AudioCircle
Audio/Video Gear and Systems => Open Baffle Speakers => Topic started by: glt on 26 May 2010, 06:15 pm
-
Hi, first time trying a diy speaker project. Currently using Newform645 speakers and want to experiment replacing the bass section with OB single or dual 15" woofer.
The Eminence alpha seems a good choice for the project. The crossover needs to be at 1KHz and I have not seen any project with such a high crossover point.
I also plan to use a miniDSP for active crossover
What would be the best choice? Thanks in advance for any info
-
I also plan to use a miniDSP for active crossover
Let us know what you think of this...I've been eyeing one for a long time. They look really cool.
Sorry I can't help on your question.
-
i don't know the eminence woofers but i used my woofers crossed at 1Khz. Mine were vintage tesla that lampizator uses in his projects. These are capable to play to 1,5Khz without any sweat. I also used the Altec 420Y and goodmans axiom 301 crossed so high. But these aren't typical woofers but fullrange speakers so the frequency response and overall characteristics might be different then Eminence. If you eminence can play so high on paper, it shouldn't be a problem.
If you can, find a pair of decent ALtec 420A or 420Y and use it as woofer crossed at 1000Hz. It is very very good speaker in this range.
-
AE Speakers Lambda TD15M, i run mine with no crossover they go up to about 3k ish
-
Ditto. Not sure about the Alpha, but many 15s have no problem up at 1Khz.
-
Wow, so much help. Thanks for all the responses. Right now everything is just on paper. I like the miniDSP because you can have room EQ plus crossover. The 4th order crossover will come in handy because the eminence has a peak at 2K (on paper). Ah, yes, the AE speaker was another I was thinking about, but it is 5X the price of the eminence alpha...
-
the eminence alpha may work up to 1KHz but when I adjusted the crossover point on my behringer to about 300Hz they didn't sound all that. My full range drivers sound far superior to the eminence at 300Hz. They may work but you may well be very disappointed with what you hear. Oh and mine are in open baffle by the way. Don't take my word on it, maybe others will have a different view.
-
Selenium even have 18"ers that can be used that high.
-
It's not about the on-axis response only. Look how a 15" driver on a 50 cm wide OB will radiate off axis:
(http://www.audiocircle.com/image.php?id=30867)
The curves are 0°, 15°, 30° and 45°. Crossing at 250 Hz would be fine, 500 Hz acceptable. If you want to cross at 1 kHz, you need a driver with the same beaming character at that frequency. That has to be a horn or a waveguide.
-
I've pushed the alpha up that high, and even with a deq, you really don't want to.
-
altec 416/515 sound fantastic at 1K, but they don't go very low and are unsuited for OB...
woofer of that size at that freq will either sacrifice response below 50hz to achieve the necessary sensitivity, or will sound like crap at 1K. I'd go with an altec and then a sub, IMO
-
Thanks for all the info. This is the best place to learn about OB
Would the AE lambda perform better at the 1K region?
Rudolf, the driver for >1K is the newform 4" ribbon driver (don't know the beaming pattern).
But listening position is about 15 degrees, so shouldn't be too concern about off-axis?
-
Rudolf, the driver for >1K is the newform 4" ribbon driver (don't know the beaming pattern).
Tried to do a radiation pattern 0-45° for a 2" line source (monopole) on a 4" wide baffle of 75 cm length:
(http://www.audiocircle.com/image.php?id=30889)
At 1 kHz the loss from 0-45° is 3 dB for the ribbon, but 8 dB for the woofer. No good combination from my POV.
You can easily try it yourself with the free simulation program Edge from www.tolvan.com/edge
But listening position is about 15 degrees, so shouldn't be too concern about off-axis?
Even when listening on axis, more than 50° of what you hear is indirect sound. You need to be concerned about off-axis response too.
What I am really concerned about is how you are trying to connect a monopole line source with a dipole point source. If you want to look into OB bass for those ribbons your aim should be a dipole line source with 5-10 drivers of 5"-7" size stacked in a vertical line IMHO.
-
Rudolf, since this is my first attempt at speaker building, I was mainly attracted at the simplicity of a single drive/OB. It seems that with a 15" driver there are many issues to consider. The ribbon I have is 4', not 4" :-)
I would then trust your recommendation for a line array of 6" woofers. Would this be a reasonable choice: http://www.parts-express.com/pe/showdetl.cfm?Partnumber=299-250 ? or this: http://www.parts-express.com/pe/showdetl.cfm?Partnumber=299-908 ?
The only thing relevant to me is that the second driver has a much higher Qts
Thanks again for all your help.
-
Rudolf, since this is my first attempt at speaker building, I was mainly attracted at the simplicity of a single drive/OB.
Well, OB is as simple or complex as any other speaker principle. From my point of view it is even more dominated by the "housing" than a boxed speaker. For a first attempt I would stick to something better explored than the road you are trying to go.
It seems that with a 15" driver there are many issues to consider. The ribbon I have is 4', not 4" :-)
A 15" woofer is perfect for OB - as long as you don't make it the midrange driver too. There are reasons why the woofer-fullrange combination is so popular for a first OB. You get a well integrated system without much complexity.
With 4' the line is even longer than I considered. I thought that 4" was for the width.
I would then trust your recommendation for a line array of 6" woofers. Would this be a reasonable choice: http://www.parts-express.com/pe/showdetl.cfm?Partnumber=299-250 ? or this: http://www.parts-express.com/pe/showdetl.cfm?Partnumber=299-908 ?
The only thing relevant to me is that the second driver has a much higher Qts
Both drivers would be reasonable. And the higher Qts would in fact make things easier. Sometimes I can't believe the great deals yo guys get from PE. The Peerless offer is amazing.
But -as said - I really can't recommend such a system as a first venture into OB.
Would be nice if some of the other contributors to this thread could comment my reservations.
Rudolf
-
I tend to agree with Rudolf and especially with LA souces they are not so easy to get right. But going DSP and wanting not to spend a fortune why not investigate in 3-way OBs. Like the MJK H-dipoles for Alpha15 and Goldwood 18" (http://www.quarter-wave.com/ - look at Open Baffle theory and Projects) and then putting a single Peerless SDS above on a say 12x28" baffle side by side of the 4' ribbon. Crossing at 200 and 1000 Hz the Peerless might use some EQ low but it will put up with it graciously. As Rudolf said the Peerless at this price is a real bargain and if later you would go for an even better unit the option is there. OK, you might have to spend some more on amplification but that you would have discovered yourself trying to press a 15" speaker up this high.
In fact I am a bit surprised about Earl Gedde's SUMMA speakers to evidently perform extremly well within your first proposed limits, so one option would be to get one of the same bassunits and design a closed enclosure from information available. But then you will not escape putting subs for the bottom.
My Alpha15s do not play over 200 Hz nowadays for reasons others have given above. It will not compete with the smaller good bass-mid or mid speakers in tone and quality.
/Erling
-
Just as a point of reference, here's measurements of 2 different 18" drivers on OB. In room, at listening position. So maybe ~30-45 degs. off axis.
Also pics of the speakers. You can see what a difference there is in the drivers and also what a difference crossover slope makes. The Eminence is a super clean driver. The Goldwood, not so much. But it's not meant to be. More intended for M.I. use than subwoofer. The Goldwood is a hi-Q driver, the Eminence, very low.
(http://www.audiocircle.com/image.php?id=30959)
(http://www.audiocircle.com/image.php?id=30960)
One cheat with the Emmy. I took out the 34Hz room bump with active EQ. No EQ on the Goldwoods.
So crossover is essential, especially with cheap drivers. With cheap drivers you often have to trap out resonances to get good bandwidth. This is true of 18",15" - right up to dome tweeters.
(http://www.audiocircle.com/image.php?id=30957)
(http://www.audiocircle.com/image.php?id=30661)
-
Thanks so much for all the info. I've decided to follow the advice here and do a 3 way: 200Hz, 1KHz, which means getting two more amps and a miniDSP at a minimum. I did purchase 20 of the $10 Peerless drivers to experiment with OB Line Array. I'll do 9 per side in a 3 parallel/3 groups series configuration to keep the impedance at 4 ohm.
BTW what is a "good" (and cheap :-)) microphone to measure in-room response?
Also, panomaniac, what are the models of the speakers?. When you say say "the Eminence is a super clean driver. The Goldwood, not so much" do you mean flatness of the response?...Thanks.
-
For a mic this is a very good choice:
http://www.parts-express.com/pe/showdetl.cfm?Partnumber=390-801 (http://www.parts-express.com/pe/showdetl.cfm?Partnumber=390-801)
Just remember that in room measurements are difficult because of reflections. You have to use your best judgment there.
The Eminence 18" is not longer made, too bad. It had very little cone break-up or other noise above its usable range. The Goldy is not in the same league. But it is not meant to be nor priced to be. It's probably meant as a bass guitar driver, where you want extra noise for harmonic richness.
The Goldwood has more noise above its useful bandwidth. That can be taken care of in the crossover, however. Yo just have to be aware of it.
-
Just saw this interesting thread . .
“Even when listening on axis, more than 50° of what you hear is indirect sound”
Hi Rudolf
50% :o
I know off-axis response is important, but didn’t feel it could be that much ~ without ever seeing a figure.
Where did you see or calculate that figure?
Would it be a little room dependent, ie with closer side walls, that figure would it be higher?
Fyi I have a pair of Lambda TD15M, and cut the 46 cm (18”) baffles last weekend. Progress will be slow, as the only wood workshop I can access is 4 hours return drive :(
But I will over time try these:
- crossing about 250 Hz to an 8” full range, as a 2 way. I thought that’d be both the easy and good answer . . until I considered the beaming above about 1500 Hz
- As TD15M's can go so high: a 2 way, crossing about 1500 Hz to back to back tweeters. Be "interesting" to hear the drop in off-axis response, over say 500 Hz . .
- Later as a 3 way, crossing about 350 Hz to a 6.5” mid, then about 1500 Hz to back to back tweeters
-
“Even when listening on axis, more than 50 % of what you hear is indirect sound”
Hi Rudolf
50% :o
I know off-axis response is important, but didn’t feel it could be that much ~ without ever seeing a figure.
Where did you see or calculate that figure?
Would it be a little room dependent, ie with closer side walls, that figure would it be higher?
With a speaker in OB half the sound energy will go to the rear side - it has to. From the energy radiated to the front hemisphere your ears will only catch a fraction directly. Most will reach your ears only after some reflection from the room boundaries. The direct sound will dominate your hearing impression, but the first reflections - reaching your ear within 30 ms - get integrated into your listening experience too because of the precedence effect. Floyd Toole describes it in depth in his book 'Sound Reproduction'.
If the frequency response of the indirect/reflected energy in the room doesn't have enough resemblance with the direct sound, your brain can't integrate it with the direct sound - probably you might hear additional phantom sources.
Your primary aim is still to get the direct sound right. But you want to make sure, that the indirect sound energy in the room does not look (and sound) too different.
-
(I knew that half an OB's sound energy goes to the rear).
But not that - of the energy radiated to the front, our ears only catch a *fraction directly.
Yet *direct sound dominates our hearing impression . .
So is it the reflections reaching our ears within 30 ms, integrated (into our listening) added to the direct fraction, that dominates our hearing impression?
Cheers
-
So is it the reflections reaching our ears within 30 ms, integrated (into our listening) added to the direct fraction, that dominates our hearing impression?
Sort of, yes. It can't be explained in two sentences :( Google for 'Precedence' and 'Haas' effect and for 'Apparent Source Width' ASW. In my opinion the room is a much more important factor in loudspeaker design than commonly believed. And I am not the first one who recommends Toole's 'Sound reproduction' as essential reading. :wink:
-
I recently got Toole's book, and will try to find time to read it (500 pp)
I appreciate that room interaction is very important, and overlooked by many. that's why I'm going OB
I didnt know the extent to which off axis was *so critical -
In fact I thought that "beaming" of a 15” say 500 – 1500 Hz, means less room interaction, so would be a *Plus . . ?
-
I recently got Toole's book, and will try to find time to read it (500 pp)
Reflections are chapter 6-10 mainly - out of 22 :roll:
I didnt know the extent to which off axis was *so critical -
I wouldn't call it 'so critical', but it should not be ignored either.
In fact I thought that "beaming" of a 15” say 500 – 1500 Hz, means less room interaction, so would be a *Plus . . ?
For most people "beaming" would be a changing directivity - increasing from 500 Hz upward for a bare 15” driver. What you really want is constant directivity. That's what a 15” dipole would deliver up to ~ 500 Hz.
-
Yes! I agree (I'd forgotten constant directivity) :oops:
"That's what a 15” dipole would deliver up to ~ 500 Hz"
Is there a reference and/ or formula for this?
Up to what Hz, would we get constant directivity from
- a 8" dipole
- a 6.5" dipole
Thanks again
-
"That's what a 15” dipole would deliver up to ~ 500 Hz"
Is there a reference and/ or formula for this?
Do a simulation in Edge under 0°, 30° and 45°. That will show you what happens off axis.
Up to what Hz, would we get constant directivity from
- a 8" dipole
- a 6.5" dipole
Dipole CD is a function of baffle size and driver diameter. Again - do some simulations in Edge. They are really helpful to get a better understanding of what is happening.
-
When looking at results in the Edge, what constitutes CD – what dB gap between 0, and 30 and 45 degrees?
Thanks again
-
When looking at results in the Edge, what constitutes CD – what dB gap between 0, and 30 and 45 degrees?
"Constant" is about dB gap width staying constant - not about absolute gap width values. In the diagram you see constant directivity up to 500 Hz. If you are more exacting, you would choose 300 Hz as the upper limit:
(http://www.audiocircle.com/image.php?id=31980)
Above 500 Hz the driver starts beaming - the dB gap between angles becomes larger with frequency.
There are different sorts of constant directivity which are defined by the directivity index. Omni speakers would have a directivity index of 0 dB and dipoles have a directivity index of 4.77 dB.
-
Thanks again, Rudolf
What are the prospects currently, of your site being translated into English?
-
What are the prospects currently, of your site being translated into English?
I have changed my "business model", moving from HTML to a series of PDFs. First PDF is a 14 page treatise on OB response patterns. Today I finished the translation of page 7. Hopefully it will be online by the mid of next week.
Soccer world championship tends to get in the way. If we win against Britain, there are more games to watch. If we loose, I refuse to speak/write English for the next few days :wink: :green:
-
I have changed my "business model", moving from HTML to a series of PDFs. First PDF is a 14 page treatise on OB response patterns. Today I finished the translation of page 7. Hopefully it will be online by the mid of next week.
Soccer world championship tends to get in the way. If we win against Britain, there are more games to watch. If we loose, I refuse to speak/write English for the next few days :wink: :green:
Based on performance over the last week I believe Germany should win.
-
I have the odd country I like to see lose, and the UK is one of them (old colonial spirit?).
(Even though you beat us 4-0) Germany, may the force be with you! :thumb:
And if the Brits were to pull it off, remember that that most of your English speaking readers are Americans :wink:
Rick (an Aussie)
-
Rudolf
Is it possible that you put your paper that is written in German text through an automatic translator into English(I understand the auto translator is not perfect in fact far from perfect)? After that you correct the content of the translated text to make it intelligible and/or as close to the German text. Since you are the author as well wouldn't it be faster this way. BTW I have no experience in translating(but I have done some translating using specialized s/w from Japanese to English) this is just my guess.
-
Rudolf,
Take your time and transalte to your satisfaction. It is good that people like you invest time to write clarifying papers dealing with fundamental OB issues. Our understanding is much improved by this kind of efforts. And the knowledge is nowhere else to be found. So we all benefit.
However your AJHorn paper inspired me to see what Dave McBeans 'Hornresp' program could be used for. In fact I think you could do the same simulations with that program. If I understand right it would be possible to do H-, U-, Ripole and (with reservations) flat baffle simulations in Hornresp with good results. See here Hornresp input parameters, schematic diagram and SPL-simu results and a comparison with Martin's model result for his H-baffle with Eminence Alpha15:
(http://img707.imageshack.us/img707/8897/hornresphalpha15input.jpg)
(http://img153.imageshack.us/img153/4123/hornresphalpha15schemat.jpg)
(http://img375.imageshack.us/img375/3482/hornresphalpha15spl.jpg)
at last MJK's model:
(http://img31.imageshack.us/img31/1084/mjkhalpha15spl.jpg)
There are mainly differences in level which can depend on different setups for the simulations but I have tried to make them as alike as possible. The thing with Hornresp, you have to go for the combined response in the Tool menu when you are in the SPL-window and input the path difference to get the right answer.
Now for even more OT. We will see Germany and Brazil in the final for this World Cup. It would of course have been different if England had kept their womenazing Swedish coach, remember Munich. But my barber is ever so proud of coming from the same small Kurdistan town as Oezil's parents. Him playing for Turkey !!! HUUUH !!! :evil: In the end he might be chosen the best player this time. And Klose makes all the difference.
/Erling
-
Hi Erling
Have you contacted Dave McBean with your idea? I wonder how it compares with JohnK’s ABC Dipole?
Germany v Brazil. The emotion! If it happens, Best of luck :thumb:
-
Hi Rick,
No I just put up the simulations so Dave Mc Bean don't know anything, if he didn't read the post. It is a byproduct of that clever program just as Rudolf has used AJ Horn for the same purpose. But it is nice that the simulationprograms do produce very similar results. I suppose that will strengthen confidence in them. I have not used ABC Dipole so I cannot tell.
OT, Regarding the World Cup I am Swedish so I can be a bit neutral. But Sweden like Germany have had a lot of immigration from Kurdistan. Best teams so far are Germany and Brazil and this time Germany can hit from an underdog position. :)
/Erling
-
Hi Rick,
did you look into your audiocircle messages box lately? I need an email adress of yours to send you an attachment.
Hi Erling,
I never looked into hornresp, but now I certainly have to. Your comparison looks really good. Now I need to get aquainted to the program to understand what you did.
WRT to the translation, I have almost finished the pre-translation and some native speaker probably will bring it into a readable shape. In addition he might look whether everything can be understood by someone with a basic knowledge of dipoles.
Automatic translation does not work well in my case since my story is not told "narrative", but simply as an explanation around pictures and diagrams.
Rudolf
-
I have changed my "business model", moving from HTML to a series of PDFs. First PDF is a 14 page treatise on OB response patterns. Today I finished the translation of page 7. Hopefully it will be online by the mid of next week.
Nice, looking forward to it.
PDF are best for printing, which i always do for worthwile stuff.
And i wish Germany all the best in this championship - it was my bet from the beginning.
Ah, can you explain how to make the angled response in Edge? Maybe there's a new version out?
Cheers,
Guido
-
Ah, can you explain how to make the angled response in Edge? Maybe there's a new version out?
Hi Guido,
no new version of Edge. It is just triangulation. If you have a certain "mic distance" and move the mic sideways from the speaker center, you can emulate any angle from -90° to +90°.
I have done a chart with the angles most often used:
(http://www.audiocircle.com/image.php?id=32337)
Thanks for the good luck wishes. Statistics dictate that a team from South America has to win 2010. Since Brasil is out, it has to be Argentina. You don't see Paraguay or Uruguay as the champion, don't you? So Germany can't win the next game ... :wink:
-
Hi Rudolf - I think Edge is normalizing to mic distance. So there is not actually a need to change the mic distance, just the horizontal distance (scaled accordingly).
In either case, it looks like the level will need to be adjusted as you move off-axis e.g. at 45 degrees, by 3 dB.
That's what I think I'm seeing when I try this anyway.
-
Hi Rudolf - I think Edge is normalizing to mic distance. So there is not actually a need to change the mic distance, just the horizontal distance (scaled accordingly).
In either case, it looks like the level will need to be adjusted as you move off-axis e.g. at 45 degrees, by 3 dB.
Hi John,
as I understand it, Edge will normalize all open baffles to +6 dB for high frequencies (and 0°) - regardless of mic distance. But the shape of the response will vary with distance, admittedly more for small distances than long ones.
For more than 15° all off axis values will be plainly wrong. Adding 3 dB at 45° is the right direction, but still not the truth. Nowadays I use Edge only for very rough estimations, if going off axis.
Rudolf
-
Regarding Alphas up to 1K.....if you are using one of the stock drivers on a poor baffle with no bracing you will be very disappointed with their upper end performance....actually, they suck terribly in the bass used that way too.
Here are the ways to make a seriously great sound:
1. Run two in parallel. This lets them move half as much and therefore less coloration/distortion.
2. Put felt on the inside of the basket struts to keep the higher frequencies from being reflected back into the cone.
3. Damp the outside of the frame with EAR SD40AL. This constained layer material is the best I have heard and this will make the speaker sound way better at all frequencies. I doubt bituminous felt or other damping material sounds as good.
4. The baffle must be at least 2.25 inches thick using 3 layers of highly refined MDF (Rangerboard, Medite 3D) using green glue in between each layer. This will be a seriously dead baffle. Way better sound. Of course, you can also use bamboo plywood or other great wood in layers using green glue in between.
5. Brace the baffle with a huge brace that also is attached to the magnets of both woofers and is solidly attached directly to the baffle and to the base. This is mind blowing and an absolute must.
6. Hook up the woofers directly to the amp....do not use binding posts, spades or the like. Simply hardwire some PCOCC wire between the drivers and hardwire your speaker wire to one of the woofs....solder them directly.
You do the above 6 things and use a highly modified Behringer DCX24/96 crossed at 1K and you will swear you are listening to an electrostat! Of course you could get even more articulation by using smaller drivers between 100 and 1K......but if you think you are getting bad sound from an Alpha 15 it is because you have not done the above 6 things.....ALL of them must be done for truly great sound.
If you check out my gallery you will see a pic of a speaker done this way. The brace is very funky that my friend did but even so he says the improvement the brace made is comparable to the last $10,000 worth of other upgrades in his system. Cost him $50 for the wood, glue and hardware. He is running two Alphas to 900 and crossing to a Beyma compression driver mounted in a Geddes waveguide using highly modded Behringer. Incredible speakers!!!