[JoshK]

8th order? Sounds like a pretty extravagant x-over. Have you checked out the phase plots? Not that I know what they would look like-just curious.

Well if we are discussing Jon Marsh's Cauer-Elliptic filter, it is a 6th order electrical, 8th order acoustic filter.  It tracks an LR8 through the xover region to around -60db, then it is 2nd order thereafter.  It has about the same phase deviation as an LR4 though.  It does require big caps in the tweeter circuit though (like 80uf...and we don't use 'lytics!).   That isn't typical for a commercial speaker as those components cost too much.

Mark (forget last initial) copied JM's topology with a two way using the RS28a & RS225 but I think he put the Fc at around 1khz.   The RS225 has a pretty nasty breakup and 1850hz is far too high IMO, 1350 is probably still too high.

[Anthony]

Where are you all seeing the breakup at 1850Hz?  I'm not denying this is a high crossover point for an 8" driver.  I'm just going by the specs:

http://www.partsexpress.com/pdf/295-366g.pdf

Looks like the breakup doesn't start until 3kHz and peaks at 7kHz.  By then the crossover should have the signal down 60dB.

Unless I'm missing something . . .

[Vapor Audio]

You need more than that to determine what's a breakup and what's just a peak in the FR ... truth is the peak around 1000hz in this graph, I saw it at about 1200hz in my own measurements, is a breakup and it corresponds with a serious rise in distortion as well.

If you want to use 8" drivers, you'd be MUCH better off with a Peerless 830884.

[BoomieMCT]

Would using a Neo8 instead of a Neo3 allow for a low enough crossover?  Can you pop the backplate off the Neo8 like you can on a Neo3?

[Anthony]

The 8" was chosen to help extend the bass response as much as possible.  The Neo3PDR was chosen because it has been shown to work really well as an OB tweeter.

Right now this is all a big experiment.  I'm also considering a Neo8PDR for the tweeter and lowering the crossover considerably.  I am also considering moving to 6" woofers and sacrificing a lot of bass.

The goal here was to get something that plays well down to 80 Hz, where I can pass it off to a separate subwoofer.

Just curious, is there any way to tell with home measurements what is a distorted breakup and what is just an output peak?  Or does that require some of the more fancy speaker modeling/testing software packages?  Right now I just get FR data through Room EQ wizard (with a 20 to 20kHz whoop sweep) and a Behringer ECM8k mic.

Thanks for the insight.

[Anthony]

From the BG whitepaper:

It is necessary to
stress that using the NEO-8 as a dipole (without
any rear enclosure) may require some signal
equalization at lower frequencies, since a dipole
exhibits a natural roll-off. In some applications
where flat on-axis response is desirable, the use
of a correction network at 12 kHz is
recommended;

so it appears you can remove the back plate.  They just warn about the rolloff and the peak you may need to tame.  They even provide a recommended notch filter circuit (in a different whitepaper)

The paper can be found here:
http://www.bgcorp.com/uploads/product_install/Radia_Pro_NEO8_Data_Sheet.pdf

[markC]

mcgsxr:
Very generous of you to offer up your drivers, but what would you listen to while I had them? I'd need them for at least 2-3 weeks to evaluate and fine tune, (if the design works). I wouldn't want to deprive you of music for that long! Oh, and I guess if I did end up borrowing them I'd Have To Visit You to pick them up.

[Vapor Audio]

It doesn't matter if you go with 8's or 6.5's ... you're not getting down to 80hz open baffle.  With some room gain, 120hz 'might' be possible with a wide baffle.  Also if you're running the Neo3 without the back cup, it's not going to like going down to 1850hz either. 

The combination of high distortion from the RS225 run that high, and the open back Neo3 running that low would likely be very painful and unpleasant.  I'm not trying to rain on your parade, but I'd rethink your design before it's too late.  Sure it would be much more complex to design the crossover network, but why not do a WTMW or TMWW 3-way open baffle.  If you want cheap, these have actually measured pretty well from a few on the Parts Express board that measured them

http://www.partsexpress.com/pe/showdetl.cfm?&Partnumber=299-776

Although 8" or 10" woofers would be much less visually intrusive.  Use the GR M-130x for a midrange - plenty of excursion for open baffle and low distortion thanks to the XBL2 motor, and then the Neo3 open back for a tweet.  You could run the woofers up to about 250hz, midrange to about 2500hz, and tweet up from there. 

If you have a Behringer mic, I assume you have something to provide it phantom power?  If so, the only other thing you need is the free demo version of ARTA.  Download it and play around, even the free version is quite powerful and extremely easy to use.

[BoomieMCT]

It doesn't matter if you go with 8's or 6.5's ... you're not getting down to 80hz open baffle.  With some room gain, 120hz 'might' be possible with a wide baffle.  Also if you're running the Neo3 without the back cup, it's not going to like going down to 1850hz either. 

How do you figure?  I'm getting an in-room F3 of about 55Hz off of that same Part's Express 12" woofer you linked to on a 18" open baffle with wings with no EQ or BSC.  I don't think that is all the woofer - any driver with a low enough Fs should get similar performance on a big enough baffle.  The two Dayton 8"ers have almost the same surface area, excursion and an Fs that is only slightly higher. 

All this being said, if an OB doesn't work Anthony can always have the two 8"ers bipole mounted.

[Vapor Audio]

It doesn't matter if you go with 8's or 6.5's ... you're not getting down to 80hz open baffle.  With some room gain, 120hz 'might' be possible with a wide baffle.  Also if you're running the Neo3 without the back cup, it's not going to like going down to 1850hz either. 

How do you figure?  I'm getting an in-room F3 of about 55Hz off of that same Part's Express 12" woofer you linked to on a 18" open baffle with wings with no EQ or BSC.  I don't think that is all the woofer - any driver with a low enough Fs should get similar performance on a big enough baffle.  The two Dayton 8"ers have almost the same surface area, excursion and an Fs that is only slightly higher. 

All this being said, if an OB doesn't work Anthony can always have the two 8"ers bipole mounted.

Well, that's a 12" driver with a wide baffle and room gain ...

See for yourself, try The Edge http://www.tolvan.com/edge/

And if you have to resort to a very wide baffle to get the bass response you need, then you're introducing a whole new set of problems.

[BoomieMCT]

Well, that's a 12" driver with a wide baffle and room gain ...

See for yourself, try The Edge http://www.tolvan.com/edge/

And if you have to resort to a very wide baffle to get the bass response you need, then you're introducing a whole new set of problems.

My point is that (EQing aside) the low frequency roll-off because of the open baffle has more to do with the size of the baffle than the size of the driver in it (assuming the driver can play low in the first place).  Many people have made the Adire DDRs and been very happy with the response.  That's two 6.5" drivers, both open baffle, with a Neo3 open back crossed at 1800Hz.  They have some bass shaping (all passive) and go below 100Hz anechoically.  The Dayton 8"er has a slightly lower Fs and slightly higher Qts then the Adire driver in the DDR so it should handle the bass slightly better.

Distortion from running the 8" references too high is an issue  , the bass extension shouldn't be. 

Am I understanding something wrong here?

[mcgsxr]

MarkC, you are welcome to them for at least that length of time - at present, with the summer weather, I rarely get down to the basement - my kids, my listening buddies, want to go outside, so music is more of a fall/winter pursuit for me.

Besides, I am in the thick of a car audio (of all things!) system install that is stealing all my time.

I can unbolt the b200's and bring them to the office - Etobicoke is closer to you, no?

[markC]

Yes, Etobicoke is closer, but I wouldn't be able to make it there during office hours. I'll e-mail you about a possible hook-up. That is Me Comming To Visit You.

[Anthony]

Okay, the speaker is built and is currently in testing.  The baffle was cut from a plywood square about 2' on a diagonal.  Since my primary goal is HT, I'm going to test the center design first -- figuring if I can get that right, some traditional MTMs, vertical, shouldn't be a problem.

<===>  is an approximate shape of the baffle.  Here it is with the drivers <OoO>    It's 16" wide (tall here), coming closest to the woofers around 2" from their periphery.  I will post a pic as soon as I can get one uploaded.

It's still breaking in, but here are the results:

Speaker is 18" from front wall, microphone is 5' on axis.  Room is irregularly shaped, drywall walls with big fluffy couches and a thick rug with padding over a linoleum/concrete floor.  Drivers are NOT broken in yet (about 1 hour of listening on them right now).

Frequency response is all over the place, but the general trend is over 72dB at 80 Hz and staying above that until 20kHz.  There are two dips to that point I'm worried about:  1.5 kHz and 2 kHz (both above and below the crossover).  At the crossover point, I'm looking at 75 dB.

It looks like both the woofer and tweeter are falling off early, but with an odd mix of local peaks summed together to make for good output at the crossover point.  Weird.  I'm hoping that goes away with break-in, but a rework of the crossover is probably in order.

Finally, the notch filter I put on the tweeter section seems to start too early in the response and is not wide band enough.  I definitely need to play around with that a bit more.

Slight bass step down at 350 Hz, which is probably due to baffle step.  A more well though out baffle design should smooth that and get me a little extra bass.  Worst case, I'll notch down the woofer section a bit above 400 Hz or so.  I'm trying to keep the crossover as simple as possible.

Again, as soon as I can get the image on my hosting site, I'll link to it here.


Finally, how does it sound?  Actually quite good.  The bass seemed too tight, probably a break-in thing.  The highs, though were very natural sounding.  I was amazed how good the BG Neo3PDR sounds.  Almost identical to my Magnepan 10.1's.  In fact, tomorrow night, I am going to test the Maggie at the same location to see how it holds up.

So this is all preliminary.  After about 40 hours of break-in, I'll know more.  I'll also play around with the notch filter and crossover to see if I can't tweak them a bit.

Thanks again for the help and support!
A

[BoomieMCT]

Two things,

1.  Try switching polarity on the tweeter and seeing what happens.
2.  Get a baseline of each driver with no electrical network.