Omnidirectional Speaker Project, any interest, help?

[dlneubec]

Dave Ralph's page shows how to take the tweeter apart. It looks pretty easy.

http://www.speakerdesign.net/dayton_...s_tweeter.html

Yup, that looks easy enough.

One thing his graphs reminded me of is the typical dip or notch in freuquency for the RS28 from about 13khz to 20khz. How come I don't see this in my front firing tests? Could it somehow be due to the small 6" dia. baffle it is mounted on, or did I just get a really good one?

[dlneubec]

Typical dome tweeter response narrows with increasing frequency so, with flat on-axis response, total radiated power is lower at HF.

Directing a relatively narrow HF cone into a broader doughnut shape means relatively low HF power must cover a broader area. Achieving flat frequency response with broad area coverage would require HF boost at the dome. HF boost could yield flat frequency response in a full 360 degree circle and would also provide a lower difference in power response across the tweeter frequency range.

Did I get that right or do I need more coffee?

At this point, I know very little to nothing about crossover design. What you say sounds good, but how do you go about boosting the HF output?

[Paul W]

For testing purposes, a resistor paralleled with a small cap, the pair in series with one side of the tweeter, will attenuate LF relative to HF.

Don't know what your crossover plans are, but most passive crossovers need a series resistor to pad the tweeter...so you could parallel that resistor in your crossover design...active EQ, turn up the treble control, etc.

[mmoeller]

I would use something like an inverse baffle step compensation. Basically just parallel cap and resistor in series. I simulated in a SPICE tool came up with 50uf and 4ohms. With a 4 ohm nominal load that should be a full 6db correction.

[JonP]

Yep, there are a few interested, Dan! I've been pretty busy past few days.. and it's late now... I'll have to read thru in depth, just kinda skimmed it tonight.

Thoughts:

Take some readings of the drivers bare without the crossover. This will answer the peak/driver cone break up question by seeing the natural break up peakx/valleys. When driving the tweeter, use a 20uF cap to protect it.

I also thought about that phase shield that the RS28's have, and if it was doing wierd things with the diiffusor. Would be interesting to compare it in and out. I'd guess that the diffusor might provide a similar effect, and it might not be needed. (or as much) And leaving it in migh create peaks or nulls.. But, it needs to be checked with real measurments.

The other thing I was wondering was where is your mic being placed for the measurments? I'm thinking that we probably need to measure at different angles to the drivers, to try to map the dispersion, but for starters picking a fixed point (45deg?) angularly from the axis might be a starting point.

As you're seeing in the SoundEasy documents, to start and measure the bare driver response on the baffle and diffusors would be a thing to do. Then you have a point to design/redesign or just look at the response of the current crossover.

I'll reread the posts tomorrow.. when I'm more awake...

Jon

[dlneubec]

Yep, there are a few interested, Dan! I've been pretty busy past few days.. and it's late now... I'll have to read thru in depth, just kinda skimmed it tonight.

Thoughts:

Take some readings of the drivers bare without the crossover. This will answer the peak/driver cone break up question by seeing the natural break up peakx/valleys. When driving the tweeter, use a 20uF cap to protect it.

I also thought about that phase shield that the RS28's have, and if it was doing wierd things with the diiffusor. Would be interesting to compare it in and out. I'd guess that the diffusor might provide a similar effect, and it might not be needed. (or as much) And leaving it in migh create peaks or nulls.. But, it needs to be checked with real measurments.

The other thing I was wondering was where is your mic being placed for the measurments? I'm thinking that we probably need to measure at different angles to the drivers, to try to map the dispersion, but for starters picking a fixed point (45deg?) angularly from the axis might be a starting point.

As you're seeing in the SoundEasy documents, to start and measure the bare driver response on the baffle and diffusors would be a thing to do. Then you have a point to design/redesign or just look at the response of the current crossover.

I'll reread the posts tomorrow.. when I'm more awake...

Jon

Hi Jon,

The first measurements I took were of the speaker with all three drivers, including the crossover. The measurements I have been doing and posting since then, have been for the individual drivers without any crossover, so you are looking at the raw driver results.

I'm measuring at 1M from the center of the upfiring driver or center of the baffle. The mic height is set at 34" high, which falls right between the tweeter and up firing RS180 baffles. All measurements so far are on axis, as much as that term applies. When the driver is upfiring and uses a centered diffuser, I'm not sure there is an on axis, so to speak. Where I have been measuring from would be on axis when the tweeter face is oriented veritcally and firing directly forward. Also, in the measurements where the tweeter is up firing and diffuser is moved to the rear, would probably be considered on axis.

I agree that off axis measurements are one of the next things I need to do. Right now I'm taking a few days off from measurements to try and read at least part of D'Appolito's book, Testing Loudspeakers, to try and get a better understaning of what I'm doing. I also want to run through more of John K.'s SoundEasy tutorial to try and understand SE a little better. I've got a lot of information to absorb.

Hopefully early next week I'll start another round of measurements. I found a couple other diffuser shapes for the tweeter I want to try. I also want to try testing the RS180 when the tweeter is firing directly forward and also the both the tweeter and the RS180 when the tweeter is firing forward, but with the baffle tilted back slightly, at a 15º to 30º angle back from vertical. I probably should take a few measurements on the downfiring Rs225 also on the current axis. Doing some comparison testing with the RS28 phase shield removed is on the list to do.

After these tests, I'm thinking then it might be the time to start testing off axis. I'd like to take the on axis tests and try to determine which setups to test off axis. Does this make sense?

BTW, I'm measuring in my living room with the speaker about 40" from both the front and side walls (to the center of the baffles).

[dlneubec]

I've been reading D'Appolito's book, a very good one, IMHO, and it got me to thinking a little more about diffraction issues with the circular baffle for the tweeter. According to what he says, I don't think it will not be a significant issue when listening at 90º to the driver axis, so when it is up-firing, it should be ok. Still, over the weekend I rounded the baffle edge over, just in case and recessed the tweeter so it is now flush mounted. However, since I'm testing front firing orientations for the tweeter, I thought diffraction might be more of an issue, though certainly less so with a smaller baffle like this. Anyway, I made a new tweeter baffle for the front firing combinations. It still is a 6"dia circle, but the edge is rounded over, the tweeter recessed flush and the tweeter offset so that the diffraction effects are not at the same frequencies all around the baffle (See photo's). Currently, it is set to a 15º tilt back from direct front firing.

I also got the bases for my Duevel-like diffusers glued up this weekend, ready to be routed on the CNC machine.

Unfortunately, I seemed to have run into a setback on the measurement front. I intended to do start some testing again and setup the measurement system and could not get a peep out of the M-Audio Transit. No matter what I tried, I could get no sound out. When I switched to my internal laptop soundcard, I got sound, so it is definitely a problem with the Transit. Removing and Re-installing it did not help. Trying a new USB port did not help. :letitout:

This can be a frustrating hobby!

[dlneubec]

I just got off the phone with tech support at Sweetwater.com, where I bought the Transit. They believe it is a defective unit and are replacing it. They are sending FEDEX to pick it up tomorrow, no charge to me. The guys at Sweetwater are great to work with.

The bad news is that the M-Audio Transit is backordered for at least a week. After talking to them, I decided to go ahead and try a different USB Audio Interface, the Edirol UA-1EX. It is the same price as the Transit, they suggest that it might be an even better unit. They are eating the extra shipping cost to send it to me Fedex 3day rather than ground, so it should be here Friday at the latest. If it doesn't work with SoundEasy, I can exchange it for a Transit later (of course I will pay the shipping then).
Edirol UA-1EX

The M-Audio has a USB and 1/8" in; optical and 1/8" out, that's it. The Edirol has a couple more options, including USB in, 1/8" TRS, RCA in; RCA, 1/8" and SP/DIF out. Both have 24bit/96kHz audio. I hope it works out.

[dlneubec]

Details, Details, Damn those Details!

John K. (JPK) has helped me to understand how the up firing woofer will be performing in 4pi space and the downfiring (about 6” off the floor will be performing in 2pi space. For that reason the upfiring woofer will be reduced considerably in its efficiency. For example, the starting drivers I have on hand for testing the concept are the RS180 (upfiring) and the RS225 (downfiring). The downfiring woofer should perform close to its stated rating of 88db, but the upfiring will perform at more like 81 to 82db rather than its 87.6 rating.

I have three options as I see it. One is to change the design to use two RS180’s and pick up some additional efficiency that way. This is doable, but does restrict other options in the design, probably necessitating a front firing tweeter, and making the whole thing taller.

Another option is to use a much higher efficiency driver in place of the RS180, with around a 94db rating. The only 7”-10” drivers I have found with that high a rating are the Pro drivers. The ones I have found so far are as follows:

B&C 8SP21 8” woofer:
Link
B&C 10HPL64 10” woofer:
Link
Eminence Beta-8A 8” midrange:
Link

I’m curious to know if anyone has used these Pro drivers in a home audio application, and how they perform and what recommendations anyone who has used them, or others like them, might have.

My third option, which has lots of potential problems all its own, is to use a polar array of 4 drivers, something like 4 RS125’s or RS150’s, firing up into a diffuser. This would bring the efficiency up, but at the expense of unknown comb filtering issues between the drivers. The drivers would be edge to edge in a 2x2 square pattern. If anyone has any experience with this type of array, I’d sure like to hear what you found.

This could end up a 3-way system, or the downfiring woofer could end up as a powered sub, something like the RS265HF, with a two-way MTM or TM top configuration.

I hope to do some more testing this weekend to see what output I do get from the three drivers I have in the test system now, to help determine more accurately what the differences in output will be.

[dlneubec]

The testing I did Friday night and Saturday seem to confirm that a single RS180, up-firing will not be able to keep up with a forward firing tweeter, so it will take either a much more efficient woofer, around 94db, or two RS180's.

I decided to test out the dual RS180 concept (see pics below). The concept I came up with was to sit a box for a second Rs180 above the tweeter firing down at the other RS180. The tweeter is now on a rectangular baffle that sits centered front to back, so the acoustic centers are very close to aligned. It is also centered vertically between the down and up-firing RS180's.

Right now I have it playing with the old 2.5way crossover built for these same drivers when they were in a traditional box MTM, with the .5 woofer the one on the top firing down. My NaO mini subs kick in around 100hz to fill in the bottom end, since the sealed Rs180's have an F3 around 70Hz or so. It sounds much bigger than it looks.

I plan to do more testing of the MTM concept this week. I want to do some more testing of the down-firing Rs225 (it will probably take nearfield to understand what to do with it). I also plan to pull one of my 12" DVC's to try in the base of this firing down, perhaps using my NaO Mini active LP to control the sub.

Looking more and more like a robot from and old science fiction, don't you think? :B

[Dennis H]

You can always pad down the tweeter to balance the SPL. Most speakers do that.

[dlneubec]

You can always pad down the tweeter to balance the SPL. Most speakers do that.

Hi Dennis,

Yes I could, but since the upfiring woofer is performing at somewhere around 82db sensitivity, to more like 91 for the tweeter, that seems like too much padding and would make for a pretty low spl overall. Even with two RS180's, I will still need to pad down the tweeter. On balance, I would prefer a more efficient system.

I have yet to test the system with the Duevel-like diffusers (which I might have by the next weekend), but I don't expect that to change the basic spl problem, unless the diffuser is effective enough to force the upfiring woofer to perfom in 2pi space instead of 4pi. I might still consider a Pro driver, at more like 94db minimum sensitivity, but I have no expereince with them and niether do many others on the DIY forums, so far as I can tell.

Next up is more measurements of the new system, including some off axis mesurements of the tweeter and upper pari of woofers and nearfield plus probably ground plane measurements on the donwfiring RS225. The I want to make a new bottom baffle to try out a 12" sub in the bottom. I have a max of about 38.6L on the bottom to work with for the low end driver, unless I want to make the whole thing about 5" taller, and it is tall enough already at about 49". I might use a sub in about 29L sealed at the bottom, take out one modular section and lower the height to 44" tall. The RS265HF 10" comes to mind. I believe I modelled it playing down to about 39hz sealed in 25L, IIRC.

In listening to it yesterday, with the 2-RS180s, the RS28 (using the old 2.5way crossover, the .5 woofer on the top and firing down) and a separate 12" sub filling in below 100hz, the sound from just one speaker was incredibly full and large sounding. I'm anxious to see what the measurements look like.

I said this to someone else as a way to try and describe it. A typical single box speaker might sound like it has 1/4" the fullness and soundstage compared to a pair of boxes. A single open Baffle (NaO Mini) sounds more like 1/3 of the pair, but the Omni's are more like 1/2 the fullness and soundstage.

[Paul W]

'nuther approach would be to array 6 or more RS180's in an outward firing ring below the tweeter. Floor loaded port out the bottom of a hexagonal+ enclosure. More 180's = roughly the cost of a sub.

[dlneubec]

'nuther approach would be to array 6 or more RS180's in an outward firing ring below the tweeter. Floor loaded port out the bottom of a hexagonal+ enclosure. More 180's = roughly the cost of a sub.

Hi Paul,

Interesting idea. I'm hoping to keep the acoustic centers and/or voice coils of the drivers aligned vertically, equidistant from all listening directions. This is one of the reasons for considering a sub rather than a multi-driver solution.

[Paul W]

Dan,
Yep..it's unusual to need to delay woofers but, if you end up going active, a digital crossover could provide the delay (as well as the HF boost for a full 360 degree tweeter).

No experience with them on woofers, but passive or active all-pass filter(s) might also do the trick.
Paul

[Patrick Bateman]

You can always pad down the tweeter to balance the SPL. Most speakers do that.

I believe this is one of the reasons that the commercial omnidirectional horns use compression drivers and prosound mids. You need a ton of efficiency to radiate omnidirectionally.

I'm using compression drivers for my current project, and even radiating into a 83 degree waveguide I need quite a lot of power.

:: PB ::

[j@ck]

Very interesting discussion.
If in a near-field setup in a smallish room, would it be a good idea to use some high efficient full range drivers firing upwards to start. Take measurements and add tweeter later on if needed?

[dlneubec]

Very interesting discussion.
If in a near-field setup in a smallish room, would it be a good idea to use some high efficient full range drivers firing upwards to start. Take measurements and add tweeter later on if needed?

There is a benefit to using high efficiency drivers, since there is a sensitivity loss because the up-firing driver is performing in 4pi space, rather than half-space.

I'm not certain about the full range driver concept. if you use a front firing tweeter just to augment the top end, then it would mean that more of the frequency spectrum would be in the upfiring mode, which might be good for "omni" effect but bad for overall sound quality.

As I understand it, as drivers get higher in frequency, they get more directional and go from naturally omni to naturally directional. That may mean that your off axis response will roll off very quickly and much of the higher frequencies will be lost anyway.

What I have found, for example, with the RS180 firing up, it starts to drop off quite quickly at about 1500Hz and drops about 23db by 4000Hz, where cone breakup starts. However, if you look at the RS180 in inifinite baffle tests, it doesn't start to drop off substantially until close to 3khz, drops around 7db to 4000hz and then the cone breakup starts.

Also of interest, is what happens in the cone breakup area. The RS180 on infinite baffle shows peaks in the breakup area of around 10db higher than the normal respnse peak, while when it is upfiring, those breakup areas are suppressed naturaly to around 8-10db lower than the normal response, which I assume will make them easier to deal with in the crossover.

[dlneubec]

I forgot to mention in the response above, that the same thing happens with the tweeter, if it is upfiring into 4pi space. The RS28 starts to drop off at 3k and drops almost 15db to 15kHz and then drops off very steeply.

The downfing woofer, however, stays in 2pi space, bacause of the proximity to the floor and performs very similar as its standard application.

The diffusers are intended to redirect that lost db, but so far, it is only partially succesful in all the tests I've run. The side effects are a much more ragged response at the top end, so much so that I'm heading toward front firing for the tweeter.

I will be getting my 13" wide Duevel-like diffuser early next week and I will test those to see if they perform any better than the ball-like ones have so far.

[joecarrow]

Hey, I know I kind of dropped out of this thread, but I'm still reading along. Have you had any thought of trying to emulate the geometry from the Bang and Olufsen Beolab 5- the stuff I linked to at the beginning? As I understand, the radiation of their tweeters is really strongest over 180 degrees.

[dlneubec]

Hey, I know I kind of dropped out of this thread, but I'm still reading along. Have you had any thought of trying to emulate the geometry from the Bang and Olufsen Beolab 5- the stuff I linked to at the beginning? As I understand, the radiation of their tweeters is really strongest over 180 degrees.

Hi Joe,

I looked at the B&O and it looks interesting. I'm not sure how I'd build something like that. It would be a pretty complex solution, especially for DIY. My hope was that I would end up with something other DIYers could do. This is one of the reasons I sort of dismissed the idea of trying to do one of the more complex Duevel hornloaded diffusers with the compression drivers. I thought it would just be too complex.

I'm sure you could do something similar in concept to try and push the tweeter sound into 180 radiation instead of 360. Maybe you could cut a 4-8 sided cone at 45º with a table saw, then round off the edges to get a relatively smooth cone. Cut the cone in half and mount it on a solid back plane piece with the edges of the back plane chamfered or rounded over and then mount the assembly right behind the dome. The parabolic cone would be difficult to accomplish and so would the little parabolic cut out area right above the tweeter. My concern would be that after all that effort would it be much, if any, better than just front firing the tweeter?

I also noted that they seem to have two drivers in the mix. I'm guessing one is a tweeter and one a supertweeter. That could be due to problems with a single tweeter rolling off in response as it gets more directional. It could also be a dome midrange and tweeter, its hard to tell from the photo.

[joecarrow]

As far as I can remember, it's a 3" dome midrange and a tweeter... I want to say a 3/4", but I'm not sure about that. The large dome mid made an impression on me when I saw them in person, but I was distracted by the overall effect such that I wasn't paying much attention to the tweeter's size.

It looks like the nearest B&O store for you is in Chicago, so I can't imagine a special trip for you to get up close and personal with them- but if you're in the Windy City for other reasons I'd recommend it. Their active room correction is really impressive. In addition to all of the fancy dispersion tactics, they also have some sort of microphone setup to feed a DSP to correct for your room. As far as how bass guitar, bass drums, and low frequency effects go, I would say that this system sounded like it had the least influence from the room out of any I have ever heard. That's not saying much considering my lack of experience with dipole bass or domestic bass horns, but it was impressive nonetheless.

If I was going to do this, I'd probably try to do a negative in plastic sculpting clay, and then cast the rest in plaster. It wouldn't look as nice, but it's a lot better than finding a CNC mill.

[dlneubec]

I bet the B&O's are impressive. I'd love to hear them, but not bad enough todrive 5hrs one way to do so. :E

BTW, I do have half of one of my Duevel-like diffusers, the one for the tweeter, that is, like the top half of the one in the picture below. I'm having the top and bottom sections of that diffuser made in two pieces that I can sandwhich together, if I want to, or I can put two bottom halfs together to work for two woofers facing one another.

I did do a little listening with the tweeter just sitting on legs above it. I used the RJB 2.5way crossover I have. I hate to say anything before I take measurements, but to my ear, it was much, much better sounding than any of the ball configurations I tried, which just plain sounded bad when compared to direct firing. This sounded a little laid back, as if it was indeed diffusing the sound such that the SPL output from on axis was lower than that of the 2 RS180's. However, it might be a good match to one RS180 or 2, with a higher efficiency tweeter.

The ball shapes seemed to cause some serious resonances between the tweeter and the ball, which were obvious when you listened to them. This diffuser didn't seem to have the same sound. Measurements will tell a lot more, but I'm a little more hopeful that these diffusers might have more potential than the ball or dome options I have tried so far.

[j@ck]

Hi dlneubec
Very insightful information. Thank you.
I have some very basic measurement equipments but no woodwork tools nor skills. Now if I can find someone do the box for me then I can start experimenting.

I just found this online, the GR-Research AV-O and O-3.

http://www.gr-research.com/kits/o3.shtm

It seems a promising design with good measurement results.

[dlneubec]

Update!

I tested top half of the duevel-like diffuser with a downfiring RS28, configured as shown above. I tried it with the tweeter mesh grid from 1-1/4" down to about 1/2" from the diffuser. The results were improved over the ball options, but only out to about 8k, after which the response got extremely ragged and dropped of serisously. Unless something else comes up to test, I'm pretty set on using a front firing tweeter.

As you might recall, in order to offset the sensitivlty loss due to 90º off axis listening and performance in 4pi space, I had to add a second RS180 in a box above the upfiring RS180 (pictured in a post above). In testing the front firing RS28, a dip at around 3kHz got much more severe after adding the second box, I assume due to interactions between the tweeter baffle and the two boxes. As a result, I have been trying different tweeter baffle shapes and vertical positions because to try and address this deep dip with the rectangular front firing tweeter and two boxes. The best response for my original rectangular baffle is shown in the fisrt graph below. There is a dip at 3khz of about 12db, as you can see.

The circular baffle, with the tweeter offset, offered some improvement. It was a little smoother on the top end, and the dip was shallower, but wider.

I then tried a third rectangular baffle with some different types of edge chamfering treatments and I think this is the best one. The second graph shows the response. There is still a dip in the 3-4kHz area, but it is much shallower and wider. In the end, I think this will suit mys listening tastes best anyway. I have been advise that A little dip in the response around 3 or 4k often can take the edge off lesser quality recording and may prove better for HT as well.

I forgot to mention that the first two graphs, are raw measurements, without crossover.

The third graph is with the two RS180's and RS28 all connected. I simply added a 2.5ohm resistor in line with the RS28 to attenuate it a little, but otherwise, they are the raw drivers combined. The graph is only good down to about 250Hz-300Hz, because of the small gated window used in the measurement.

I also pulled the RS225 from the bottom and installed a PE 12" DVC subwoofer (the more expensive one) bottom firing and used short legs by a couple inches. I'm using a 300w plate amp to power it. In this scenario, I will end up with either a 2way or 2.5way MTM on top with an activly contriolled sub on the bottom, probably crossing at about 100k.

Just to get an idea how this all sounds, I connected the RJB 2.5ay crossover I already had that was designed for the RS180/RS28 combo, but in a traditional front firing box and let the bottom end roll of naturally. The sealed RS180's have an F3 of around 74hz. The top box was setup as the .5 woofer. For the 12" sub, I used the active LP filter designed for this sub for my NaO mini's. It pushes the sub to a QTC of .5 , 20hz alignment with the top end rolling off at 100hz.

I can't tell you how excited I am about how good this sounded, especially since the crossover is not designed for this configuration. The sound is so incredibly open, it is hard to describe. The off axis performance it extremely good, though you can hear the highs drop off as you get off axis, sinc the RS28 is front firing. The RS180's on the other hand, seem to sound the same at about any position, so the omni effect seems to work very well, without any diffuser at all.

I'm supposed to get my other Duevel diffusers on Tuesday, so I will test the response with them, but I have a feeling that what I have now will be the final driver layout and arrangement, unless I decide to test a single high efficiency Pro-type driver instead of the two RS180's.

Next up is crossover design, which should be interesting, since it will be my first! I'm anxious to see how much better it will sound with an optimizes crossover. :T

I'll post some more photo's this afternoon of where I'm at.

[dlneubec]

More Progress Photo's below:

1) Overall
2) Sub installed base and box apart (again!)
3-5) New Tweeter Baffle, various angles

Note: I added those wool insulation strips (1/8"x5/8") just to see if would help smooth out the response. I don't know if that is the right kind of wool, but it was all I could find locally. The graphs I posted above are without it. In the measurements taken after, I thought that it might have helped smooth the tweeter, a little, but where I saw the most change was in the cone breakup modes of the RS180, above 4kHz. The SPL and amplitude of the breakups was reduced, so I'm assuming what I'm seeing above 4kHz is a combination of cone breakup, the interaction of the two RS180's firing at one another with the tweeter baffle in between. I have some F13, 1/4" wool felt on its way to test this further. I might coat the entire tweeter baffle with it. I might also try some other sound absorbing materials on the back of the tweeter baffle also.

[ThomasW]

Like the Energizer Bunny, Dan keeps going, and going, and doing..... 8O

The plot is much flatter than I expected given the newest configuration.... :T

It's always a tell-tale sign of pending success when wool-felt starts arriving on the scene.... :B

[dlneubec]

Like the Energizer Bunny, Dan keeps going, and going, and doing..... 8O

The plot is much flatter than I expected given the newest configuration.... :T

It's always a tell-tale sign of pending success when wool-felt starts arriving on the scene.... :B

Well, I had and have a lot to learn, so persistance is much needed. I probably couldn't have picked a much more challenging design to start with.

Not being crossover design aware yet, I assumed that I needed to work on the physical design and get the raw response as good as I could first, thinking that it would make the crossover design easier and more successful, especially with something as unconventional as this.

I've gotten some really good advice from John Kreskovsky along the way which has been a great help.

What I need soon is a name for this project. One idea I had was to call it the Mentor1, because it has taken lots of folks mentoring to be able to get this far and will take more to get it done, or the Tyro Project (another word for a novice).

Anyone have some good name ideas?

[dlneubec]

Hi All,

Just a little update. I've been messing with some 1/4" (f13) wool felt on the front firing tweeter baffle to see if it would help smooth the response of both the tweeter and the RS180's. I now have felt covering the back of the tweeter baffle and probably about the top and bottom 1/3 of the front and it has most certainly helped in both cases.

Also, what looks like cone breakup in the 3k and up area is probably a combination of breakup modes and diffraction as the RS180 begins to get more directional at higher frequencies. This is apparent because the felt had the most effect on the freq. resp. in this area.

Below are the final measurements I will take into crossover design (I hope). The RS180's are farfield and nearfield merged at 290hz. The nearfield was taken in the final setup, with the tweeter baffle in place, both RS180's playing. I believe the ripples you see in the response at around 200Hz are a result of diffractions with the tweeter baffle that would not show up in a normal nearfield measurement. The results were recorded with 6db less added to the graph, so that I can add each RS180 as a single driver in SoundEasy (per JPK's instructions).

the next step, crossover design, will be a huge one, since it will be my first! :unsure:

BTW, I've decided to finish this design using the RS drivers in the current configuration through crossover design and build, rather than get into trying any PRO drivers, at least to see how it turns out. I think it will be more acceptable as a DIY project for others if I stick with well known, standard, and relatively inexpensive drivers, plus I already have these drivers, so why not try.

[dlneubec]

Hey all,

I've finished adding 1/4" felt to the tweeter baffle. A couple of pics are below.

Also, I did farfield measurements of the RS28 and the two RS180's at 0, 15, 30 and 45 degrees to the right of axis. They are plotted on one another in the graphs below. Blue is on axis, red is 15º, green is 30º and black is 45º. The off axis of the RS28 looks pretty good I think. The off axis performance of the two RS180's indicates just how well the omnidirectional response is on the up and down-firing drivers. The four plots are virtually identical through about 2kHz.

It's also interesting to note that I did semi-nearfield measurements centered between the RS180's, but at the front edge of the speaker (6.5" back from the center of the drivers). The nearfield response is much flatter and more extended with the tweeter baffle in place than with it removed.

[dlneubec]

I'm working away on crossover design in SoundEasy, which is a big challenge for me on my first crossover design.

In the meantime, I have my Duevel-like diffusers. I added two of them stacked to use as diffusers for the dual RS180's and moved the RS28 to the top of the box. Below are a couple photo's.

Also below are some measurements of the Duevel-like version. First, is a nearfield-farfield merged response for the dual RS180's. They are set 6db lower than the tweeter response (next graph) so that when 2 drivers are added in SE, they will sum 6db higher. The RS180 response is more extended than it was with the other configuration. The tweeter response is also flatter. Bear in mind this was with a 5ms window, rather than 3.04ms I used before and that generally makes the response a little bumpier overall.

I'm wondering if this configuration is worth further exploration (crossover design). It looks like it might accept a little higher crossover point, which may mean a little more extended true omnidirectional response.

[ThomasW]

To me the Duevel-like diffusers look the be the better concept, I'm curious as to your listening comparisons.... :T

BTW, I was wandering around the PE board, and it looks like the competition is heating up .....:wink:

Image not available

[dlneubec]

To me the Duevel-like diffusers look the be the better concept, I'm curious as to your listening comparisons.... :T

BTW, I was wandering around the PE board, and it looks like the competition is heating up .....:wink:

Hi Thomas,

I saw that on the PE board this morning. I wonder if he has taken any measurements on it. I'd be willing to bet he has all kinds of problems with the tweeter response, rolling off drastically above about 7-8kHz or so, resonances between the tweeter and the ball, plus diffraction problems from the tweeter and woofer baffles, etc. Also, he is going to have issues with needing to add delay to woofer since it's acoustic center is so far ahead of the tweeter, not to mention the roll off problems with the woofer, when listening close to 90º off axis.

Until I saw the measurements and tried lots of different configurations, it was impossible to figure out what was happening.

As to the Duevel-like diffusers, I'm convinced they can help some with the low frequency driver dispersion and perhaps the low end of the tweeter, but my measurements show they are of little help in the response above the 7-8Khz threshold. I think that's why Duevel uses a compression driver and a real horn for the tweeter in all but their lowest end systems.

One major problem with using two woofers in the arrangement I showed above with the back to back diffusers is not the woofer response, which is, in fact, extended a bit, perhaps allowing a higher crossover point, it is with where to put the tweeter and not get into delay complications or tweeter height issues, etc. My plan is to finish the design as is without the diffusers, which I think will probably be unnecessary. This keeps the acoustic center of the tweeter aligned and centered between the two RS180's and the height of all drivers function somewhat like they are the same. That way the sound is much more like it is coming from a single full range driver, than multiple drivers.

My major problem at this point is the crossover. I've done a bunch of designs on my own. Then I took the best of all I could find on the net for the RS180/RS28 combo, including the NatP, the Modula MTM, Dave Brown and Roman Bednarek's (2way and 2.5way) and attempted to adapt them to this design. In all cases, to get a good impedance response, 30db or higher suppression of the breakup modes of the RS180 and a reasonably flat response, it seems to push the crossover down into the 13-14Khz range and I'm concerned that could be a problem for the RS28.

Below are a couple of my crossovers and an adapted one as examples. Anyone care to offer some ideas or thoughts on these designs, since they are my first? I could sure benefit from some advice at this point.

[ThomasW]

I was just goofing on you with the picture from the PE thread. Obviously that design is plagued with issues....:wink:

Do you have access to a DQX2496? That would make your crossover questions easier diagnose.

I can see the plots but can't read the component lists. The plots are so similar I can't see how there would be major sonic differences between them.

[dlneubec]

I was just goofing on you with the picture from the PE thread. Obviously that design is plagued with issues....:wink:

Do you have access to a DQX2496? That would make your crossover questions easier diagnose.

I can see the plots but can't read the component lists. The plots are so similar I can't see how there would be major sonic differences between them.

I don't have access to a DQX2496. :cry:

I'm finding it difficult to get readable graphs and lists within the 100k limit, at least with screen captures, which have to be resized and compressed to fit. I'll see if I can post some plots on Photo-bucket with links that are more readable.

[dlneubec]

Photo-bucket was even worse. Here are some images that I cut and pasted to make smaller. I hope they are more readable.

[ThomasW]

Attachments are to limiting for this, so just post links to bigger files

[dlneubec]

Any free services out there that will let you post larger files to be linked to?

Here is one more attempt at just one done a different way. Is this any better?

[dlneubec]

That one looks readable to my eyes. Here are the other two.

[ThomasW]

I can read the component list on those if I squint.....

Places like photo-bucket, putfile and others allow links to bigger files for free.

[theSven]

And the problem here in lies that by asking people to use external photo storage many of the photos from builds are now lost. Thankfully under the new site owner storage is not an issue so upload away!

[Dennis H]

Save them as gif files and they will fit easily and be much more readable too. It will reduce the colors to 256 but after that it's a lossless compression that doesn't distort the image.

[dlneubec]

Save them as gif files and they will fit easily and be much more readable too. It will reduce the colors to 256 but after that it's a lossless compression that doesn't distort the image.

Thanks Dennis! That worked like a charm. I also turned off the cad nodes which removes the dots.

See the three files below as gifs.

[JonP]

My major problem at this point is the crossover. I've done a bunch of designs on my own. Then I took the best of all I could find on the net for the RS180/RS28 combo, including the NatP, the ModulaMTM, Dave Brown and Roman Bednarek's (2way and 2.5way) and attempted to adapt them to this design. In all cases, to get a good impedance response, 30db or higher suppression of the breakup modes of the RS180 and a resonably flat response, it seems to push the crossover down into the 13-14Khz range and I'm concerned that could be a problem for the RS28.

Wow, Dan! All this progress, despite having Thanksgiving thrown in the works!

Me, I've had numerous work situations taking up time, (including a 7 day biz trip on 2 days notice) etc... and I'm not building anything..

Looks like great progress in discovery on how these things work.... and learning SoundEasy didn't slow you down to a crawl! It's looking very good, those response curves are pretty flat. Very interesting how things are turning out different than the example of the Duval speaker... this is real Science in action.

On other notes... 1300-1400Hz for a RS28... As you probably know, that is sort of pushing it, though some stay fairly low distortion down below that. It would be good if somehow someone did a testing to find out how variable they are.. or if PE has gotten them consistent with greater manufacturing experience... I think they're on sale now, there's always buying a few and testing for "good" ones, then selling the others to folks who are crossing higher? SE enables you to test for HD problems... you can check yours out, and others if need be. (reminds me, I have a pair that I still haven't done that to yet....)

Another thing... if you are up to a little more experimenting, and have a stereo amp handy, try playing with the DF function of SE... it will really be fun to listen to your crossover designs as you develop! I've done this, it's fairly easy, and if you're just simulating one 2 way, you don't need a multi channel sound card, only one stereo pair for outputs... Not sure if you can measure the response (with the same computer) due to time delays... that's where the DQX2496 would come into play.. (create filters, D/L into digital crossover, measure with your PC and SE)

Though.. the just released V13 of SE is supposed to have much lower latency than before.. I was seeing 1-1.5 seconds of delay, :E while the new one should be in the 100-200mS range. I'm wondering if that could be dealt with by SE's MLS system. And if my computer would run out of clock cycles trying to filter and measure at the same time...

Great work...

[dlneubec]

Hi Jon,

The RS28's I'm using are the AS or shielded version. The sample response plots PE publishes on these look quite different than the unshielded version. I don't know which has the best potential for crossing low, but the sample FR of the shielded version is flat down to 1kHz, while the unshielded sample starts rolling off quite a bit sooner.

I'll look into the digital filter section in SE. I may have to contact you with some questions as I get into it. It would be great to b able to simulate crossovers before you buy the parts, and at no extra expense. :T

I've also not looked into doing any distortion measurements. I believe I read a note from Zaph, maybe on the Madisound forum, saying that this is greatly enhanced in V13.

[mattmmm]

omni diffusers

Hi,

I'm new to HTGuide, I had a bit of skim read through your posts, really great you guys are thinking outside the box to make something different that sounds good. I've finished some DIY speakers for omni sound with diffusers also, right now I'm not even using the diffuser, just using a forward firing mid woofer and also an upfiring woofer which provides omni sound and a forward firing tweeter. the speaker box is not that high off ground and tilted forward so the upfiring midwoofer is not too off axis, I also lowered the X over frequency to 1.8Khz so the tweeter takes over where the upfiring midwoofer can't provide off axis.

I orginally tried with the diffuser, but found that it had a boxy sound (affecting frequency response between driver/box and diffuser etc). I also originally had the tweeter down firing like Duevel, but I didn't have the measurement equipment to check it right, the tweeter is also quite small compared to the Duevel tweeter etc so might be hard to get right for different reasons; shape size distance etc...I didn't bother experimenting that much. But anyway I think (aswell as other speaker designers) that a front firing tweeter is still a good idea even with omni speakers, generally speaking I guess higher frequency sound is more directional and typically often absorbed more by furniture etc than the lower freuqncies that bounce around the room...

I'm actually thinking of experimenting with the diffuser again, trying different heights (having the diffuser further away/angled from driver may get rid of frequency response problems?) just with midwoofer, tweeter will stay front firing, the diffuser also looks cool. With a front firing woofer and tweeter plus omni woofer I think safety in getting best of both worlds.