View Full Version : The Miniliths: A project based on the Seas CA21REX, Neo8 PDR & Neo3 PDR

07-23-2007, 12:37 PM
So, I had a couple of weeks to start some work on my Neo-based project. Since I already have a pair of Monoliths (my last project), and I'm considering building the Megaliths (an MTMWW tower), I thought the Miniliths name would fit within the "series".

Anyway, as always, here are my objectives for this project.

1) An accurate 3 way monitor speaker.
2) As small as possible.
3) Mostly for nearfield listening (as huge PC speakers), or semi-full range speakers (need a sub to sound their best).
4) Able to be biamped (my main idea is to use two T-Amps with a bit regulated power supply).
5) Accurate - as flat as possible (a huge issue, considering the FR of the Neo3).
6) Monopolar.
7) High WAF.

I have the following issues.

1) FR of the Neo3 & Neo8.
2) Crossover slopes and cut-off points.
3) Diffraction issues.
4) Mounting the Neo3 and 8.
5) Size of the cabinet.

More to come...

07-23-2007, 12:39 PM
I decided to go sealed. Considering that my available 8" woofers (in case I didn't like something about the Seas) could work well with about 12 liters (the Dayton 225, and the Audax AP210ZO), that's what I used.

Here's the layout and size.


1" MDF all around, veneered. Nothing fancy, just something practical and nice.

According to Edge, diffraction shouldn't be an issue with drivers this big. We'll see.

07-23-2007, 12:48 PM
Aren't there always issues? :W

Sounds like an interesting project, Javier- anything special in mind for the size and shape of the cabinet?

I'll be interested to see how this turns out for you.

A lot of three ways popping up lately... :B

07-23-2007, 12:51 PM
Aren't there always issues? :W

Sounds like an interesting project, Javier- anything special in mind for the size and shape of the cabinet?

I'll be interested to see how this turns out for you.

A lot of three ways popping up lately... :B

Hi Jon!

Yep, I tried to immitate a MartinLogan in a smaller size (yeah, things don't always work that way, as you'll soon see!).

And a lot of issues came up. Good thing is, I'm in no rush to finish this speaker.

07-23-2007, 12:53 PM
Here's the plot of FR for the cabinet. Nice, not to deep a response, but OK. Measured with the Woofer Tester 2, I got a nice 0.64 Qts.


Here are the finished baffles, with wiring included.


07-23-2007, 01:10 PM
According to my estimates, the speaker would be an 89 dB/W model...

But something's definitely wrong here. Either I got a pair of macho woofers, or...

Here's the woofer's FR plot, according to SEAS.


Almost 90 dB, considering baffle step. Not bad!

The Neo3 and Neo8 graphs have their lowest point at 90 dB, so I assumed I wouldn't have any problems here. Well, not really.

Here's the FR of my first iteration of the crossover.


Hey, not bad, considering it's the Neo3 and Neo8 PDR...

Here's the impedance plot:


A little bit too low for my taste, but still not bad... hey, there's something wrong with the bass unit. See if you can guess what the problem is...

Well, here is the problem:


See that resistor on the woofer leg of the circuit?

Yep. The woofer is more efficient than the mids or tweeters. By a whopping 3 dB.

This can't be happening.

Right now, the only thing I can think of is that I overstuffed the tweeter and midrange cavities. I used a lot (A LOT) of dacron. Basically, the Neos chassis are pushing against it.

I surely hope this is the problem. I need to correct it and take measurements again.

Here's the response without the padding resistor.


And my problem turns out to be a bit more tricky. Even if I the tweeter and mids respond well to taking out some dacron... the impedance on the woofer drops below 2 Ohm when trying to use a 4th order electrical filter.


What do you think about this problem? I think I need to do a couple of things.

1) Re-measure the speaker without so much dacron.
2) Re-design the woofer / mid portion, shoot for a 2nd order filter at 800 Hz.

If this doesn't help, I'll switch drivers and go for the Dayton RS woofer... with an 8 Ohm impedance and lower sensitivity it may help.

07-23-2007, 01:59 PM
Have you verified the voltage at the speaker terminals?

Also, at what distance are you measuring?

Do you know what the baffle conditions were for the b&w measurement you've provided are? If it's effectively dipole....


Dennis H
07-23-2007, 02:10 PM
Is the planar baffle sealed on the back? If so, I think B&G recommends at least at least 5" of depth and light stuffing for the big RD planars which cover a similar freq range as the Neo8.

07-23-2007, 02:12 PM
How are you handling BSC?

07-23-2007, 02:14 PM
Hi C,

I did verify the voltage... well, I think that's what JustMLS does when in 2-channel measurement mode. It has never failed me before.

I measured at 1.2 meters. That distance seemed to help regarding the FR of the planar drivers.

The baffle conditions are: 10L box, anechoic. I don't think the baffle dimensions are shown on the graph (I got it from the SEAS web site).

Dennis, the baffle is sealed, indeed, but it has only a 1" depth. Yikes. I surely hope the Neo8 doesn't require a 5" depth - if not, I'm in serious trouble!!!

Thanks a lot for your help. :)

07-23-2007, 02:16 PM
Did you splice the nearfield data of the woofer with farfield and forget to add in BSC? The woofer data may be incorrect.

07-23-2007, 02:29 PM
How are you handling BSC?
Via the woofer circuit. As far as I can tell, there's no BSC on the midrange (I'm crossing over at 700 Hz).

I calculated the correct woofer response by doing the usual things... nearfield, farfield, KHF tool...

07-23-2007, 02:32 PM
Did you splice the nearfield data of the woofer with farfield and forget to add in BSC? The woofer data may be incorrect.
Heh, that was exactly the first thing I thought of! :) Still, I just re-checked (just in case) and I'm using the corrected woofer response.

07-24-2007, 02:02 PM
Since I won't be able to re-measure the speaker until later this week, I spent some time playing with the x-over layout.

I changed the woofer's 4th order electrical to a 2nd order electrical, and tried to hit a 4th order acoustical slope. The result was great from an impedance standpoint. Then I shaped the response so as to create a downward slope (a bit of warmth).

In case the measurements are correct, I'll be using this crossover as a starting point for the definitive version.

According to my measurements, off-axis response will be pretty good, although it's only a simulation. I'll take 30 and 60 measurements this weekend, to be on the safe side.

So far, pairing the Neo3 and the Neo8 has been most helpful. The Neo8 is far easier to work with. Also, removing the Neo3's small baffle and using a sealed box extended its FR by at least 600 Hz!



New circuit

07-24-2007, 02:16 PM
I bet you could get away with softer slopes mid-tweeter. 2nd order acoustic perhaps. Might help with various things and will obviously simplify the circuit.


07-24-2007, 02:26 PM
I bet you could get away with softer slopes mid-tweeter. 2nd order acoustic perhaps. Might help with various things and will obviously simplify the circuit.

Thanks a lot for your input (again!) C. The reason I rolled the Neo3 so fast was because I've had bad experience trying to work around its FR, and when I saw the Neo8 was actually flatter, I decided to mainly use the 3 almost as a supertweeter.

One thing I've noticed is that after removing its back enclosure and giving it some room to breathe, its FR measured a bit better.

I'll try 2nd order, just like you said. I tried doing the same between woofer and Neo8, and the results weren't that great... at 700 Hz, actually, the 8's roll-off was basically 2nd order acoustic!

07-24-2007, 03:09 PM
Yeah, I'd still keep the target frequency where it is - you'll still get primary contribution from the Neo8, but you may pick up response levels a bit more and even out some of the peakiness that exists in the top octave.


07-25-2007, 10:58 PM
Forgot to add - I usede Meniscus' Audio metal plates for mounting. I can't recommend them enough - they look as if they were an integral part of the Neo's. Awesome product at a great price.

Anyway, I've been playing with 2nd order slopes. I have an issue... the drivers' aren't in phase, so I have even worse of a frequency response.

I'm thinking about emulating the crossover and listening to it with both crossovers. The 2nd order ones are a lot simpler.

I also need those off-axis FR plots...

07-26-2007, 12:03 PM
Interesting project Javier. I guess all the bad stories about B&G Neo drivers tell only that they need a bit of trial&error to get good performance out.. many people don't bother to try enough.

I came across this:

Looks like Neo8 with quite a low price. Do you guys have any idea what's the story behind that one?

07-26-2007, 02:06 PM
Anyway, I've been playing with 2nd order slopes. I have an issue... the drivers' aren't in phase, so I have even worse of a frequency response.

Keep fiddling. Perhaps the target frequency has to shift a little. Or try switching to a series hybrid. I often find that I have to fiddle quite a bit before things really gel, but when they do it's SO nice. How far off is phase? Maybe a 3rd order slope would be a better fit. Or an asymmetrical mix, 4th on the Neo8, 2nd on the Neo3. Or...

This is the fun part, you know?! :)


07-28-2007, 12:39 AM
OK, I finally re-measured the speakers and took off-axis measurements. I then compared the measurements to the original ones, and used the off-axis measurements to find out which version of the crossover looks better (2nd order vs. 4th order).

1) The original measurements are correct. The Seas woofer seems to have a higher sensitivity than the Neo drivers. I measured voltages, measured every curve, etc. Removing the stuffing from the drivers seemed to make their responses a bit more ragged - sensitivity was unchanged.

2) Off-axis, the 2nd order filter has some major suck-outs on the mid-tweeter crossover region at 30. While, at 0, both drivers sum perfectly well, at 30 they look to be out of phase.

3) Off axis, the 3rd order filter has a flare at 5 KHz. I had to shape the response curve to mitigate it as much as possible without affecting FR on-axis.

So, for now, 4th order acoustical seems to be the best bet. I'll still keep on playing with 2nd order... or different topologies.

I'll post some graphs later today, or tomorrow.

07-28-2007, 08:39 PM
I should have suspected it. There's a reason there aren't many 3-ways being designed in this forum.

3-ways are hard!

Especially when the drivers are physically big, and they don't have nice horizontal dispersion.

At this point, I'm seriously considering just unplugging the Neo3 PDR.

The issue?

Off-axis response! Due to their horizontal placement and (I assume) the distance between drivers, off axis response is terrible!

Depending on the listening axis, response develops major suck-outs - when I use a 2nd order x-over, the only difference is that they appear at different angles.

I seriously don't think this issue can be solved, unless I raise the x-over frequency and make the Neo3 irrelevant. But I'd like to avoid the peak in FR from the Neo8.

Any ideas?

(The only FR with nearfield / farfield is at 0).


(My next speaker WILL be a 2 way).

07-28-2007, 08:46 PM
looks like the tweeter phase should be reversed, with that null off axis, so maybe instead of doing that, you could shoot for a lower crossover point to avoid the cancellations you are getting in the higher frequencies.

07-29-2007, 11:41 AM
So far, no luck.

I spent all night playing with X-over points. From 4KHz to 12 KHz, and every time I get a null at 15 off-axis.

At this point, I can't help but realize this design is flawed from the start.

The reason why I placed the drivers side by side instead of top-down was because I saw a MartinLogan which used this baffle layout. I suspected they did this because of the limited vertical dispersion of the Neo8, and thought it was a good idea.

Now I realize that a better idea would have been to do a T-M-W and design using the M as the listening axis, letting the Neo3 fill in the upper octave with its wider vertical dispersion.

I'm still wondering whether to build this speaker, or re-design with a new cabinet.

07-29-2007, 02:23 PM
So far, no luck.

I spent all night playing with X-over points. From 4KHz to 12 KHz, and every time I get a null at 15 off-axis.

At this point, I can't help but realize this design is flawed from the start.

The reason why I placed the drivers side by side instead of top-down was because I saw a MartinLogan which used this baffle layout. I suspected they did this because of the limited vertical dispersion of the Neo8, and thought it was a good idea.

Now I realize that a better idea would have been to do a T-M-W and design using the M as the listening axis, letting the Neo3 fill in the upper octave with its wider vertical dispersion.

I'm still wondering whether to build this speaker, or re-design with a new cabinet.

If you are using 4th order acoustic slopes in a 3-way for HP, LP, and bandpass, all drivers should be in positive polarity. It appears that you have the tweeter reversed, and although on axis is flat, it is not where it should be for a technically correct system.

So, reverse that tweeter polarity like I suggested earlier, before you scrap this project because I think that is the problem.

07-29-2007, 02:27 PM
Hi Jed,

The graphs I posted were for the system with everything with positive polarity. I made an error in my original measurements - the tweeter polarity was reversed.

I just modeled the speaker in a TMW arrangement, and I see there's no tweeter cancellation if designed with the listener axis on the Neo8. When I model it as I have it right now, I can see that the cancellation is there (memo to myself: model before building!).

As it is, right now I'm wondering whether I should build the speaker correctly, or just live with the FR for a while.

07-29-2007, 02:37 PM
Hi Jed,

The graphs I posted were for the system with everything with positive polarity. I made an error in my original measurements - the tweeter polarity was reversed.

I see, that tweeter phase still looks wrong to me. Are your slopes true 4th order, or more like asymmetrical?

07-29-2007, 02:45 PM
True 4th order.

07-29-2007, 02:49 PM
Would you mind posting a screenshot of the system when the tweeter is in negative polarity?

Another option might be a steeper slope between mid and tweeter.

07-29-2007, 04:37 PM
Would you mind posting a screenshot of the system when the tweeter is in negative polarity?

Another option might be a steeper slope between mid and tweeter.

Sure thing. Great idea on the steeper slopes - I tried slower ones.


07-29-2007, 05:18 PM
Thanks, but I can't see your pics, broken links.

07-29-2007, 07:24 PM

Here they are. That's strange - I was certain I had uploaded them before.

07-29-2007, 08:33 PM
Here they are. That's strange - I was certain I had uploaded them before.

Lookin better and how is the off axis now?

07-29-2007, 10:32 PM
Lookin better and how is the off axis now?
It isn't still how I'd like it to be, Jed.

Still, I've noticed something. I quickly built a higher order crossover, and the cancellaton isn't anywhere as severe as a 4th order (I went for a 6th order), the FR is flatter... and I found out that when tweeters are placed on the outside, cancellation doesn't occur at all. Tweeters on the inside are problematic.

This could be the solution to this problem. As long as the listener is in the sweet spot, the speakers will sound good. Since I already knew vertical dispersion would be poor, the "sweet spot" listening won't be an issue.

07-29-2007, 10:45 PM
Sounds like you are getting to the bottom of it. :T

07-30-2007, 12:27 AM
Jed, you are probably right, and I have to thank you for it. :) Quite frankly, if you hadn't guided me through it, I'd have probably forgotten about this cabinet.

Anyway, a 6th order filter seemed to help a lot. I'll probably be refining the crossover point a bit during the week, but I don't expect the topology to change that much.

I have very minor cancellation now, considering the size of the drivers, I guess this was pretty much unavoidable.

Anyway, I may try to drop a bit the x-over frequency on the tweeter, but nothing too serious (500 Hz or less, perhaps).


(FR as measured with the tweeters on the outside).

07-30-2007, 12:38 AM
:T Glad I could help out. I'm sure they'll sound as good as they look. I haven't tried those BG mids/tweets yet myself, so let us know your impressions.

07-31-2007, 10:05 AM
I forgot to check one thing... impedance.

It drops below 2 Ohm.

I'm designing a new cabinet with side-firing woofers. I simulated in LSPCad the required distances to avoid phase cancellation. And yes, the tweeter is now on top of the mid. :)

I'll build it in el-cheapo MDF, just in case I sc**w something up again.

(Hey, if I'm using relatively expensive / nice drivers, there's no use in paying good $ for a crossover when I'm not sure of the results! - plus, the new cabinets may look very cool).

I'll post again when I sort these issues out.

07-31-2007, 10:11 AM
I forgot to check one thing... impedance.

It drops below 2 Ohm.

I'll post again when I sort these issues out.

Ouch, oh well you tried to salvage those boxes. Live and learn as they say.

07-31-2007, 11:58 AM
They always say, you learn more from your mistakes... I may have been unsure about whether a side-by-side alignment would work before - now we all KNOW the answer far more specifically. :P

You can leave a little more volume in there for the mid this time too.

Having fun still I hope!


08-01-2007, 10:02 PM
C, it's always about having fun :)

I designed two new cabinets. This time, I'll have them built out of raw MDF, and choose between them. One of them is a standard 3 way tower, T-M-W. It's quite probable the listening axis will be on the center of the Neo8, but we'll see. I'm sure this one will work, since I already modeled it extensively.

The second one is a rather unusual design. It has the Neo8 and 3 mounted in the front baffle, and the 8" woofer firing sideways. According to LSPCad, at 200 mm from the front baffle, and with inverse polarity, the woofer sums up with the Neo8. Since LSPCad isn't taking into consideration the size and shape of the actual box, I'm not sure this box will work. I hope it does, since I think it'd look very cool.

Any thoughts about it? The boxes themselves were not that expensive, so I'll have them by next week.


08-21-2007, 11:14 AM
After finally deciding the baffles were all wrong (and the woofer enclosure could be a bit bigger), I built two new enclosures.

One of them had a side-firing woofer, the other is a normal TMW baffle.

Anyway, the one with the side-firing woofer was GREAT... on axis. Off axis (obviously enough now, not so obvious back then), FR depended on whether you were listening directly to the woofer or not. Since the x-over point was around 700 Hz, the speaker was unusable.

The TMW baffle showed some promise, so I decided to work on it... and after a couple of weeks, this is what I ended with. Other tha vertical off-axis response (which I already knew wouldn't be that great, due to the tallish drivers), I think the design turned out quite fine.

One of the things I love about it is that it uses no resistors whatsoever. Levels were adjusted using the inductors' resistance (and, quite honestly, they aren't that critical - but the values I have for them are the most exact ones).

I don't have any pictures of the baffles, since they are being finished right now, but here's the final circuit and lots of graphs.

Another thing - these drivers' response off axis is... ahem, well, not really flat. Things start looking much better at 15. I can't imagine designing a speaker using these drivers without considering off-axis response as well as on-axis.

(Note: Off-axis response is not an LSPCad simulation - it's taken from real 15 and 30 off-axis measurements. This is the only way to see how critical off-axis response seems to be with these drivers).



On axis:

15 off axis:

30 off axis:

Power response:

08-21-2007, 11:17 AM
Transfer function:


08-21-2007, 02:35 PM
I really appreciate how you get an idea and then try it out no matter how unusual it is. That's the way to get new cool inventions. As someone wise said, most breakthroughs in science didn't come after "Heureka, I did it" but "hmmh, that's interesting".

The sidefiring could work if:
1) It had lower xo point and/or
2) smaller woofers and/or
3) woofers firing to both sides

Here's some examples where sidefiring woofers are used:



09-18-2007, 12:00 AM
After what seemed like months of hard work (oh, wait - it really took that long), I finally finished my speakers

I'm absolutely blown away at the Neo drivers in this configuration. By combining the 8 and 3 , I avoided most of the issues (especially regarding FR) I had with the Neo3 alone.

The Neo8, crossed at 800Hz, is simply superb.

How good is it?

Here's an idea - I was listening to the "Brothers in Arms" XRCD2 disc today, and I was sure I had blown a speaker. It had a very strange distortion I had never heard. After swapping channels, speakers, amplifiers, versions of the album, and listening to both speakers side by side I realized something - the distortion I was listening to came from the bass guitar itself. It was so real and pure, I thought the Neo8 had blown up! :E

Now I have a question... can a dynamic driver sound as realistic as the planar ones? I've listened to many, but I had never confused real life distortion with recorded distortion. And I don't think I've ever heard a speaker with such dynamic transients, either (except for my old MartinLogan Scenarios, which were quite quick with transients, but sounded terrible).

I'm sold on these drivers. I'll try to post some pictures and more information tomorrow.

02-15-2008, 12:53 AM

After living with the speakers for a while, they became my reference units. The Neos are really something - I suppose their attack - decay times make them sound "different".

Bad thing is - I blew a CA21REX.

Oh. Well, I didn't "blow" it. Or maybe I did. I was listening to the Blue Man Group, and one of the woofers started making a rapping noise. I was pretty sure I blew it. But when I looked at it, it was fine (the cone, that is) - the tinsel leads came apart from the cone, and the voice coil wire broke down. No matter what I did, I couldn't fix it.

And, I realized Madisound doesn't carry the CA21REX anymore. I think Solen still does, but it's a different CA21REX - apparently, Madisound had an "Innersound" branded model with slightly different specs. Anyway...

Looking for a replacement woofer, I found the SEAS CA22RNX. Close enough, but the frame is a bit bigger, and my grilles are tightly built around it. So it wouldn't be very useful. And the T/S & efficiency parameters aren't even alike.

Metal cones (like the RS-225) were out of the question. I simulated the in-box response of it, and efficiency was way down, and the crossover became extremely complex from taming the upper peaks.

I settled on the SEAS P21RF/P. Same size as my old CA21, the T/S parameters matched the enclosure.. although efficency will go down at least 3 dB. It's a champ, though - in box speaker response is beautiful - I bet many people could get away with a 2nd order acoustic response with just an inductor.

Anyway, I just took some new measurements of the drivers in box. Much to my surprise, the Neo8 changed its performance quite a bit - it has more output in the lower midrange (around 800 Hz or so). The Neo3's response is as terrible as always (since those were used units, I'm not surprised).

I've modified the crossover so that it can use the 21RF/P unit. I need to take some off-axis measurements before posting it, though - I need to make sure the off-axis response takes care of the jagged FR of the Neo8 and Neo3.

Well, at least now the design is current again - since it's using available drivers.

BTW - now I know why SEAS is a bit more expensive than Dayton. I don't know about motor specifics, but SEAS and Peerless shure have beautiful FR curves... so smooth and easy to work with.

02-15-2008, 10:14 AM
The new Seas CA22RNY (http://www.seas.no/index.php?option=com_content&task=view&id=173&Itemid=165) could be a replacement for the old CA21REX. The chassis is of course the new 22cm type.

02-16-2008, 11:24 AM
Very interesting thread and project!!! Could you post a picture of the finished speaker?


02-16-2008, 02:21 PM
Minkuni, you are right, the CA22RNY has the efficiency and the FR needed! Vas is bigger, though - I think I'd have needed a bigger enclosure. The thing that hurt me the most was losing a couple of dB's - I liked the efficiency of the finished speaker.

Magnus, thank you! Here are some pics with the old CA21REX:




02-16-2008, 05:09 PM
Really, really nice!!! This was clearly the most interesting Neo-design I've seen.

Did you make the subenclosures for the Neos deeper in this version? What is the size of the whole speaker?


02-22-2008, 01:39 PM
I've been looking for the speakers' drawings everywhere...I believe I lost the file somewhere! I need to re-draw them and post them here. I'll probably do such a thing this week-end.

Anyway, the Neos' chambers are 1". The front baffle is 2", so I just did a cut-out on the front 1" baffle. It really does help the low end of the Neo3, extending it quite a bit. Then again, its FR is so jagged this isn't a big benefit. The Neo8 is far easier to work with, in my opinion.

I finished modeling the new crossover for the SEAS P21. Good news are many - the design is simpler, FR is flatter, and the parts are cheaper. Plus, the load is far easier on the amp. Bad news, obviously, are the loss of around 4~5 dB's in sensitivity.

I actually like the sound of the P21 a bit more than the paper version. Bass is a bit more solid, and since the crossover was reworked, the addition of resistors to the design meant I could achieve a midrange response a bit flatter, plus I contoured the high frequency so that off-axis response is far better. One big minus of the original design was a small sweet spot - the new one has a wider one.

It's good to have these speakers back. I didn't remember how much I love the Neo8 sound.

I'll post the new crossover sometime during the week-end.

04-03-2008, 06:21 PM
How are things going with the new crossover? Is it stable yet and will you post it and the measurements?


06-29-2008, 04:23 PM
I was very lazy posting this... sorry about that.

Here's the new crossover for the SEAS P21.

I've been living with this speaker for quite some time now, and I prefer it to any other designs. The Neos have a certain sonic signature to them (more accurately, they don't have any sonic signature, in my opinion) that makes me go back to them all the time. They are a joy to listen to - never tiresome, aggressive or misbehaved in any way. The only shortcoming I really dislike is the fact that they don't have a great off-axis response. I tried to minimise this problem in this design by aiming for a 30 flat frequency response, and it helps.. when you are at or near the sweet spot.

These have been my reference speakers for quite a while.

Impedance: Quite a bit more benign than the first design. The design is about 4 dB less sensitive than the original one, though.

Transfer functions: smooth transitions all around.

Frequency response at 30: The Neos measure badly at 0, but at 30 they behave nicely.

Final crossover design: Most of the parallel caps are shown like that because that's the way I built the crossover (I wanted to use a couple of stuff I had from previous projects).

06-29-2008, 06:11 PM
I've always wanted to listen to these BG's. Too bad you're a bit far from any of the US diy meets, it looks like a great speaker!

07-01-2008, 02:54 AM

Does this mean you have to listen to them 30 off-axis to have a flat response? Do you have a measurement at 0? :)

07-07-2008, 06:08 PM
Well? ;)

11-20-2008, 11:00 PM
Time to revive an old thread.

I still think these are the best speakers I've ever designed, and they are in my "reference" system. Time has only made me love more and more the Neos - especially because my living room has hard floors and many reflective surfaces, and apparently the Neos controlled vertical directivity makes the speaker sound much better than other designs with better vertical dispersion.

Anyway, I did create a new version of the crossover. Reason being, the speaker sounds extremely well when listened at a very specific sweet spot - ears at the Neo8 height. At that point, the midrange is extremely well rendered, and the highs blend well. Unfortunately, when you stand up, the speaker loses its coherence pretty fast.

The new version of the crossover drops the crossover point to the Neo3 down to 3.5 KHz, which helps a lot in this respect. The speaker now sounds a lot better around the room. The midrange still sounds great - I believe the trade-offs (the midrange may be a bit better with the first version) may be worth it if you are going to be listening to the speaker around the room, instead of sitting still.

I'll be posting the design tomorrow. Crossover topology is the same - only some values changed.

11-22-2008, 11:47 PM
Javier- glad to see you back with more info on the Miniliths. I'm a big fan of three ways and interested in building either your Miniliths or the Statement Monitors as my first DIY project. I will need five for my multichannel SACD / HT theatre set up.

Before I got the urge to build I was very interested in buying :E five of the Eminent Technology LFT-16s or (even) the LFT-8s http://www.eminent-tech.com/main.html. I think both are tremendous values for the money, but again I have the bug, although not one iota of woodworking capability.

Keep up the good work. On my part I've just finished reading Chapter 2 of Speaker Building 201.


11-23-2008, 01:23 PM
Hi Russ,

Here's the design.

The changes I made to the original are:

* Lowered x-over point from the tweeter to the midrange: 3.2 KHz, LR-4.
* Added a series filter to the tweeter. The Neo3 has a rising response in the top-end, and although you could avoid using this filter, I found using it makes the speaker sound a bit more natural. The exaggerated highs are very obvious when you remove this piece.

Here's the schematic. I used Dayton polypropylene caps all around, except for the shunt ones, which are basic non-polar electrolytics. The big inductor on the woofer is an Erse 14GA with an iron core, in order to keep sensitivity higher.


The impedance is fairly benign, with a single 3.5 Ohm point in the midrange.


The frequency response is very similar to the original design's one, but the power response is vastly improved. The tweeter's response is better controlled in this version, too.


I'd definitely call this the better of the two designs. After comparing it to my own commercial speakers (Monitor Audio RS8, Energy Connoisseur 7) and my own DIY designs, these are my favorite speakers.

11-24-2008, 09:38 AM

Thanks for posting this. I have a pair of Neo8's begging to be put into service and your writeup gives some good guidelines on what to try with them.


02-12-2009, 03:13 AM
Hi Javier,

Any uodates on the neo project, has any one else built these. I have pair of neo3 pdr's and 8 looking for a home. Are you still happy with your final design. Looking forward to your reply.

Best regards


02-12-2009, 09:34 AM
Hi Stuart,

They are still my favorite speakers. You may have to play around with the padding on the Neo8 / Neo3, depending on your listening distance (a bit more padding if you are listening to them farther away).

Their mids are outstanding...

02-12-2009, 03:01 PM
Hi Javier,

Thank you for the reply! I want to use them with a HiVi D8.8+ , this is an amazing driver and is very quick and dynamic, the bass section will be 28litres and will go flat to 33hz in that cab. Do you think the neo's will work ok with it. What is the volume of the box for the neo 8. Lastly is it ok if i use the Neo3pdr that has its own back cup?



02-12-2009, 03:29 PM
I don't think see why the D8.8 wouldn't work. The bass crossover would need to be redesigned, but the overall topology would remain the same, since I believe the HiVis also have a smooth response (not as smooth as the P21, but still nice).

I'm not really sure about the Neo8 volume. The enclosure, though, is 3/4" thick and the size of the Neo8. The same thing goes for the Neo3. I'd suppose not removing the Neo3 back cup wouldn't matter with a crossover at 3.2 KHz. I saw more extension without the cup, but you wouldn't really be using it at all.

02-12-2009, 03:59 PM
Hi Javier
Here are the specs of the D8.8+. Do you think you could help me with the woofer x over design, or the change in values of your x over.

D8.8+ Bass-Midrange

Shallow mineral filled polypropylene cone;
Fatigue resistant foam surround;
Inside vented magnet system with huge Ferrite magnet;
Long-throw symmetrically driving motor structure;
Extreme power handling voice coil with 100mm diameters black anodized aluminum former;
Progressive spider with decompressed rear volume.
D8.8 represents a design where the voice coil is positioned over an undercut central pole piece with a large thermally stabilized Neodymium magnet at the bottom and high energy ceramic magnet on the top. A specially shaped, CNC machined steel cup completes the magnetic structure from the outside, creating a shielded magnet system with symmetrical flux density distribution along the gap. The long-throw voice coil experiences a symmetric driving force and substantial decrease of inductance and back electromotive force modulations. The whole magnet structure is optimized using Finite Element Analysis. The result is a radical improvement of driver linearity. The cone has a shallow, high integrity structure, which is extremely rigid and dynamically stable.
The large voice coil diameter allows more uniform cone excitation, thus improving phase and transient characteristics of the driver, resulting in clear and accurate reproduction. The vented design, effective and uniform cooling of a high temperature rated voice coil provide a dramatic increase of power handling.
D8.8 is excellent as a woofer for high power "no compromise" audiophile systems. It also may be used as a subwoofer in various configurations.

Products Graph:

Mechanical Drawing


Frequency Response

Bottom View

Top View

Part View 1

Part View 2

More picture Woofer D series: Please choose... D5.8 Bass-Midrange D6.8 Bass-Midrange D10.8 Hi-end Woofer D8.8 Hi-end Woofer D8.8+ Bass-Midrange D5G Bass-Midrange D6G Bass-Midrange D8G Bass-Midrange D10G Bass-Midrange

D8.8+ Bass-Midrange
Nominal Impedance (Z)(Ω) : 8
Resonance Frequency (Fs)(Hz) : 31
Nominal Power Handling (Pnom)(W) : 150
Max Power Handling(Pmax)(W) : 300
Sensitivity (2.83v/1m)(dB) : 86
Weight (M)(Kg) : 4.2
VC Diameter (mm) : 100
DC (Re)(Ω) : 6.5
VC Length (H)(mm) : 23
VC Former : CCAW
VC Frame : Kapton
Magnet System : Inside
Magnet Former : Ferrite
Force Factor (BL)(N/A) : 13.3
Gap Height (He)(mm) : 8.0
Linear Excursion (Xmax)(mm) : 7.5
Suspension Compliance (Cms)(uM/N) : 425
Mechanical Q (Qms) : 6.16
Electrical Q (Qes) : 0.44
Total Q (Qts) : 0.40
Moving Mass (Mms)(g) : 62.7
Effective Piston Area (Sd)(m2) : 0.0241
Equivalent Air Volume (Vas)(L) : 34.9
Cabinet Type : Vented
Recommended Box Volume(Vb)(L) : 40
Tuning Frequency(Fb)(Hz) : 30
-3dB Cut-Off Frequency(F3)(Hz) : 28


2006 HiVi Inc.(USA) All rights reserved.

02-12-2009, 04:35 PM
I'd need to mount it on baffle and take a measurement. Or, have infinite baffle measurements of it. I don't think I've seen them on line... I'm not sure they exist. Do you have a link to a PDF file for it?

02-13-2009, 01:02 AM
Hi Javier,

Here is the link - www.swanspeaker.com/product/htm/view.asp?id=60



02-13-2009, 11:17 AM
Hi Javier,

Forgot to mention, when the language pack installation screen pops up just click x and the file will load up!



05-06-2009, 05:54 PM
Very nice design!! What is the software you are using to design the speaker?


05-07-2009, 11:21 AM

I did it all on LSPCad 6, JustMLS, the KHF Tool.... the Edge baffle step compensation and the XLS Spreadsheet by Paul Verdone for finding diffraction effects.

It's still my favorite speaker. After building a couple more designs (including Zaph's Dayton RS 3.5 way), I find the Neos have the best midrange and highs of them all. They don't call any attention to themselves - they are smooth, unfatiguing, clear, detailed... I'm currently using them with a McIntosh MC-240 tube amp, and they behave beautifully with them. I used them with a Channel Islands Class D amp, a Rotel RB-1070... even a small T-amp. No problems whatsoever. They are the best drivers I've yet used.

I'm currently building a Jed's Tombstone clone, with an Usher 8137A, TB W3-1335SBs and Seas 27TAFN tweeters... if they sound as good as these ones, I'll be extremely happy.

09-03-2009, 04:18 PM
I am looking at doing a similar project using 2x rs225s8 in parallel.

DO you suppose that this would provide a good sensitivity match?
With your xo, will the BG NEO8 reach 110db?

09-07-2009, 05:54 PM
2 RS225's would be, I think, a bit too sensitive for the Neo8, perhaps 1 dB or so (which could be "fixed" by using a lower gauge inductor in the woofer circuit).

I've actually thought about changing this to a floorstander two woofer system. I'm sure they'd sound nice.

09-07-2009, 08:04 PM
I am debating the possibility of building a very robust tower and was wondering specifically whether or not the neo8 would handle high volumes. I don't listen at rock concert volumes often but I would like to have the ability to take these speakers to smaller venues.

I am guessing that I may gain some sensitivity because I would be using the neo8 in dipole mode. I realize that I am in for a treat trying to equalize the low end. I also need to decide on a baffle size. I am seriously wondering whether or not this may be asking too much of a single neo8 per side.

Could you tell me whether or not you have tried these in a situation where they were really cranked?

04-11-2011, 11:22 AM
It'd seem I'm stuck in the past. In some ways, I am. I was fortunate enough to get back the first speakers I ever built and the Miniliths... for free!

The guy I sold them to got a Bose Lifestyle system, so he didn't need them anymore. Heh.

Anyway, I compared the Miniliths to:


Just to give you an idea, this is the FR, in room of the Miniliths (aqua trace) vs. the Megaliths (red trace).


So, apparently, the Megaliths' power response is far better, the bass hits far lower, and the midbass is less bloated.

Why on earth, then, do I still prefer the midrange and the high end of the Miniliths?

It just doesn't make sense. The Neo3 & Neo8's FR is a lot more ragged. I'm using 2 TB W4-1337's instead of a single Neo8, and two RS270's (!) vs. a single Seas 8".

And, know what? I could live perfectly happy with the Miniliths instead of my huge towers. All I'm missing is the deepest bass (which is far better on the big towers).

The Neo8 and Neo3's are champs. I wasn't wrong when I said I preferred this design to any other speaker I owned. 4 years later... I could say I still do. Wow.

04-11-2011, 04:33 PM
In my ears the Neo3 used right is one of the best sounding tweeters out there, especially considering the very reasonable price. :P

04-12-2011, 12:13 AM
Yeah, the Neo8 and Neo3 are... different. I hate to use this term, but definitely, they sound "fast". As in "transients are incredibly sharp", waaay faster than my cone drivers. And that is with the Neos in an enclosure, I can imagine in open air they are even... well, faster.

One thing my cone speakers beat the living dynamics off the Neos is, well, dynamics. The Neos can't go very loud without audible stress.

11-03-2011, 02:16 AM
Do you have a schematic of your final enclosure for this design?

Andres Silva
02-06-2012, 10:37 PM
Hi (hola) Javier,

I live in Mexico City too, If I write in English is because this is the languaje HTguide forum. Please tell me the amplifier you use with your minliths? The minilits works better with tubes or solid state amps?

Thanks and saludos,


5th element
02-06-2012, 11:04 PM
Well the neo planar drivers are some of the best measuring drivers within their pass bands too. I wouldn't say that the frequency response of the miniliths is 'ragged' we're still talking only +-2dB from 400hz and up. There's nothing wrong with the minilith design compared to the megalith except for maybe the bass driver running a tad hot.

If you prefer the midrange/treble from the miniliths then I'd just go with it, they have every right to compete with and perhaps exceed the performance of the megaliths. Each of them handle the midrange and treble in quite different ways, it isn't really a surprise that you have a preference for one.

02-07-2012, 07:10 AM
Javier, it's likely about the difference in distortion products. That, or some aspect of the power response that sounds more natural. As 5th notes, there are reasons the Neo3 and Neo8 are well regarded.

02-19-2012, 11:47 PM
Funny how these things work. This thread was digged up a couple of days ago, and I modified and re-painted the Miniliths, to better fit the room they are in now. And yeah, there's a circuit change involved, and yeah, to my ears, they are better than ever

All of the changes have to do with the midrange - woofer interaction. The P21 is an excellent driver, and its bandwidth is quite wide, so it needed something faster than a 2nd order electrical.

It's interesting that I've built at least 5 more designs after these ones, but I always come back to them. I just saw how many crossover iterations they've been through - 5. IT seems that everytime I learn something new, I mod them - but no other designs. I suppose that, after all these years, I wasn't wrong, and I really do love the Neo drivers. I'm listening to Gaucho right now, and Donald Fagen has never sounded better. :thumbup:

I think it's time to properly document these speakers. I know the P21 is no longer available, but the tweeter - midrange crossover can be reused with other drivers.

Tomorrow, hopefully!

02-23-2012, 10:47 PM
Sneak peek: I took off axis measurements from 0 to something like 80 in 6 steps. Here's the result. I like what I see.


02-24-2012, 12:26 PM
This makes me very excited to get to the project I'm collecting parts for that will use the Neo8 as a mid. I won't be using the Neo3 though. :)

02-24-2012, 04:30 PM
Hi C,

Actually my latest redesign of the x-over is based on what I've been able to learn about power response and off-axis response. The Neo8 works pretty good in this regard. I *think* (correct me if I'm wrong here) that the driver, being so slim and yet so tall helps smooth its horizontal off axis response a lot when compared to a standard cone midrange. That, and the fact that it simply sounds so great - if I were asked about this design, I'd say I love it because of the Neo8.

02-24-2012, 05:17 PM
Low distortion midrange drivers are crucial... the ear's pretty sensitive in that region.

5th element
02-24-2012, 06:13 PM
I can only echo what Jon has said. I was really happy with the way my W15CY001s sounded and then I swapped that out for the B&W FST midrange unit.

At normal listening levels the distortion looks like this.


Which is ridiculously low and turning it up beyond that, to what I consider loud, only really increases the second harmonic up to a flat 0.1%

Perhaps you should perform a sweep of the Neo8 + Neo3 at your normal listening levels and compare it vs the megaliths, it might show up something interesting.

03-04-2012, 11:06 PM
How about the Neo 10 in an upgraded design? Looks like BG is using it in a new commercial speaker they offer and it is flanked by the Neo 3's. Looks like a dipole.

03-04-2012, 11:10 PM
My bad. Their new Radia speaker is an MTM with two Neo 10's around a Neo 3, then filled in below by 6.5" bass unit.

05-20-2013, 01:06 PM
Any updates to this? I have a Neo3 and Neo8 .. looking for a woofer (hopefully something that is still in production)

09-09-2013, 12:29 PM
Do you have a schematic of your final enclosure for this design?

I'm also would appreciate this information. A list of key dimensions would also work if a drawing is not available.

Thanks and good work!

09-09-2013, 11:07 PM
Any updates to this? I have a Neo3 and Neo8 .. looking for a woofer (hopefully something that is still in production)

I have decided to go with the Hivi 8.8+ as the woofer. The Neo3 and 8 will be passively crossed and then an active crossover to the 8.8.

I already cut the woofer boxes .. hope to get the tweet/mid tops done this weekend.