Need speaker ideas/help/advice/critique

[Saurav]

Hi,

I've been lurking here for a while, and know some of the members on different forums. This looks like a pretty friendly environment for DIY newbies, so I figured I'd sign up. I hope to learn a lot from the time I spend here.

Anyway, on to my questions. My system currently falls kinda into the hi-eff/SET paradigm. Here's a not-too-recent shot of the speakers:



Changes since the picture was taken: The upper baffle is upside down now and reversed, so the tweeter sits below and to the outside of the midrange. Also, there are a couple of layers of polyster batting on the top surface of the cabinet.

Details: The woofer is my old Adire HE12.1. So it's an Eminence Beta 12CX in a ~ 24x14x15 sealed cabinet. The midrange is an Audax PR170M0, and the tweeter is a Fostex FT17H. Crossovers - active LR4 at 80Hz to the sub (NHT1259 in a sealed cabinet), active LR4 at 500Hz between the woofer and the midrange. The XO between the midrange and the tweeter is passive - 12uF/8ohm zobel and a .20mH coil on the midrange, 2uF cap on the tweeter. Amps: A Gainclone (~ 25W or so) on the woofer, and a 3.5W SET on the mid+tweeter. The mid/tweeter efficiency should be around 95-96dB.

This has all been set up by ear (i.e. handwaving and guesswork) and running sims using Speaker Workshop on manufacturer-supplied FR/impedance plots. I've run some minimal measurements on the final speakers using an SPL meter, but nothing worth mentioning. I have the parts to build a microphone and preamp, and will get around to doing that some day. Yes, I know I should do that before asking for advice

Anyway, questions. The thing I've been dissatisfied with is the treble. I really like the sound of the midrange driver, but I've heard smoother treble in many speakers. Now I have a pair of Fountek JP-3.0 ribbon tweeters that I plan to try here. These should match the efficiency of the mid fairly well. I plan to start out with the same XO. I'll probably also start out by mounting the ribbons on the same baffle (assuming they fit), probably directly above the existing tweeters (remember that the baffle is flipped).

Does that sound reasonable? It will place the ribbon at ear height, which is important based on everything I've read. It will also place the ribbon almost beside the midrange driver - I'm not sure if that will be a problem or not. Should I put together a new baffle for this? If I do, what kind of mounting would be recommended as a starting point? Maybe put the mid above the ribbon, but lined up with it vertically? Should they be in the middle of the baffle or offset to a side?

Assuming I get my measurement jig set up, any advice on measuring the ribbon tweeter? I know I shouldn't send a low frequency to it, and should always keep a cap in series. I can run FR measurements on it that way, I guess, but wouldn't that affect impedance measurements? Also, about measurements in general - I don't have the luxury of mounting these on a stand in the middle of a field. Can I get useful results using Speaker Workshop with a microphone in my living room?

OK, that's all the questions in my mind at this point. I'll probably think of more as I go along. And once again, hello to everyone here.

Saurav

[Jerry Pease]

what you doing around here???? It's a small world after all.....

[GrahamT]

Welcome Saurav,

I can't answer most of your questions, the pooh-bahs will be around to help soon I'm sure. One thing I can say is that you might be able to use a program called Baffle Diffraction Simulator to design the placement of drivers on the baffle.
There are a few people here who really know their dipoles so you should be in good hands.
Now start building that mic, jig and preamp.

Graham

[Saurav]

Is that the evolved version of The Edge? If it is, I've played around with it a bit. I asked the designer about my specific case, where the open baffle sits on a box, and he didn't really know how I should model it. Do I include the surface area of the lower enclosure or not? Or is the effective baffle size just the top part? I have no idea. Which gets back to measuring it

Thanks for the welcome, guys. And yes, I noticed the recent dipole threads, so I figured this is a place where I could get some good advice.

[ThomasW]

Hi

The standard use of a PR17 is as a transition driver (upper midwoofer) going to a large dome midrange. Sorry to say that there's going to be beaming and breakup modes at any reasonable frequency where you can safely XO to a small ribbon.

IMO you need to rethink this speaker. Drivers should be used where their cone operation is most linear (pistonic). Forcing drivers to operate where it's convenient for the user and ignoring the physics involved is asking for trouble.

My guess is that your lack of satisfaction with the treble has as much or more to do with the choice of the passband for the PR 17 as it does with any tweeter.

Some 20yrs ago I worked for a company that designed numerous systems using the PR17, so this isn't idle speculation on my part.

[Hank]

Suarav, is that YOU? From HT Talk? Heed Thomas' advice. The biggest learning I've had on this forum is that tweeter crossover frequency is critical, mainly because of cone breakup modes in just about, unfortunately, all midwoofs. Those modes mostly occur in a frequency area too high to cross over to most tweaters, leaving you with just a few tweaters that go low enough to cross over below cone modes of a few midwoofs. Pistonic operationg limitation is closely related.
On reflection, it makes me wish some companies had been putting as much R&D dollars into new tweeter motors as they have been investing into mid-woofs. Wouldn't a great-sounding tweet that could be crossed at 1200 Hz AND was affordable, sell like mosquito repellant at a nudist camp?

[JonMarsh]

On reflection, it makes me wish some companies had been putting as much R&D dollars into new tweeter motors as they have been investing into mid-woofs. Wouldn't a great-sounding tweet that could be crossed at 1200 Hz AND was affordable, sell like mosquito repellant at a nudist camp?

Uh, that's called a Seas 27TDFC, though you still need a kind of steep slope. And it won't do 93 dB/watt, like a Focal Tc120dx2. But what do you expect for $28.50 - the world?

Hank and Thomas are right; what's often laid at the feet of the tweeter is actually a problem with the midrange or midwoofer. Easy way to get a handle on that is listen to the system WITHOUT the tweeter, and see if the upper range sounds smoothly rolled off, like a nice equalizer tone control, or if there's some grit and grunge coming from the midrange driver. The cleanest top end I've seen on a 6-1/2" is the Peerless HDS, and it starts getting into energy storage problems at 2 K and above. That's OK, if you have a fairly steep crossover around 1.6 - 1.8 kHz. But that's low for a ribbon to go- if you look at the ribbon tweeter nonlinear distortion plots, waterfall plots, and ETC curves, the Arum Cantus units, for example, get hashy below 3-4 kHz, and start to show a lot of spectral contamination (IM distortion).

Well, you know what they say, one man's tweeter distortion is another man's "air".


Seriously, check out this site for some projects using box woofers and dipole mids. Some good info and research, though I personally think you need to take the dipole thing down to at least 60-75 Hz. First dipole speaker we built was only dipole from 60 - 300 Hz, but boy, what that did for the clarity in the bottom end through upper bass.




DIY projects

[Saurav]

Lots of good info here. Let's see...

From the plots, I know what the Audax looks like above say 1.5kHz. So far it sounds good to me though. I have listened to it without the tweeter, that's pretty much how I decided on the zobel/coil combination. I've seen lots of people tell me that I should cross that over lower. I've also seen people use it up even higher than I'm taking it. But I agree, the ideal thing to do would be to XO the Audax lower. But if I want to keep this a 3-way, that means a different tweeter, right. And I just bought the Fountek Which sounds pretty good, BTW, definitely cleaner than the Fostex.

What driver would you recommend to go between the PR170M0 and the Fountek JP-3.0? Assuming I was insane enough to try a 4-way (well, 5-way if I include the sub). I would need mid/high 90s efficiency, preferably a high enough Q to be usable on an open baffle, and affordable, say under $150/200 for the pair.

Dipole bass... yup, heard a lot about that too I know this is backwards, monopole bass with dipole midrange. My objective was to reuse my HE12.1 cabinets/drivers, which is why I did it this way. Some day (in a new house and new listening room) I'll build a dipole woofer.

Also, assuming I decided to keep my driver complement for now, what would you recommend for placement? I could try a 'real' crossover too, and move it down to say 3kHz or so. That would probably need a 3rd or 4th order on the ribbon, and maybe a 2nd order on the midrange.

Thanks for the link, Jon. I'll check it out, and probably come back with more questions. That's not a site I've seen before.

[JonMarsh]

Well, you're possibly going to need a different top baffle, perhaps taller. The tweeter should be at a height close to ear level when you're sitting down- I'm guessing you want to optimize for your normal seated listing position. The vertical dispersion of the Fountek ribbon will be about the same as a comparable size dome tweeter would be- so, imagine a 2-1/2" or 3" dome tweeter- not very big vertical window. You could try crossing it over directly to the Audax- nothing less than a 4th order L-R at say, 2.5 kHz, probably. With as flat an impedance curve as they have in the low end, you could probably use something close to a textbook filter on the Fountek. The Audax you'd need to have decent measurements in that baffle. Using the FRC tools wouldn't work because you can readily translate the dipole baffle into a modification of an IEC test baffle, regarding it's affect on the response.

A more robust tweeter to use in this application would be a Focal TDC120TD5 or TC120DX2, as they're about 93.5 dB/watt sensitive, and wouldn't have the vertical dispersion issue. Wouldn't be hard to do a decent 1.8 kHz crossover for one of these- but you already bought the Founteks.

Do you have measurement gear? A a good basic setup would be a Behringer ECM800 mic (very inexpensive) with an MAudio balanced mic preamp with phantom power (less than $100), and SpeakerWorkshop software (free download on the web).




Speaker Workshop





~Jon

[ThomasW]

The Fountek and the Fostex are really supertweeters. Any driver that can't go lower than 5K isn't really a tweeter.

The proper way to decide what goes between the PR17 and any supertweeter is to measure the PR17 and see where the 1st cone breakup mode occurs. The XO point should be below mode. Once the XO point is known then you can chose the an driver.

The other option is to chose a tweeter with a broader passband and forgo the supertweeters. The downside to that is tweeters that go down low usually aren't high efficiency, or if they are really efficient, they're very expensive. (think compression drivers or very expensive ribbons like the Raven 3 or Stage Accompany)

[cjd]

How 'bout the BG Neo 8? Or two. Might not even need that Audax... If the Eminence goes up that high.

It's in my finalist list for a dipole setup with probably some of the 10" Dayton RS drivers on the bottom (and some Ascendant drivers below that). Not sure what to use to fill in the very top since the Neo 3 seems to be not nearly as clean as the 8.

C

[Saurav]

The Ravens are definitely outside my budget. This is the first speaker I'm trying, I'm definitely not at a level where my design/construction abilities would justify that kind of expense.

I thought of something else last night - is there any driver that can replace the Audax? I looked through the Fostex full-range catalog a while back, but didn't find anything that had a high enough efficiency, high enough Q, and was reasonably priced. If I could find a different midrange driver that could go up higher, that might solve my problem too, right?

Well, you're possibly going to need a different top baffle, perhaps taller. The tweeter should be at a height close to ear level when you're sitting down- I'm guessing you want to optimize for your normal seated listing position.

Understood. That would put the tweeter at the bottom of the top baffle. So the midrange driver would go above it... vertically aligned with the tweeter, I guess? I looked through the website link you provided, and there's a lot of stuff there that's similar in concept to what I'm trying to do. Thanks a lot.

Do you have measurement gear? A a good basic setup would be a Behringer ECM800 mic (very inexpensive) with an MAudio balanced mic preamp with phantom power (less than $100), and SpeakerWorkshop software (free download on the web).

Working on it. I have Speaker Workshop, I've been playing with it for crossover simulation purposes. I've bought Rod Elliott's preamp PCB, I still need to buy the parts to build the preamp and buy/build a mic. I was thinking of the Panasonic units, since everyone seems to use those. How much is the Behringer?

The proper way to decide what goes between the PR17 and any supertweeter is to measure the PR17 and see where the 1st cone breakup mode occurs. The XO point should be below mode. Once the XO point is known then you can chose the an driver.

Looking at the graphs published by Audax:



The first cone breakup looks like it occurs a little below 2000Hz, based on the blip in the FR and the spike in the waterfall, right? One octave below that is 1000Hz. Audax recommends a lower crossover point of about 500Hz (and I wouldn't want to go much lower on an open baffle anyway). So that doesn't leave much of a range that this driver is used over. The other school of thought I've seen is to ignore the breakup modes in the FR, because it sounds good even with them. Which is a school of thought with dubious merit. It does sound good to me, but I know enough to know that I shouldn't be trusting my ears.

A zobel brings the 3-5kHz SPL down to about the same level as the 500-1000Hz range. So maybe I could pick an XO point around 2.5-3kHz, that's one option. I'm doing this more to have fun and learn rather than build the perfect speaker, so I know I'll be going through a few iterations of drivers/crossovers before I finally decide that it's done.

[taz13]

Part Number 248-625 from parts express
BEHRINGER ECM8000 MEASUREMENT MICROPHONE The ECM8000’s linear frequency response and omni-directional polar pattern allows you to carry out measurement and alignment tasks with minimum hassle and maximum precision, making it a perfect complement to any real-time analyzer. Its rugged construction ensures years of trouble-free use.

$49.99 EA

[ThomasW]

Look at the glitch in the impedance curve about 1.5-1.6kHz. That's the first cone breakup mode. The XO point should be below that mode for optimal performance.

You're sort of backed into a corner here. There are reasonably priced tweeters that can be used that low. None however will meet your needs/wants for high efficiency.

So chose your compromise, bigger amp and low buck tweeter, or high priced, high efficiency tweeter.

As noted the ECM-8000 is certainly be best deal in mics out there. It does need phantom power (15-48VDC). The little Behringer 802 mixer is a good low buck device that can be used as a mic preamp and has phantom power.

Through Jan 30, 2005 Behringer has a $15 rebate on the ECM-8000

http://www.behringer.com/01_news/rebates_us.cfm?&lang=ENG

Since watts are cheap we don't pay much attention to what's available in the world of HE drivers. There are midwoofers we know that can be XOed higher than 1.5kHz but they aren't HE. One example Jon already mentioned is the 6.5" Peerless HDS

[Saurav]

Look at the glitch in the impedance curve about 1.5-1.6kHz. That's the first cone breakup mode. The XO point should be below that mode for optimal performance.

Which is worse, and means that this isn't really a useful driver, because that gives me a passband that's less than 2 octaves.

You've sort of backed into a corner here.

So I see I'd hate to lose the amp, I've put a fair amount of money into building that and I know I won't be able to sell it for anywhere close to what I've spent.

Is there any other midrange driver that I could use instead of the Audax that might fit better with the tweeter? Or I could go fully active, use the SET for the midrange driver and something else for the tweeter. Maybe one of those Sonic Impact digital amps - have any of you heard those? But that would mean losing the ribbon tweeter.

Should have planned this a little better...

[Saurav]

P.S. Is this the Behringer mixer you were referring to?

http://www.partsexpress.com/pe/pshowdetl.cfm?&DID=7&Partnumber=248-602

[ThomasW]

You need to decide what world you want to live in. The 'HE' world, or the 'regular' world where 100+watt amps are the norm. Unfortunately the fundamental elements of these worlds are in essence incompatable. So one must pick one or the other.

$39 Class D amps are toys. Audio gear particularily power amps, should be an investment, because good ones have the potential to last literally for decades. Now that doesn't mean one needs to spend a ton of cash. There are plenty of reasonably priced kits, or one can buy on the used market

Yes that little 802 mixer is fine for a low buck mic pre-amp, with phantom power.

[Saurav]

You need to decide what world you want to live in.

LOL!!! Very true. That's good advice, and I thank you for it.

[Saurav]

Well, since I had nothing better to do, I decided to see what I could do for measurements. This is with the RS SPL meter (with a correction file from the ETF website) going into the mic input of my laptop's soundcard. Software was Speaker Workshop, and I did my best to follow instructions for setting the correct gate times.

This was more an exercise is figuring out how to use SW for measurements than anything else. I'm not sure if I should trust the RS meter above about 5kHz, especially because some of the dips in response showed up no matter what driver (or comination of drivers) I was measuring. The Fountek is sitting on top of the lower cabinet, beside the top baffle (which has been moved to the side). The "crossover" is put together with whatever parts I have on hand. I'm sure the response will change significantly when I get a real mic/preamp, and probably also when the Fountek goes onto a baffle (right?). My measurements for the Fountek don't match the ones posted on the website that Jon provided the link to, which is one more reason I don't trust the treble response of the RS meter.

The dip at 1300Hz is definitely the baffle, and some other people who tried a PR170M0 on similar sized baffles had similar dips. So the next baffle I build, I'll have to take care of that. Other than that, I don't think I can draw too many conclusions from this. At least the driver sensitivities seem to be decently matched. It sounds better than any of the previous iterations I've tried. The treble still feels a little splashy, which many be because I've been listening to these with a hole in the treble for a long time, or I need a higher order filter on the ribbon, or something else entirely.

[JonMarsh]

Yes, you do need that Fountek in a baffle. And be careful of what crossover you run with it- no low order stuff at low crossover frequencies, you don't want to feed it any significant level below 2 kHz.

That site I referred you to has a nice evaluation article on open baffle midrange testing- should give you some good ideas as to panel layout. That guy and I independently came to the same conclusions regarding asymetric layout for dipole midrange.

Sounds like you're having fun...

[cjd]

What are your XO slopes? As a curiousity, did you try reversing the polarity on the Audax ever?

Your measurements look significantly better than anything I've managed so far, and I have the Behringer gear. I must be missing a step in the SW setup procedure (actually, IIRC, I kinda missed the steps to calculate the sound card/cable impedance/capacitance/etc.) Everything rolls off pretty fast on the top end, and my ears insist otherwise. I did discover a set of speakers I used in HS look to have a polarity problem - huuuge null. Pretty funny.

Oh yeah... the BG Neo 8. Or two.

C

[Saurav]

Yes, you do need that Fountek in a baffle.

It's raining here, and all I have to work on is my porch Hopefully it'll be dry this weekend, and I can get my jigsaw out.

And be careful of what crossover you run with it- no low order stuff at low crossover frequencies, you don't want to feed it any significant level below 2 kHz.

I have 2.6uF of capacitance on it, which given its ~ 7 ohm impedance gives me a first order filter at about 8500Hz. And my amp is 3.5W. Do you think that's OK for a rough mock-up (which is all this is at this point?). Once I have a real mic I'll experiment with pushing that lower and going with a higher order, right now it seems pointless because I can't really measure the effects.

That site I referred you to has a nice evaluation article on open baffle midrange testing- should give you some good ideas as to panel layout.

Yes, I saw that, I'll go back and read it again. Tapered wings seems to be the way to go. I do have the Audax placed asymmetrically on the baffle, but it looks like even that's not good enough by itself. One edge of the driver is about 4.5" from the baffle edge, which is roughly 1/2 the wavelength for about 1400Hz, so I think that's probably what I'm seeing.

Sounds like you're having fun...

Always

What are your XO slopes?

2.6uF on the tweeter, 0.22mH on the midrange. No zobel at this point.

As a curiousity, did you try reversing the polarity on the Audax ever?

I tried reversing polarity on the Fountek. Heck, I had it reversed until I read the website which said the red terminal was ground. When I had asymmetrical slopes on the two drivers, reversed seemed to work better. But all of this will probably change once I actually put the tweeter on the baffle.

Your measurements look significantly better than anything I've managed so far, and I have the Behringer gear.

Have you seen Claudio Negro's website? He has an excellent tutorial on SW. I discovered last night that my new laptop only has a mic in, and no line in. So I was getting just noise for a while, until I realized that I could tell SW that there was no reference channel. Then it started working. Up until then I guess it was expecting to see the reference signal on the other input, didn't find it, and so was confused.

I must be missing a step in the SW setup procedure (actually, IIRC, I kinda missed the steps to calculate the sound card/cable impedance/capacitance/etc.)

I didn't do that either. Make sure the signal levels aren't overloading SW, again Claudio's website tells you how.

Oh yeah... the BG Neo 8. Or two.

I looked at that again yesterday. Could I hook up a pair in series? That would get me the sensitivity I need, as well as 8 ohms impedance. Not sure what it would do to any Q-related issues though. That's $200 in drivers though, which means I'll have to save up for a while.

[cjd]

The Neo's can certainly be wired in series. I forgot that tube amps are extra picky that way. That Fountek sounds good though, doesn't it?! And, what's a couple hundred bucks more on the ol' plastic? Besides, I bet you could sell your Audax drivers for a pretty good price bringing the total to a reasonable sum. Not sure what the big shots 'round here think of the Neo's though. 8)

As for SW, I went through Claudio's stuff. I think it was just user error. Kitchen remodels don't wait for speaker tinkering, you know?! Almost done. Gotta clean up silicone caulk (it doesn't go on as easy as it seems it should) and touch up some paint where I bumped the wall installing my in-wall cabinet (just built). Then put everything back, wash all the dishes (drywall dust gets everywhere).

C

[Saurav]

Not sure what the big shots 'round here think of the Neo's though.

I'm interested in that too. And the Neos will add another problem into the mix - vertical placement. Now I'll have 3 drivers with limited vertical dispersion, and my listening height is pretty much at the bottom of the top baffle, or just a little higher.

Maybe I should buy a taller couch

[ThomasW]

We tested and briefly listened to the Neo8s when they first came on the market. Then we returned them for credit ......... :wink:

That being said quite a few people like them and our testing was to determine how a line array of them would compare to a RD75. For our needs the Neo8s weren't as good ...... YMMV..... :B

[Saurav]

My mileage is also at a very different budget. A pair of RD75s would cost as much as most of my system

Will the Neo8 work on an open baffle? I haven't found T/S parameters or an impedance plot anywhere... not sure if T/S even applies here. I'm assuming these are open backed, and have a back wave?

The more I think about it, the vertical placement will probably be my biggest issue with these. I could do an MTM with these and the Fountek, maybe put the Fountek to the side and the BGs one on top of the other. But that still means that my listening height needs to be about 6-8" higher than it is. If I want to move these drivers lower, I'll have to build a crazy looking baffle that hangs off the side of my woofer boxes. Which could be cool, but I don't think I have the skills (or the tools) to build something like that. I dunno.

[ThomasW]

Sure they work in a dipole. HERE's a link to their specs.

2/side with the Fountek in the middle might work well.

With PE if there are no 'use' or wear marks they can be returned. We used duct tape to attach them to the baffle, and alligator clips to attach the power leads. That way we were able to run our tests and return the drivers 'as new'

[Saurav]

I'd found that page. There's no mention of any kind of Q though. Maybe that doesn't apply to this kind of driver? It was my understanding that one of the main requirements for a driver to work in an open baffle is a relatively high Qts.

Duct tape is a great idea. I've been using alligator clips on my speakers for a long time.

Thanks for all the help.

[thylantyr]

I've done alot of listening tests with the PR170M0, various PHL drivers, Seas Excel, and Focal kevlar and W cones and while
the Focal 6W midrange has a good looking chart, it always
scored dead last in percieved sound quality listening tests.

The consistent top scoring midranges were the PR17 and PHL 1120
midrange even though their charts are not the best.

The moral of the story is that you should never forget about
the human variable when doing loudspeaker design. It's
hard to judge final sound quality by just looking at charts.
At some point you need to listen to the drivers.

To fix the sonic problems I would do this;

1. Optimize the enclosure for proper listening height if you haven't
done so already.

2. Find a tweeter that operates well as low as 2 - 2.5khz using
a 24db/octave crossover.

3. The low pass on the PR17 is not your sonic gremlin, therefore
the crossover frequency/slope is not critical.

4. Temporarily, wire up all your drivers in a tri-amp setup using
a flexible electronic crossover. Set generic crossover frequencies
as already indicated. Listen to music and tweak the settings
to whatever crossover frequency/gain your ear desires. Note the settings
and then if you want, build a passive crossover using the preferred
numbers.

5. Based on your testimony, the Fostex FT17H doesn't appear
to be a great sounding tweeter. I would ditch it and use something
that is smooth down to 2khz.

6. Last (personal preference), I would ditch the low powered
amplifiers and use higher powered amplifiers to raise your clipping
headroom, avoid distortion.

[Brian Bunge]

Saurav,

I have a bunch of Neo 8's. If you decide to try them I'll make you a good deal on a couple.

Also, can you post a link to Claudio's site? I've got a mic and a preamp but haven't had much luck getting things set up with my laptop and a few programs. Maybe SW is the answer for me w/o having to spend any more money on this test setup at the moment.

[Saurav]

Brian,

Speaker Workshop tutorial:

http://planeta.terra.com.br/educacao/claudionegro/english/

There's also a Word manual out, but I haven't really looked at it:

http://www.audiodiycentral.com/resource/SWmanual/Speaker%20Workshop%20Manual%201.0.zip

I have a bunch of Neo 8's. If you decide to try them I'll make you a good deal on a couple.

Cool. I'll get in touch with you if I decide to try that. The first step is getting the Fountek on the baffle and seeing how that works with the Audax.

thylantr,

For now, the amps are definitely staying And so is the lower enclosure, I don't look forward to building that again. I'm going to be trying out wings on the upper baffle (why didn't I think of duct tape before?) as well as cutting a hole to put the tweeter on there. If I had a spare amp lying around and a flexible 3-way crossover, I'd definitely be prototyping this with an active XO. But building those just for testing purposes seems like a lot of money to spend. Some day I might go fully active, but for now the mid-tweet XO will be passive.

[Saurav]

Got the tweeter on the baffle, and fiddled with the L/C values some more. This is better than what I had before, I think. Everything's still first order slopes. I think playing with baffle wings and/or applying dipole EQ should take care of the 750Hz hump and 1200Hz dip. Then a 4th order filter at around 2.5kHz should bring the 3-5kHz peaks of the Audax down a bit, and I'll add a 4th order ilter on the tweeter as well. If nothing else, this should tide me over while I research the Neo8s.

[JonMarsh]

Looks like you're making progress. I bet it'll be smoother once you tame that area between 2 and 5 kHz.... :B

How does the power response look? What happens when you go off axis horizontally to the right or left?

~Jon

[Saurav]

Haven't checked that yet. I was kind of saving the real testing for when I get a real mic, if that makes sense. Is it OK to move the microphone? The instructions I've seen say to rotate the speaker, but if I re-calculated the gating times for the new mic position, would that work?

Based on Audax's graphs again, I think I need the 2.5kHz LR-4 crossover to get the off axis response to work. By 5kHz the off axis response is down quite a bit. It's also not very much in phase at all, reversing the tweeter just produces a broad dip. Hopefully the next XO will make that better too.

[JonMarsh]

You'll want to locate the tweeter and mid on a common vertical axis to get consistent off axis behavior. Yes, you may as well wait on getting the ECM8000 before going to crazy with more testing and tweaking.

It's more reliable to rotate the speaker, not the mic; that the the room effects you may be experiencing remain constant- one less variable.

[Saurav]

You'll want to locate the tweeter and mid on a common vertical axis to get consistent off axis behavior.

Hm... good point. Right now they're offset horizontally a little, because I put the new tweeter where the old one used to be.

How important is it to flush mount these? I'd have to buy a router to be able to do that. I also saw a website (was it yours or Thomas'?) which showed how to build felt diffraction rings/pads. Would something like that help?

Also, what other baffle layout would you recommend I should try? My listening height puts me at the top half of the tweeter/bottom edge of the midrange. And this is the right speaker, i.e. the tweeter is to the outside.

[cjd]

Not flush-mounting can have surprising results. I would definitely do so. A tool-less person can do a flush mount by making a 1/8" hardboard (masonite) overlay for the baffle board. If it's not quite thick enough, you can use poster-board or similar to shim it out a tad. The tweeter is critical, though if the mid sits out too far it can cause issues with the tweeter still. Thomas's site does have good info on using the felt as well. Check a place that installs floating floor for felt scraps (you'll have to peel the backer off) and build it up. Or buy some of the stuff - it's not that expensive. Scraps are usually free though, so you're happier to experiment.

And, that layout would give the best response if you rotate the whole project 90 degrees. Also, IIRC this is mostly a one person hobby (the sitting and listening part - I *know* the alligator clip mess certainly is)? In which case, off-axis response may be less of an issue. Choose your poison.

For now, you could try laying the woofer unit on its side to give you more height to play with. Tweeter at ear level is usually the recommended layout. Have you done any experimenting in the Baffle Diffraction Sim yet? (FRD Consortium site). Moving a driver just a cm can have a surprising result. I find often 1/2" offset from center is the best spot for a tweeter, though I'm not sure why that would be at the moment.

C

[Saurav]

Also, IIRC this is mostly a one person hobby

Definitely

And, that layout would give the best response if you rotate the whole project 90 degrees.

In which direction?

A tool-less person can do a flush mount by making a 1/8" hardboard (masonite) overlay for the baffle board. If it's not quite thick enough, you can use poster-board or similar to shim it out a tad.

Hadn't thought about that. I could try something like that. Or I could try the felt thing, that might be thick enough to fill in that gap too.

For now, you could try laying the woofer unit on its side to give you more height to play with.

I did think about that. The bottoms of those cabinets aren't veneered, that's my only reason for not doing that I'll probably go that route though if I decide to try the 2xNeo8 or something like that.

Have you done any experimenting in the Baffle Diffraction Sim yet?

Back when I first built these, yes.

[cjd]

Since I have some time on my hands...

It seems that:

Total baffle height of 42", 14" wide
Mid centered, 15" down from the top
Tweeter offset .75", 9" down from the top

Gives you the best diffraction results I could get with a little inspection. That deals with some *really* significant peakyness in the tweeter you get otherwise, though there's still a ~1db peak at ~6300hz it's otherwise pretty much flat from 2khz-10khz (or +/- 1/2dB). The mid has comfortably flat response with a slight peak at 1650 and then standard baffle step roll-off, with a small shelf at 1000hz and peak loss at 605hz.

Just a quick glance and some tweaking. This says nothing of what wings or a top with non-parallel sides would do.

C

[Saurav]

OK, I was thinking of another diffraction software tool. Yes, I've played with this one too, but a long time ago.

I should be able to do this. The box is 14" wide, so that's taken care of. It's 24" tall, so 42" means I need an 18" upper baffle. 9" from the top puts the tweeter at 33" from the ground, which would put it a little bit above my listening height, but probably not by too much.

My current upper baffle is 8" tall. And I was wrong when I said my listening height is at the bottom of the midrange, it's actually closer to the middle of the midrange, so about 30-31".

When you mentioned rotating my baffle, what were you referring to? If I rotate it clockwise, the tweeter is on the lower inside corner. Rotate it the other way, and the tweeter is on the upper outside corner. Which one did you suggest? Either of these is esy for me to implement, since I can swap between left and right and turn the baffles upside down, so any rotation can be achieved easily.

[cjd]

Your pic had the tweeter between the mid and the woofer. This is similar to many designs of the 70's, which are best with the tweeter on the top. Don't worry about it - it's not as optimized as the 14x18 mid/tweet baffle. FWIW, I would move the mid as close to the tweet as you can get it - if that means altering the numbers I had for the mid, that's fine. I didn't find exact measurements on either, but for a quick pass it was close enough. Moving the mid a little up or down doesn't make huge differences.

And, 30" listening height?! 38-40" is the norm I thought! Do you sit on the floor and listen usually?

C

[Saurav]

What can I say, I'm 5'5"

OK, I got my wife to measure my listening height (yes, she is a very understanding woman), and it's about 35". I sit on a fairly low couch/futon.

My upper baffle is 12", not 8", I was doing my math based on that. So your 18" baffle should definitely be doable, and put the tweeter at listening height.

FWIW, I would move the mid as close to the tweet as you can get it - if that means altering the numbers I had for the mid, that's fine. I didn't find exact measurements on either, but for a quick pass it was close enough. Moving the mid a little up or down doesn't make huge differences.

So keep the tweeter position the same, and maybe move the mid closer.

[cjd]

(yes, she is a very understanding woman)

No kidding. I think we had that figured out already.

So keep the tweeter position the same, and maybe move the mid closer.

You got it. Center to center spacing rules, y'know? And do what you can to get both drivers flush with the final surface. If that means building up layers of hardboard or something (the Audax looks like it has a thicker rim?) then do that.

C

[Saurav]

OK, this is a lot better, I think. Added wings (cardboard and duct tape for now) and turned the baffle upside down to bring the tweeter closer to ear height. Changed the crossover too. Schematic and response are attached. I'm going to ignore the spike above 10kHz until I get the ECM8000. I know the JP-3.0 has a rising response, but not like that. I added the coil in series and played with some parallel resistor values mentioned on that website, but I'll leave this fine tuning for later. I really doubt the tweeter's response is actually that high. And if it is... well, I'll deal with it when I get there.

Edit: The third plot compares two different calibration files I've found online for the RS SPL meter. The differences above 10kHz seem large enough that any meaaurements there are suspect, I think.

The ECM8000 is on sale for the next few days, right. Time to spend some money I don't have