Hey guys...we need a little rallying here... - Page 22

Again a viable option, and for all we know they will be unfinished resin and need final prep and painting anyways.

Ok, so it seems we have a consensus on look, only problem is this #4 WG is the one I stretched from the narrower profile N=2.5 SEOS Jzagaja posted a few pages back.

Now we need to figure what the corresponding transition radius is and see if Jzagaja can model a similar WG and if it performs well on paper.

Here is the original EOS with a 12cm radius:
http://www.avsforum.com/avs-vb/showt...9#post19556489

The SEOS's radius unknown:
http://www.avsforum.com/avs-vb/showt...1#post19641991

While searching through the threads I ran across a model Jzagaja posted of an 12" OS-JMLC hybrid, maybe this expanded to 15" in EOS may be the sweetspot?
http://www.avsforum.com/avs-vb/showt...9#post19556489

As penngray says, the other reason is to let the CD handle more of the critical midrange, as it's more efficient and better behaved at those freq than the woofer.

And of course, the lower the XO, the more the woofer's inevitable breakup modes are attenuated.

Yes.

Not sure I understand the WG/baffle blend options being discussed.

I'd think, and it looks like the case in all of the pictures, that the WG in any incarnation has its periphery parallel to the baffle and it flush mounts, meaning that its blend radius is fully defined.

Correct?

If not, can someone sketch out these WG termination schemes being talked about?

As far as mounting flange, I like the nonrectangular ones that follow the shape of the WG.

I think it has a cool artsy look and would want the option of leaving it exposed.

Though that may not happen if the WG is filled with foam.

Speaking of which, anyone interested trying that out?

I am thinking of mounting my WG in free air, I am not sure about the foam though. But, I know I could probably make it look awesome / odd and have it appear that it is "Floating" Above the speaker, and get it to appropriate ear level (for me) and It will still function well.

Though, my only concern is, with the baffle removed, once the sound exits the WG it will be in open air, and not have a baffle behind it, which could change the sound.
I know that the baffle determines some aspects of xover design (BSC) but, aside from the basic principals of that, I do not know how much it can/will affect the sound from a WG.

Obviously, if this will end in disaster, then I will simply mount it in a box, or flush with a forward baffle (Like #2 in the lineup)

I can't check all possible configurations. This is hobbywork for me. This days I prefer large format JMLC horns setup. Iwata-300 seems good compromise:

http://www.diyaudio.com/forums/group...ml#post2401608

I haven't listen to OS with hi-end drivers. Some experience with waveguides tells me SEOS-15-9060-N3 is a good compromise.

Like I said earlier subjective difference between waveguides will be small but if you cover 300Hz-20kHz with compression driver(s) the difference is more pronounced. That's why we are involved in bass compression driver project.

H_radius = 12cm, V_radius = 3cm.

Thank you for your response. So by your estimation at this size and XO point there will be little difference/benefit between an OS/ EOS/ JMLC profile? I understand the reasons for wanting a lower playing horn, and it seems you and others already have this range covered with some great looking products. I am currently building a router duplicator so I can make a few hardwood Iwata's. However this project stemmed from the want/need for a better or ultimate baffle mounted mid-size horn that will play well with a 12 or 15" driver in a moderate size cabinet. From the graphs you have produced it does seem there is some benefit to either a modified JMLC-OS or an OS with a large transition radius in order to flatten frequency response/negate vertical nulls while maintaining control within the desired pattern frequency. Is the 12cm radius on the original EOS profile which you modeled for Penn, the "optimal" radius for these purposes, or was it just a starting point? Also would an OS-JMLC hybrid with 15" width have a greater depth than the EOS/SEOS, thereby increasing VC offset?

Now that you have provided the vertical radius value for the SEOS I overlayed and aligned the resized horns and measured them. Based on my calculations for the vertical values in comparison to radius, the image I made of the stretched SEOS n=2.5 is equivalent to a ~8.25cm vertical radius. Say we call it 8.5cm, will this fair worse than the EOS's 12cm in the vertical?

I believe If you could model a 90x45 SEOS N=2.5 or 3 rH=12cm rV=8.5cm, including off-axis response's as you originally did the EOS 15", that will allow the group to directly compare and help make a final decision. Regarding your suggested SEOS-9060 I believe some purposefully want a more limited vertical directivity to reduce ceiling and floor bounce in their theatre's, and also it seems many prefer the shorter look of the modified SEOS I created, therefore if a going to a 60deg vertical will change these dimensions or lessen the transition radius it may not be as preferred.

Thank you for your time spent to help forward this project.

Sounds like a part number for an existing part - is it?

If that's the case, isn't all that's needed is getting one to make a mold from?

Erich, are you interested in that?

Assuming jzagaja's is OK with that of course, and I would understand if he isn't.

I think that's just the SEOS 15" with the 90/60 aspect and he thought the N-3 was better than the N-2.5 which was posted earlier.

JMLC has some advantages but constant directivity is not one of them. I'm pretty sure that is a common goal amongst the participants in this thread. I don't know much about applying the JMLC formula to the roundover portion though. That might have some value.

WTF?!

We are doing OS here, NOT JMLC.

[The debate between the two of them has been going on for years.... ]

I was replying to jzagaja's reference to it as his preference. Trust me, I have no interest in JMLC's.

Sorry for bringing up JMLC, the FR graphs he posted early in this thread of a hybrid OS-JMLC showed a more even FR (negated a large null) compared to straight OS at the expense of a dB or 2 in HF pattern. Just covering all possibilities for the best outcome, I was not aware of it's shortcomings in a CD application, and maybe in hybrid these are negated. In addition I happened across the Geddes thread on DIYA this evening and it seems most all the latest discussion has already been fought out over there, so while everybody may not agree on each point it seems a general consensus can be found.

Therefore is what I asked Jzagaja to model a few posts up your guys' ideal solution? Shall we start a re-vote, or are all those who originally voted for the EOS-15 happy with the following?

90x45 SEOS N=2.5/3 rH=12cm rV=8.5cm??

If not, please put forward any changes/suggestions.

Yes but only with waveguides. JMLC is a horn with varied directivity. They are cleanest but suffers from sweet spot. Minphase horn is something in between (check #551):

http://www.google.pl/search?sourceid...+gelsenkirchen


8cm was used in Mummies and is so-so, 12cm is better and 14cm is best. Simulator takes only one plane so it doesn't mind EOS or SEOS unless I feed a plane in between. Forget about rV>3cm, WG will loose directivity control.


It's geometrical construction so can be equal or longer. SEOS is very short (12cm), OS-12 is almost 20cm.

Thanks, I was not aware of that regarding straight JMLC, and was going off your "hybrid" OS-JMLC in which you stated integrating the JMLC roundover onto OS may be optimal. Will have to read up on the Min Phase, interesting concept.



This is what we need to know. So does rV </= 3cm only apply to the SEOS, since the original EOS-15 design you modeled had a rV=12cm. In an earlier post you wrote:

"Roundover >5" removes axial dip but from the other hand slightly reduces directivity control above 60deg. Below is EOS-15 vertical plane directivity. Coverage is 45deg and roundover same as in horizontal plane."

So if SEOS is limited to a 3cm I assume the axial dip returns. Also, does this mean that the only way to get the "look" of the taller stretched SEOS that everybody likes is to screw up the performance of the horn?




Good to know. Thanks.

SEOS N value= 2.5 (the value of the original I stretched in the mock photo) or 3 (as was the latest suggestion from Jzagaja as a happy medium between the 2.5 and 4 he orginally modeled)

rH = transition radius Horizontal

rV = transition radius Vertical


The rV=8.5cm is what I came up with after overlaying the mock-up SEOS #4 everybody liked against 2 known values. However in his post this morning Jzagaja stated a vertical radius larger than 3cm on SEOS will lose directivity control. Therefore depending on his reply to my follow up questions, we may be stuck with a short and wide horn as in mock-up #5 if we stick with SEOS, and be better off going with original EOS-15 which had a 12cm vertical radius and maintained control to 45deg while eliminating the null common with standard OS due to its extended vertical radius.

First graph is simulation of an OS with coverage 50deg as in EOS vertical plane. Result - "axial dip" is not an issue with such coverage, optimal rV=3cm. Second graph is simulation of an OS-12 with roundover 4.8cm and measured response.

In all honesty, I think we are the point where the rubber hits the pavement and it is rather difficult to determine which design choice is best. I'll be honest, I think any of these versions with different super ellipses and roundovers would work for everyone from Zilch (workable CTC) to Penn (xover at least as low as 1khz). I guess I'm saying I'm fine with anything in the range discussed.

I know most of us here want to get this rolling to the next phase. This is all exciting and awesome that we are this far and no one wants to lose any momentum that this project has.
My only issue with just jumping into this is that prototyping is expensive, and we are getting someone to do this for us.
We do not have the luxury of simply tossing out a design, saying "oh, that's not quite what I wanted" and then simply making another, and another and another.
Yes, we might end up making 2, but that should be because our first design should be 95+% and we know that simply changing X a tiny bit will make it better.

Tossing around ideas, coming up with new ones and shooting them down is all part of the design stage with anything.

If we want to do this, and do it right, let's ALL take the time to get it as close to perfect (for what we want to do) on our first take.
Discussing the sizes, weighing the options of A versus B costs us nothing (but a little time.) So why not make sure this is what we want to do.
(That is unless we have some time constraint that I am un-aware of.)

Quality first, then we can "Put the rubber to the pavement."

my two cents.

Thank you, still a few unanswered questions, if you don't mind. I'm a bit confused as you stated above rV=12cm is good but 14cm is best, and now we're at 3cm is optimum. In which regards is your original statement of rV=14cm true, is it only regarding the "axial dip", or in general including diffraction and the elusive HOM's, only OS not EOS/SEOS?

-Therefore does this "axial dip" than re-reverse itself at higher radius, because the original EOS-15 you modeled with rV=12cm has similar characteristics of the rV=3cm EOS, and also outperformed the straight OS model you presented back then.

-Does the rV<3cm rule only apply to SEOS, and not EOS?

-Was the reason you originally used such a large rV due to the fact you were trying to meet the original parameters of a 15"x10" dimension?

-Is there a way to expand the desired physical vertical height of the SEOS to ~8" without decreasing the performance (by increasing the vertical directivity maybe?)

There is always a risk of splitting hairs to far in the discussion. Also "windows of opportunities" never stay open for ever. The longer the discussion goes on the more the inherent risk of failure increases.

Lets remember someone is ready and waiting. There is no time frame set but I would never risk prolonging the discussion and risk that window closing for unforseen reasons.

I agree with all of that I just believe we are at the point where the differences are so small that no amount of simulation or discussion is going to improve on it. We are at the stage of trade-offs and fortunately very small trade-offs. The only way we could make it better beyond the range of choices we have now would be to prototype and measure. Even then, what makes one designer happier wouldn't do the same to another designer. We are at the point where we are picking a winner regardless so we might as well throw a dart at the dartboard.

If you look at the success of the QSC WG, I'm confident this one will also be successful. It is very similar, but improved in a number of ways with no real downside other than higher cost.

So, on roundover and super ellipse shape, your guess is as good as mine and both will be right.

I'd like to get the information in around the first of the week. It would be really nice to get this figured out by Monday if that's possible. Obviously if there's some questioning, then we can wait a little longer. But sooner than later now that we're this close.


I'm hoping that the new talk about N=2.5/3 rH=12cm rV=8.5cm (or whatever values) doesn't confuse the issue more. I'm claiming some ignorance on that stuff until I get a better handle on the terms. If any of that changes what we've been discussing for the past couple weeks, then I'd hate for it to cause some issues when brought up at the last minute if others aren't sure about it.

If those values are only used to change the look of the waveguide, but alter the sound in a negative way, then I think they should possibly be scrapped.



I'll go back though the thread tomorrow if I have time and write down most of the main characteristics that we want. If anyone else can also make a 'list of demands' for the prototype, that would help as well.

Minimum required roundover for removing "axial dip" depends on coverage. With 90deg coverage it is 12cm, with 50deg - 3cm and these values were used in SEOS-15-9050-N4. Now if you need vertical height to be 8" we need rV=4.5cm and 60deg coverage. Should I draw and simulate SEOS-15-9060-N3?