow baffle tends to peak more... Food for thought... 14cm is very close to my actual baffle width. (15.24cm)

[Rudolf]

This is what a consistent dipole system would look like:

Image not available

And this is the radiation pattern of a good monopole system:

Image not available

Maybe I haven´t fully understood why you want to build an OB system in the first place. Is it about the "boxless" sound or about the dipole radiation pattern? It is my fault that I did not ask this in my first post.

[brianpowers27]

Maybe I haven´t fully understood why you want to build an OB system in the first place. Is it about the "boxless" sound or about the dipole radiation pattern? It is my fault that I did not ask this in my first post.

I do want to build a dipole. I may not end up with a perfect dipole, but I certainly am looking forward to the reduced radiated energy and interaction with room modes. I am also looking forward to the improved soundstage due to reflections.

I understand what you are getting at with the typical power response(radiation pattern) for both systems. I am attempting to build a functional dipole system. I may consider tapering the top even further, in order to make the design conform more closely to the ideal dipole radiation pattern. I am having a hard time matching the ideal dipole design criteria with a design that fits my aesthetic style. Let's face it, if it doesn't look good no one is likely to appreciate it but me. I am considering a 3rd and smaller facet to house the tweeter. I will model this up later.

The graphs that you produced for me are extremely helpful. I attempted to do some myself using the u-h frame mathcad sheet from the FRD consortium. I was able to get the very top two graphs to function correctly. The rest didn't work. I will have to try again.

Bare with me. I am looking for ways to get closer to the ideal. I always appreciate everyone's help. Don't be afraid to post.

[brianpowers27]

Do you think that roundovers on wings and the main baffle would be a good idea(Assume they are symmetrical on both sides.)?

[CraigJ]

This is what a consistent dipole system would look like:

Image not available

Rudolf,

Thanks for the excellent information. Is there a diagram of the speaker that goes with the above image? Our weather is finally warming up, so this might make for a fun experiment to compare with the Isiris Jr.

Craig

[JonMarsh]

Yeah, I'd be real curious to see the tweeter that manages the off axis dispersion in the green trace - the plot for green trace in the box off axis example does look like a somewhat average dome tweeter roll off plot (compare that with the SS Illuminator tweeter, though... )

Would have to be a LOT of rear hard wall refections from a rear firing tweeter to fill that in for the dipole case. Would also expect to see some comb filtering due to the time path differences- what's the smoothing used in this measurement? Inquiring minds would like to know... also what that front wave looks like if the wall behind is somewhat absorbtive.

[brianpowers27]

Yeah, I'd be real curious to see the tweeter that manages the off axis dispersion in the green trace - the plot for green trace in the box off axis example does look like a somewhat average dome tweeter roll off plot (compare that with the SS Illuminator tweeter, though... )

Would have to be a LOT of rear hard wall refections from a rear firing tweeter to fill that in for the dipole case. Inquiring minds would like to know... also what that front wave looks like if the wall behind is somewhat absorbtive.

I too doubt that any tweeter could get the ideal response in the 15-20khz range... I am using a .75" dome vs a 1" dome. This should help top end dispersion somewhat. I am guessing that the environment will change the wave considerably. The two 4 ohm tweeters will be wired in series with a variable l-pad in the middle.

WHen I build, I will measure. I have a couple of different test environments that could produce measurements for your question.

1. Basement = concrete floor, 3 walls have exposed insulation with a 2'x4'x1.5' pack of rockwool 60. It is quite dead.

2. Living room. 14',24',8' All walls are painted plaster. The room no longer has sound panels...

[Rudolf]

Yeah, I'd be real curious to see the tweeter that manages the off axis dispersion in the green trace

Sorry to say that I missed to set the proper depth distance for the tweeters in that diagram. Both front and rear dome were set at the same depth. :banghead:
When dialing the depth distance to 2.4 cm I get this response, which of course is less ideal:

Image not available

The tweeters are 0.5" Visaton CP 13 - the smallest I could find. They are simulated on a baffle of 5 x 3 cm with the driver at 3 cm height.

Would have to be a LOT of rear hard wall refections from a rear firing tweeter to fill that in for the time path differences- what's the smoothing used in this measurement? Inquiring minds would like to know... also what that front wave looks like if the wall behind is somewhat absorbtive.

Those are not measurements, but simulations for anechoic room conditions. I do not know the smoothing value, but certainly below 1/6octave.
You are certainly right to question the real world value of my proposal. I don´t know about any in-room-research into the specific importance of +2 kHz frequencies for the localisation of phantom images, or about the need of power response above 2 kHz to be uniform.

[brianpowers27]

You may need to explain the tweeter depth settings to me.

I am planning on mounting the rear tweeter on a separate .5" board with a .5" standoff between the main baffle and the rear tweeter board.

My design will be similar to the Orion++

[brianpowers27]

You are certainly right to question the real world value of my proposal. I don´t know about any in-room-research into the specific importance of +2 kHz frequencies for the localisation of phantom images, or about the need of power response above 2 kHz to be uniform.

You could be the first to research this.

SL liked the sound fo a rear tweeter, I suspect I will too. On the Orion, the baffle doesn't qualify for true dipole, in the treble region due to the baffle width. My guess is that the average tweeter dispersion mimics dipole behavior, especially when horn loaded/waveguided.

[Rudolf]

Brian,

I don´t have to bare with you. You have to bare with me. Actually you are helping me to make my point more clearly and explain it more focussed - which is dearly needed.

When SL and JohnK developed their dipoles they did not have access to a new generation of small sized neodymium tweeters. They had to deal with 10 cm diameter mounting plates and 2.5 cm deep magnet assemblies. That left them with (at least) 10 cm wide baffles in the tweeter region and a distance of at least 5 cm between the mounting plates if they wanted to mount the tweeters back to back for the best dipole position.

On the right side of the following diagram I simulate this situation for the already mentioned Visaton tweeter SC 4ND (now without any HP filter).
On the left side you see the same tweeter on the smallest baffle possible with the shortest distance possible from mounting plate back side to mounting plate back side (my "depth distance"):

Image not available

You may see that the polar response situation at 2.5 kHz in the right diagram now can be reached at 5 kHz in the left diagram. My interpretation: The upper limit of true dipole radiation can be moved one octave up when utilising the smaller size of the new neo tweeters.

The old tweeters did not physically allow to have a true dipolar radiation pattern above 2 kHz, so SL and JohnK abandoned to try it at all. But if they had to start with a new neo tweeter, would they not try a smaller baffle and a closer distance????

Some additional thoughts:
There is no doubt at all that a rear tweeter in dipole configuration is an assett. I never heard anybody (including me) saying that the rear tweeter was not an achievement.
It would be fine to think of the tweeters as a separate unit above the main baffle. (Narrow) baffle width is at a premium, so you don´t want large roundovers between the drivers. Anything below a 1 cm chamfer doesn´t help the acoustics, so you may forget about it altogether.
Don´t try to put any additional separation distance between the front and back tweeter like SL did. It might help to get a smoother transition in the crossover region, but it will ruin the dipole response in the tweeter region completely IMHO.

[Saurav]

Would you care to comment/speculate on using a CD waveguide on the (front) tweeter? Assuming perfect drivers/design, could this be used to create a front radiation pattern that approaches your ideal simulation? The rear radiation would be completely blown, of course, but it seems to me that the front 180 degrees should have an even polar response up to a higher frequency (essentially, where the response narrows beyond the waveguide angle), above which it should fall with increasing off-axis angle.

Is it more important to keep the off-axis curves flat? Or to prevent them from crossing each other?

[Dennis H]

FWIW, here's what The Edge predicts for a B&G Neo3 hanging in space -- no baffle other than the metal frame. 0, 15, 30, 45, 60 approximated by dragging the mic to the side. It looks pretty usable to me.

[AJINFLA]

Would you care to comment/speculate on using a CD waveguide on the (front) tweeter? Assuming perfect drivers/design, could this be used to create a front radiation pattern that approaches your ideal simulation? The rear radiation would be completely blown, of course, but it seems to me that the front 180 degrees should have an even polar response up to a higher frequency (essentially, where the response narrows beyond the waveguide angle), above which it should fall with increasing off-axis angle.

The front is more like 1/2 a sphere, not just 2D 180 degrees. It is possible to get uniform horizontal response with a multi driver WG setup but due to piston source directionality, non-ideal unless you use narrow bandwidths and carefully shaped baffles.

Is it more important to keep the off-axis curves flat? Or to prevent them from crossing each other?

Ideally the spectra of reflections would be a mirror of the on axis with tapering HF. But real rooms are neither perfectly absorptive or reflective, they are non-linear. Propagation loss is also not linear wrt spectra, so placement and dimensions will factor in.
Basically try to make the off axis as uniform as possible. Try to place them as ideally as possible in a non ideal room. Then sit back, relax and try to enjoy the music without obsessing over details too obsessively :W .
No matter what you think, these will not be the last speakers you make .

cheers,

AJ

[Rudolf]

Would you care to comment/speculate on using a CD waveguide on the (front) tweeter?

Earl Geddes once commented on his CD waveguide thread over at DIYaudio, that if he would ever do a dipole, it would be two of his waveguides mounted back to back and their mouthes connected by a large roundover. I´m sure this would give an excellent dipole. But you would need two waveguides to achieve this.

Assuming perfect drivers/design, could this be used to create a front radiation pattern that approaches your ideal simulation?

Not really with a front WG only. Hidden in my diagrams is the dipole eight. You always will find a blue horizontal line along the 0 dB axis which represents the 90° radiation. I know that this is highly idealised, but practically I get 12-16 dB difference in the frontal hemisphere between 0° and 90° all the way from 50 Hz to 2 kHz. And 10-12 dB in the back hemisphere from 90°-180°. That is with ungated MLS measurements in room.
The constriction at 90° can only be achieved with the symmetry of a dipole radiation pattern.

The rear radiation would be completely blown, of course, but it seems to me that the front 180 degrees should have an even polar response up to a higher frequency (essentially, where the response narrows beyond the waveguide angle), above which it should fall with increasing off-axis angle.

It would be no easy task to maintain a consistent frontal radiation pattern while the rearward pattern is a mess. Especially for angles above 45°. May be it can be done, but it certainly would be beyond my abilities.

Is it more important to keep the off-axis curves flat? Or to prevent them from crossing each other?

This is easy. If the off-axis curves are somewhat parallel, you can always equalise them to (somewhat) parallel flat lines. But if the curves cross each other, every attempt to flatten one curve will make all other curves worse.

[Rudolf]

FWIW, here's what The Edge predicts for a B&G Neo3 hanging in space -- no baffle other than the metal frame. 0, 15, 30, 45, 60 approximated by dragging the mic to the side. It looks pretty usable to me.

Dennis,
you caught me in the act. ops: The Neo3 looks pretty useable to me too. I did not realise it has such a good ratio of radiating area to "baffle" size. May be I was to fixed to ribbons when I claimed that planars would be bad OB tweeters. Thanks for correcting me.

[Saurav]

I guess I should clarify that I'm thinking of a tweeter on a WG, with a dipole midrange/woofer below it (i.e., my current design). For lower frequencies, the 90° response will be created by the dipole effect. For very high frequencies, it seems to me that even a monopole tweeter should have very little output at 90°. Where I'm hoping the WG will help is from the mid-tweeter crossover frequency, up to where the tweeter response starts narrowing (10-12kHz, in my measurements).

Though honestly, I haven't measured my speakers' response all the way out to 90°. I'll do that the next time I get a chance to drag them outside (maybe the weather will cooperate this weekend).

Anyway... I feel like I've hijacked Brian's thread enough at this point. Thanks for answering my questions

[Dennis H]

Hey Rudolf,

It's a little hard to know what to set the membrane size to as the PDR version uses some foam or felt inside so the effective HF membrane is smaller than the LF membrane. I used 60x87mm for the baffle and 25x37mm for the membrane in the sims. I tried reducing the membrane to 12x12 thinking you could do that with external damping. That gave far less HF rolloff but it also gave curve crosses.

[brianpowers27]

Project update:

I just purchased the Jasper Jig. The cutting went very smoothly. Pictures to come soon.

I restructured the top part of the baffle so the tweeter area is 2" wide and about 3" tall.

[brianpowers27]

I guess I should clarify that I'm thinking of a tweeter on a WG, with a dipole midrange/woofer below it (i.e., my current design). For lower frequencies, the 90° response will be created by the dipole effect. For very high frequencies, it seems to me that even a monopole tweeter should have very little output at 90°. Where I'm hoping the WG will help is from the mid-tweeter crossover frequency, up to where the tweeter response starts narrowing (10-12kHz, in my measurements).

Though honestly, I haven't measured my speakers' response all the way out to 90°. I'll do that the next time I get a chance to drag them outside (maybe the weather will cooperate this weekend).

Anyway... I feel like I've hijacked Brian's thread enough at this point. Thanks for answering my questions

This seems quite relevant to me. This type of hi-jacking is relevant.

It seems that most waveguide loaded tweeters trade low end efficiency for top end dispersion. I recently waveguide/horn loaded a cheap vifa b25 tweeter. I found that it gained me tons of low end efficiency.

I believe that the change in dispersion might not lead you to a true dipole. The limited dispersion may mimick a dipole but the lack of top end dispersion is ultimately forcing the individual tweeters to create their own monopole power response vs a dipole cancellation.

[brianpowers27]

Well I finally built it and tested it. I zipped the measurements and attached them to this thread. I will have to wait until later to test the feasibility of the system, as a whole.


Images not available

[brianpowers27]

Here are a few graphs. o, 45, 90 degrees for each unfiltered component.

[Rudolf]

Brian,

those graphs look really good to me. I am sure you could take the system as it is (with added crossovers of course) and have a nice dipole speaker.

Just two places where you could get even better:

First: Every cm you can cut off the tweeter baffle or back-to back distance will improve the tweeter graph further. But I don´t know at all if it will make a difference for the ear at all. When equalising the tweeter response I would concentrate on the 15° angle, definitely not on 0°.

Second: The 90° response of the midwoofer indicates that you have a 1.5 kHz peak at the rear response. I had very good results fighting such a peak by changing the driver mounting position on the baffle. I see that you haven´t chamfered the midwoofer hole yet. Possibly this might cure the peak already. If not, you may try more radical approaches.

The following graphs should illustrate what I mean. All measurements at 30 cm distance from the dust cap. First a 3" fullrange driver in free air:

Image not available

You can see that the 0° and 180° response are quite different.
Next the driver is mounted in the conventional way on a 15 cm wide baffle:

Image not available

You see what happened.
Only when I mounted the driver to the back of a baffle, the 180° response became acceptable:

Image not available

I don´t think that the same cure is applicable to all drivers, but you´ve got the picture.

[brianpowers27]

Image not available

I don´t think that the same cure is applicable to all drivers, but you´ve got the picture.

What does the front of the baffle look like? Did you chamfer the opening?

I will have to rethink the tweeter mounting. I have already broke one tweeter from bending the terminals. The terminals stick straight out... I need to find a low profile connection.

Also, what do you mean when you say this?

When equalizing the tweeter response I would concentrate on the 15° angle, definitely not on 0°.

[ThomasW]

Also, what do you mean when you say this?

He's talking about EQing for 15 degrees off axis instead of on axis

[brianpowers27]

He's talking about EQing for 15 degrees off axis instead of on axis

So in theory I would be equalizing just the tweeter this way, while attempting to design the xo for the other parts based upon on axis measurements.

[Rudolf]

What does the front of the baffle look like? Did you chamfer the opening?

Above measurement was on a thin (3 mm) masonite baffle. Mounting the driver on a 16 mm thick baffle from the backside and chamfering the opening 45° took off the 2 kHz peak too:

Image not available

I've got three different 3" drivers (FRS8, FF 85K, FE 83) with baskets that look much alike. But they've got quite different 180° responses. It could be tricky to find the mounting which will subdue any resonances or reflections best.

Also, what do you mean when you say this?

ThomasW is right.

Sorry about the broken tweeter.

[Saurav]

So in theory I would be equalizing just the tweeter this way, while attempting to design the xo for the other parts based upon on axis measurements.

I think he means designing the XO with the 15 degree data for all drivers. But I could be wrong... if he means use the 15 degree curve for just the tweeter, then I'm curious about why you should do that.

[Rudolf]

So in theory I would be equalizing just the tweeter this way, while attempting to design the xo for the other parts based upon on axis measurements.

Yes. Thats the beauty of the narrow baffle approach: As long as you stay below the first dipole peak, most dips and peaks in the response will be at fixed frequencies, regardless of off-axis angle. If you EQ on axis, it will help to EQ off-axis too.

Only if you have to EQ above the first dipole peak, dips and peaks will change their frequency depending on off-axis angle. You see it in your tweeter response above 4 kHz. Equalising for 0° will make the response at 45° worse and vice versa.

[brianpowers27]

I just ordered a 1" cutting surface chamfer bit. I will have to wait until this bit comes in before doing any more measuring. I plan on trying the rear mounting for the mid and chamfering the front as pictured in Rudolf's post.

I ran some basic simulations on Sunday. I had to pad the mid and tweet a good bit to match with the woofer. I also found that a BSC circuit helped to get solid response in the low end.

I am likely to build the final project with .75" pine boards. I found 16" boards for a resonable price. I have a nice cherry finish that I have been using lately. It is the same one in the following picture.

[brianpowers27]

I am listening to the maiden flight of the finished cabs. So far the simulation is working out well. More details to come. :T

[brianpowers27]

More data...

[brianpowers27]

I have gone through a few xo modifications. THe most notable problem was the bass boom between 100-300hz. I remedied this by altering the midrange rolloff and I am awaiting a 3.6mh coil for the woofer.

Note: I am going to need to alter the baffle/wings/feet in order to reduce a few of the buzzes I hear. Part of the buzz comes from the wires slapping against the cabinet. Part comes from the driver frame buzzing against the chassis. Part comes from the speaker minutely "walking accross the floor". I will likely add foamies underneath the legs and gasket tape between the baffle and the drivers. I may even consider somehow adding mass between the baffle and the floor... (Hmmm.)

Listening impressions:
Holy cow! These may be the best imaging speakers I have ever heard! I hear things in 3d vs 2-d like I expect from mono poles.I took a total shot in the dark on driver selection considering everything was either on buyout or on sale. The bass response is quite adequate for most listening. The monopole bass does reduce room boom. A monopole sub would help out if crossed around 60hz.

[brianpowers27]

The pics I promised...

[Dennis H]

Glad you're liking the sound of your speakers, Brian! I'm not sure what you mean when you say "the monopole bass does reduce room boom." Did you mean dipole bass?

[brianpowers27]

Glad you're liking the sound of your speakers, Brian! I'm not sure what you mean when you say "the monopole bass does reduce room boom." Did you mean dipole bass?

Oops! You are right. That is what I wanted to say.