First Design project - could use some (probably a lot) of advice

[technodanvan]

The below was posted at TechTalk earlier today, however there are a bunch of really super smart guys here and I'd love for your input too. I've got big ideas and tend to get carried away pretty easily, but today I think I finally settled down, recognized my limitations and lack of experience, and settled on trying my hand at designing a 'simple' TM bookshelf speaker as a first project to call my own. I would like this first design to be as successful as possible so I absolutely welcome any and all advice and recommendation.

-------------------------------------------------------------------------------------------------------------------------------------

Since I have completed a number of other people's projects I figure it's about time that I sit down and design my own for once. I was sort of hoping to keep it relatively simple and use well-known drivers with well-documented measurements, and maybe even use one of the PE boxes so I could focus on crossover design. However, this happens to coincide with my wife's request that I build a pair of speakers for her, with the following criteria:

1. They must be relatively small and unobtrusive (I have defined this as being no more than 0.25 ft^3)
2. They will be used for ambient music in a largish kitchen/dining area
3. They MUST integrate purpleheart somewhere into the cabinet
4. They must use the Tang Band W4-1879

The last requirement came from the "mistake" of letting her know what website a good deal of my parts come from. Basically, she liked the way it looked. Like a good husband, I said I could make it happen and ordered a pair for her before they became NLA. To be fair, I like the way it looks too....though I'm more concerned about the impressive magnet structure and not so much the cone/phase 'plug'.

Anyway, it is what it is at this point.

I happened to finish a set of boxes for some CSS EL70s about the same time these were delivered so I dropped them in to see what I had. Obviously the bass wasn't right since the box wasn't tuned for them, but overall they sounded surprisingly good. Much better than the EL70, in my opinion. Off-axis response was noticeably poor, as was overall treble reproduction when turned up a bit. If I were running them full-range with a dedicated seating position I could probably live with them in a proper box. However, since Requirement #2 basically says these need to be able to fill a room I'm thinking the addition of a tweeter would benefit them. Plus it's not much of a first design without a crossover, right?

I initially had ideas of all these extravagant things I could do with integrating a small amplifier, maybe make a tiny OB, etc, etc. But what it comes down to is that I don't really have the time or equipment for experimentation in non-standard alignments, and more importantly, I just don't have the knowledge or experience to try something like that yet. So I'll be keeping the box a simple, ported affair on the order of 0.25 ft^3.

-------------------------------------------------------------------------------------------------------------------------------------

That said, here's what I need in terms of advice/recommendations (for now):

1. First and foremost, I have no means of measuring the drivers. From my understanding TB is notorious for publishing specifications that aren't particularly close to reality. Either this means I can get them measured by someone in the Vegas area (I have no friends here ) or I purchase the necessary equipment. Personally I think I'll be sticking with this hobby for a while, so I think having the equipment on hand would be a solid decision. For a bare-basics amateur, what should I start with? Is the Omnimic V2 what I should be looking at, or is there something else I should start with?

2. I could use some recommendations for a nice tweeter to pair with the driver. I have a single Fountek NeoCD1.0 and a pair of Dayton RS28A tweeters sitting NIB, but I'd rather go with something I haven't used yet. <$50 each, unless a little more gets me someplace special. By no means does that mean I want to spend that much, if $25 gets the job done then that's cool too. I understand that realistically I should wait until I see the true limitations of the driver, so if that's the answer then that's the answer.

3. Perhaps most importantly, I need to know if I should be thinking about anything else. Or rather, have I already skipped important steps?

-------------------------------------------------------------------------------------------------------------------------------------

That's where I'm at for now. I've done a lot of research into the driver and haven't really seen anyone use it outside of a couple small full-range towers. I'm using the published specs to try to zone in on what box size to go with, but for the time being I feel pretty stuck without measurements. What else is there (for now)?

[JonMarsh]

Where will the box be placed in relation to adjacent boundaries? This makes a big difference in how you'll need to approach the crossover. Eg, were it flush mount, built into the wall, or a cabinet, then you wouldn't likely need any baffle step, simplifying the crossover design and reducing the inductor size.

Have you researched posted measurements? One thing I'd want to see is the impedance curve and distortion curves- underhung means it should linear up to a moderately high frequency as regards HD, but you know, in God we trust, all others (including TB) we verify...

Last, if you want some IB type measurements, or in a specific box, you could PM me about arranging that- 0.25 isn't very big to ship.

Funny, my GF's daughter lives in Vegas, and is coming back out here for the first time in 4 years...

Oh, and with a driver this small, plan on carefully bevel rabetting the back side of the front panel!

[technodanvan]

Where will the box be placed in relation to adjacent boundaries? This makes a big difference in how you'll need to approach the crossover. Eg, were it flush mount, built into the wall, or a cabinet, then you wouldn't likely need any baffle step, simplifying the crossover design and reducing the inductor size.

Have you researched posted measurements? One thing I'd want to see is the impedance curve and distortion curves- underhung means it should linear up to a moderately high frequency as regards HD, but you know, in God we trust, all others (including TB) we verify...

Last, if you want some IB type measurements, or in a specific box, you could PM me about arranging that- 0.25 isn't very big to ship.

Funny, my GF's daughter lives in Vegas, and is coming back out here for the first time in 4 years...

Oh, and with a driver this small, plan on carefully bevel rabetting the back side of the front panel!

Thanks for the reply Jon. In all likelihood these will be shoved up against a wall, or at least within 6 inches or so. They should end up sitting on an open shelf and not shoved into a cubby or anything like that.

I've done some research but can't seem to find any independent measurements of the driver. It seems there was enthusiasm to buy them when they were on closeout at PE, but after they were gone there wasn't much discussion about actual numbers. The most information could be found over at diyaudio, but most of that was just the full range guys going back and forth on whether it was money well spent and no objective information.

I really appreciate the offer for getting them measured, I'm sure you're much better equipped than I to do so! I really would like to have a go at doing it myself, even with inferior equipment/software/overall brainpower. I guess I just don't know where best to invest my money to that end. I won't forget the offer though, it'd be simple to just buy one of the 0.23 knock-down cabinets from PE so we could get a box for measuring put together.

Fortunately I've had some recent experience in cutting out the back of some baffles this size so it shouldn't be too much of a problem. I'm still working out the details on wrapping up two pairs of Speedsters.



I feel like maybe I'm spending too much time with 4" drivers. Between half a dozen W4-1720 woofers, a pair of EL70s, and a full dozen (maybe more!) W4-1337 mids...I need to move on to something else.

Anyway, if I were to actually make the baffle out of purpleheart then I'll need to buy a new cove bit, that stuff is hard! I'll probably try to find some veneer instead.

[fbov]

Considering the application, I would run them full range. Your usage pattern has no defined listening position, so there's no "axis" per se. You may still want to add a contour filter, but you won't need BSC...

... and that's not the kind of "speaker designing" you want to do, but it gets a high score for WAF.

My normal recommendation is to reverse engineer someone else's design before starting out with one of your own. If only as a paper exercise... helps you find out what you don't know.

For measurement hardware, Omnimic is a great product, but you only need to spend ~1/3 as much for a basic acoustic measurement system, starting with a mic:

http://www.parts-express.com/pe/showdetl.cfm?partnumber=390-801

and mixer or other source for "48V phantom power" (not phantom powered)

http://www.musiciansfriend.com/pro-audio/behringer-xenyx-802-mixer

or

http://www.amazon.com/Blue-Microphones-Icicle-Converter-Preamp/dp/B001EW5YQS/ref=sr_1_15?ie=UTF8&qid=1360705003&sr=8-15&keywords=Blue+Microphones

And freeware measurement software... I use REW

http://www.hometheatershack.com/roomeq/

For little tweeters, hard to beat PE's little neo's, ND16, ND20, ND28...

Hav efun,
Frank

[cjd]

If you don't go omnimic, get a proper dual channel USB soundcard with built in phantom power.
however, that won't include easy to use software. might have to eBay that.

HOLM looks nice. I jumped for ARTA and am glad I didI.

I have a mixer I used to use for the phantom power I'd let go cheap., holler

[technodanvan]

Considering the application, I would run them full range. Your usage pattern has no defined listening position, so there's no "axis" per se. You may still want to add a contour filter, but you won't need BSC...

... and that's not the kind of "speaker designing" you want to do, but it gets a high score for WAF.

My normal recommendation is to reverse engineer someone else's design before starting out with one of your own. If only as a paper exercise... helps you find out what you don't know.

For measurement hardware, Omnimic is a great product, but you only need to spend ~1/3 as much for a basic acoustic measurement system, starting with a mic:

http://www.parts-express.com/pe/showdetl.cfm?partnumber=390-801

and mixer or other source for "48V phantom power" (not phantom powered)

http://www.musiciansfriend.com/pro-audio/behringer-xenyx-802-mixer

or

http://www.amazon.com/Blue-Microphones-Icicle-Converter-Preamp/dp/B001EW5YQS/ref=sr_1_15?ie=UTF8&qid=1360705003&sr=8-15&keywords=Blue+Microphones

And freeware measurement software... I use REW

http://www.hometheatershack.com/roomeq/

For little tweeters, hard to beat PE's little neo's, ND16, ND20, ND28...

Hav efun,
Frank

Thanks for the input Frank!

I thought about just running them full range as well. They would certainly fulfill the purpose and I'm sure my wife would be happy as long as they made noise and looked nice. That said, I probably use the kitchen more than her and I'm not sure I'd be happy.

Also, from an aesthetics standpoint the box would be the same same size whether a tweeter is used or not. I'm always thought small single-driver bookshelves looked goofy. Maybe it's just me.

Yeah, it's probably just me.

Anyway, I've done a little "reverse engineering" to learn how Jeff Bagby's spreadsheets work. I'll need to do it again since it's been a while, however it seemed pretty straightforward and made logical sense after some toying around. I certainly don't pretend to be a professional, but I think I know just enough at this point to be able to see problems as they appear and ask questions as I go. Maybe. I hope.

I had forgotten Dayton just sold the mic, if you guys say it's better to just go with that and alternate free (or at least less expensive) software then I'll go with it. Instead of a large mixer, would something like the below work?

http://www.amazon.com/Art-ART-USB-Dual-Pre/dp/B002KEAT78/ref=sr_1_8?s=musical-instruments&ie=UTF8&qid=1360776715&sr=1-8&keywords=usb+mixer

I only have a laptop at the moment and would like to keep its peripherals as portable as possible.

If you don't go omnimic, get a proper dual channel USB soundcard with built in phantom power.
however, that won't include easy to use software. might have to eBay that.

HOLM looks nice. I jumped for ARTA and am glad I didI.

I have a mixer I used to use for the phantom power I'd let go cheap., holler

Thanks for the reply cjd!

Is there a substantial benefit to using ARTA over either HOLM or REW? I just ask because it seems that while the latter two are freeware, where ARTA has a license fee in order to save any results. Obviously this isn't a huge deal right now since I can just run with the free ones in the meantime, I'm just curious if ARTA has something else going for it.

If your mixer happens to be portable (like laptop sized portable) then I'd be interested!

--------------------------------------------------------------------------------------

I've been narrowing down my choices for a tweeter and have a short list right now (including Frank's excellent suggestion) of the following:

Dayton ND28 - $24
Vifa NE19 - $34-$42 depending on material of dome
SB Acoustics SB29 - $60
SB Acoustics SB26 - $35

I think I've seen most of these used in various projects and Zaph has data to get me started on most of them. I'm looking at getting one with a relatively small frame, though it's really not a necessity. The SB29 sticks out as by far the most expensive tweeter on the list but I've only heard very glowing things about it.

Thoughts on the above? Are there other obvious contenders that I should be looking at?

[BOBinGA]

+1 on the SB29. I just finished a build with it and love it. But I don't think it would be the best choice in your case. With a 4" woofer, you can cross at about 2500 hz and most tweeters will do that comforatbly. The reason I used the SB29 is because I crossed below 2000 and the SB29 has no problems with that low of a cross. With a 2500 hz cross, you could use a 3/4" and I would guess the Vifa NE19 would make an excellent match.

I use ARTA and find it pretty easy to use. If I didn't already have a mic, I would probably go with ARTA and the USB mic without the Omnimic software because I'm cheap. But certainly, the omnimic software would be an excellent choice.

-Bob

[technodanvan]

My own research on the SB29 turned up the same thing, it'd probably work but there are probably other models that would also suffice. That said, it may allow for more flexibility in the design. Given the frequency response of the SB29 down to 2000 Hz and below I could toy with different crossover points and different slopes, just to hear the difference. It might be overkill for the design itself, but the SB29 might give me some options to help learn how this is done.

I am tentatively crossing off the NE19 series, but only because they look awfully cheap with those little plastic "guards". If someone has used them and found otherwise please let me know, but I can't just wrap my head around those things.

Still looking at the SB26 and added the Seas 27TFFNC/G surface-mount tweeter as an option as well.

[---k---]

I also have a Dayton Mic and M Audio Moble Pre in the closet that has Phantom Power. And an older copy of Sound Easy that I won at Iowa. It is all gathering dust. I originally thought I would get into the design, but never did. I would be willing to part with it if you whatever doesn't work out with CJD. I don't know if drivers are still available or anything for the Mobile Pre. If you're interested, PM with an offer. I'm reasonable. I would probably need to test the soundcard to make sure it still works.

[BOBinGA]

I certainly don't want to dissuade you from using the SB29. I should mention that I used the large faceplate/ferrite version, but I'm pretty sure that the small faceplate/neo version is just as good. After looking at the NE19 and the regular dome SB26 again, I would agree that the SB26 (small faceplate) would be my choice. It should be able to reach down to 2k easily with a resonance filter to suppress the 950 hz impedance peak.

-Bob

[technodanvan]

I also have a Dayton Mic and M Audio Moble Pre in the closet that has Phantom Power. And an older copy of Sound Easy that I won at Iowa. It is all gathering dust. I originally thought I would get into the design, but never did. I would be willing to part with it if you whatever doesn't work out with CJD. I don't know if drivers are still available or anything for the Mobile Pre. If you're interested, PM with an offer. I'm reasonable. I would probably need to test the soundcard to make sure it still works.

I might get with you in a couple of days on that. CJD should be sending me a mixer shortly so I'll have power in a week or two, but I'll still need a mic for sure and one can never have too much software, right? The Mobile Pre sure looks nice though, big fan of M-Audio soundcards. I miss my old 24/96...

Anyway, I appreciate the offer and will hopefully get with you by next paycheck.

I certainly don't want to dissuade you from using the SB29. I should mention that I used the large faceplate/ferrite version, but I'm pretty sure that the small faceplate/neo version is just as good. After looking at the NE19 and the regular dome SB26 again, I would agree that the SB26 (small faceplate) would be my choice. It should be able to reach down to 2k easily with a resonance filter to suppress the 950 hz impedance peak.

-Bob

No worries at all, I'm weighing all of the options and appreciate all the input I can get. I'll hold off on the tweeter purchase until I have all the measurement stuff up and running, but I think the SB26 will be the one. In the meantime, I'll keep an eye on the classifieds to see if another option comes up.

I initially figured that a small flange tweeter would be preferable, but given that I'll probably end up crossing fairly low is it that important to get the driver spacing so close? The only reason I ask is the driver is asking for a biggish box, so I'll have plenty of real estate on the baffle to fill up. Also, I have been given a design style to shoot for:

http://www.stereophile.com/content/joseph-audio-pulsar-loudspeaker

It might look goofy with a small tweeter. Then again, it might not.

That speaker is a little bit larger but I think it can be done. I'm thinking if I can get the main body a nice semi-gloss white then purpleheart sidepanels might look pretty sharp. Then I could keep the entirety of the box construction BB or MDF, which would be easier for me to do.

[technodanvan]

Minor progress - a pair of SB26 small-flange tweeters have been ordered. Probably no other updates coming anytime soon since my time in the woodshop is limited, but at least I have the parts that matter now.

Next step is to source some bamboo; I need to stop showing my wife pictures of other people's stuff. I'm looking at the two of you Jon Marsh and Jon W!

[technodanvan]

You know, when I first pulled up this thread I figured it'd be around two years old. Time flies I guess! One of my goals for this year was to finish these speakers to the best we (me and you, dear reader!) can do. I seemingly don't have any of the original files for this, the rationale for including a tiny passive radiator, or even any sketches. What I do have are the speakers themselves (shown below) and recollections of the following:

1. The radiators were 'tuned' to a specific frequency, but I'd need to run a sweep to figure out what it is.
2. I seem to remember reading about overstuffed cabinets in use with passive radiators.
3. The cabinets were primarily made from scraps of bamboo and purpleheart, so I had some constraints on dimensions used.
4. These were originally intended for use on a desk up against a wall, however I no longer have an immediate use for these so designing them as such probably wouldn't make sense now.

So anyway, I've never finished a design of my own before and could use some pointers on where to begin and general hand-holding throughout the project. I have an Omnimic and other software packages, though I'm mostly unfamiliar with them.

[Carl V]

attractive....good luck with the next phase

[JonMarsh]

It took me a bit to realize the first post on this dates to 2013- In one sense I'm happy to see I'm not the only person that has projects that get set aside for quite a while to get revived when time permits. But so far I'm still short of the time permits part!

Nice looking little cabinets, and I'd say the combination of bamboo and purple heart works well!

Of course, you're initial PE link for the TB drivers is long, long dead. They do have a number of drivers with that construction that are sort of full range, But realistically, about the only way I could approach this (were it myself) is to treat them kind of like black boxes, do the basic characterization of the drivers in these enclosure IN THE KIND OF LOCATION you will use them, and go from their with the crossover development.

Now, I'm still in the middle of my move, and it's funny, because I decided I had to get an alternate measurement setup up and running, not as "premium" as what I usually employ, but something that will give me confidence in the results. A few of the different mic and measurement setups PE offers were on backorder, so I ended up going with something more like a blast from the past, a CLIO Pocket setup. I used to use CLIO ages ago, when it employed MLS measurement techniques and was a PC plug in board. Pocket CLIO is an outboard USB powered box with A/D and D/A, and configured for both doing impedance measurements and acoustic measurements. Though it's a tad pricey, it was shipping, and it has Mac software (another big plus for me, though the software feels like it was written for the PC and hard coded graphics ported straight over).

In case you're curious, here's a sample impedance curve I captured with it. There's a lot of features and functionality, but that means there's a bit of a learning curve, too.



But more to your situation, I'd say you've got a few unavoidable steps to go through at this point:

• Collect impedance data for the woofer+PR and the tweeter. The former will guide you in understanding what your LF behavior will look like.
• In a location similar to how you will use them, take acoustic data of both drivers; use a window on the tweeter that is adequate for something like 500Hz cut off (need to understand it's roll off) and a window good for 100Hz for the woofer.
• Optionally, you might also be curious to do nearfield for the woofer and PR, say at about 3", and this will give you a good idea of how well the PR alignment is working.
• Look at the measured impulse response conversion (if your software does that) and you can determine the time orgina offset of the drivers- that will give you insights into what crossover type to use.

This delay offset for the woofer should be included in your modeling software for the crossover evaluation. I like VituixCAD- very flexible and capable. Free, too, but I frequently donate to Kimmo when I finish another project with it, so I'm sort of using it in subscription mode, having originally been a paid user.

The reason I bring that up is that the degree of time offset can be a guide to what crossover type to use- and which to try modeling.

Most two or three ways I do use my funny 3rd order network design that REQUIRES a certain amount of time offset at the crossover frequency for it to sum flat- I call it a quasi 3rd order LR, as the drivers are -6dB and in phase at the crossover point. This was the case for the NatlieP, and for the Wavecor Ardent, as well as others.

If you have very little time offset, then a classic butter worth network is simple to do, and works well as long as things are in phase and in time alignment. For a low output system, 2nd order is likely just fine.

This is the 10 thousand foot view- there are obviously a lot of details, but those are dependent partly on the exact tools you have, and how you intend to use the speakers.

But best of luck, you've got some nice looking cabinets and drivers there!

[JonMarsh]

Z axis time origin evaluation simplified.

OK, was thinking about this last night, and here's another Tip/Hint- to take care of the time origin alignment evaluation.

What ever raw measurement you make will have time delays for each driver, which include the measurement microphone distance and the difference in Z-axis acoustic origin.

There's a well known method to figure that out if your measurement software isn't up to it directly. OK, well known to me and others skilled in the art, so to speak.


Here's how you can do it:

• You'll need to take three measurements- the woofer only, the tweeter only, and parallel the woofer and tweeter and measure the combination
• Import the FRD or similar format files from your measurement tool into the calculator Tool module of VituixCAD, and calculate a minimum phase version of the file, exported with a new name, for all three files.
• Create a basic simple simulation file with just the source and both drivers (woofer and tweeter) as outputs- no components. Just wire the inputs together. You're effectively connecting the drivers in parallel as for the 3rd test.
• Compare the frequency response of the summed output against the measurement you made of the summed output. Add in a small delay on the woofer and play with the value until the simulated frequency response for the summed output matches your measured output. Then you have the correct time offset and physical delay equivalent, which can be used to inform your crossover decision.

Kimmo has a nice new online user manual page for VitiuxCAD 2 now-

https://kimmosaunisto.net/Software/V...D_help_20.html

Here's the calculator tool main page:

https://kimmosaunisto.net/Software/V...alculator_tool


Here's a screen image from that page, how the tool looks for a typical waveform math problem.

[technodanvan]

Thanks for all the input Jon, I'll be referring to that this weekend (hopefully!) as I try to figure out some of the software I have. By the way, it's probably not evident in the photos but these are slanted back slightly and the tweeter recessed and woofer surface mounted. I'll be curious to see how much time alignment is necessary.

The current woofer link is here. It's dang pricy now, but I'm pretty sure they were around $60-$70 at the time of purchase. I think I read you may be looking at its smaller sibling as one of several candidates for your future line arrays? The magnet structure is pretty impressive but reduces the available volume I have substantially. Cramming everything in there will be a challenge.

[JonMarsh]

Yeah, the W3-1878 is very similar in design execution, but for some reason it's only a bit more than half the price! (around $87), and a few bucks cheaper on the commercial side. Yeah, I would call that dang pricey, too!

It makes the Visaton B 80, the priciest part I've been looking at for the line array (3.3"), look pretty dang reasonable! (about $50 each)





The other pricey candidate is the TB W3-1364SA.




You can see a certain similarity in that I'm favoring underhung Neo motor designs with small motors that don't mask the rear wave much.

That and others are all on hand, so we just have to see how things go when I have finished up enough of the moving tasks and earlier SMJ priorities to look at this in detail.

[technodanvan]

I don't know why, but that Visaton looks pretty inviting.

Ever so slight progress over lunch. I installed DATS and ran a couple impedance sweeps. I'll see about getting the Omnimic installed later.



Edit: Man that's difficult to see. I'll see if I can brighten that up a bit.



I'm gonna go out on a limb and assume the big peak is our little driver while the small peak is the PR at 57 Hz. Given the small size of the cabinet and driver, is that a reasonable target?

[JonMarsh]

Examining a ported or PR reflex system is a little more complicated than that. The lower peak is the out of phase resonant peak of the PR. The dip between them is where the PR is loading the woofer and reducing cone motion. That seems to be occurring between 70 and 80 Hz.

What PR are you using, and do you have any extra mass loading on it? (Many PR's come with weight sets, which you can bolt on to alter the tuning parameters- however, just bolting on more, though it may lower the tuning, may make the response less uniform and do funny things to the transient step response.). I like to analyze PR setups with either Unibox or VituixCAD. If you can provide the PR part number and enclosure internal volume, it shouldn't be hard to model the system and fit it to this measured impedance curve (you were mentioning heavily stuffed - this damps the PR response but also reduces the PR output.)

A little math might go a long ways towards clarifying what you've got on your hands here.

[technodanvan]

The PR appears to be a Peerless 830878. This model is really inexpensive and does not come with any weights or even a method of attaching washers. What I did was JB Weld a threaded rod and nut to the backside of each PR - and this is the only weight I ended up adding. It turns out that I was pretty much at my target just by doing this, so additional weight was unnecessary. I'll see if I have some spare rod around to estimate the weight.

I'll also pull the drivers from the boxes and measure driver parameters over lunch (assuming I get one) and pull up Unibox to play around. The box is about 1.7 L not including the magnet - I'll measure that when I pull the drivers.

Edit: 1.3 L including the magnet.

[technodanvan]

So I've been playing around in Unibox trying to fit what I'm seeing on the impedance graphs to a bass response and it was going all over the place. I decided to try another project that I've been sitting on to see if I'm just missing something within the software, so I figured I'd share it here for now.

Again, the below is NOT from these tiny little speakers but instead 4x DIYSoundgroup Anarchy 708 drivers in a vented 80L box. Parameters are averaged from DATS measurements of four drivers (they were NOT broken in though). I was comparing this to Jim and Curt's Statements as the box size is similar (100L) and I was expecting bass performance to be comparable (or better). It is, in that these can play louder, but it also isn't as I cannot seem to approach an F10 of 20Hz as published for the Statements - it just rolls off much faster. Granted, when I model the Statements I also don't get an F10 of 20Hz, but then I'm stuck with using published specs for that.

The smoothest graph has a tune of ~40Hz, below.



I can get the model to extend bass at ~4-6dB down with a tune between 24-30 Hz, with 30Hz shown below. I'm afraid this would sound boomy or uneven in a room though?



Note these were set with a nominal power rating of only 25 watts - anything greater and I run into excursion limitations. Anyway, I intend to build a test box at some point and initially set it up as a subwoofer to see what happens. It's been fun playing around with the software, now I suppose it's time to get back to the matter at hand....

[JonMarsh]

about the Anarchy 708...

Good morning Danny- It's a beautiful day in the neighborhood, and I've got a little time to comment about the Anarchy 708 and your Unibox plots. (I've been busier than a drop of water on a hot griddle, I never knew retirement could be so much work, but doing in the age of COVID increases the Olympic degree of difficulty... plus these little side projects with Steve ALWAYS take up more time and resources than you initially expect...)

Please go down to the bottom of the post and read the disclaimer, then proceed with the meat of the post! :W

And yeah, for the sake of clarity, I'm going to start with the basics for the benefit of those not familiar with the 708.

I haven't read the posts on the other forum or the rationale behind the 80L 4x design, so I'm coming to this cold. Let's just start off with what the 708 is and what it bring to the party- which also necessitates a bit of discussion about what it isn't.

I like the 708/704, and maybe even a bit more, it's baby brother the 558/554. It is a mid woofer, not a subwoofer. It is a mid woofer with one specific technology, plus a well designed cone, that creates it's "enhanced" performance envelope. That specific technology is an XBL2 motor.

Now, lets review the 708 characteristics. Heres' the factory spec basics plus the measured results from the original Voicemail magazine review.



Key points to take note of -

• Spec'd and measured Fs agree fairly closely, and there's no reason not to use the factory number, maybe round it up to 46Hz.
• 46Hz is not the Fs for a subwoofer, espcailly not for one intended to output 30Hz and below. It's a pretty typcial midwoofer number.
• OTOH, the Xmax, at 12mm, is not a typical midwoofer number. That's half an inch. Are we really going to play this puppy at half an inch P-P, or are their other points we need to consider?
• Maybe that's where the XBL2 thingie comes in, and we ought to take a closer look at that...

I think some folks focus on XBL2 as a method to get a bigger Xmax number, but IMO, I think that is missing one of the real benefits- after all, why make a moderate Fs mid woofer with XBL, since the driver is not tuned for 20-30 Hz, like the big woofers made with this technology?

Remember PuriFi's big push is primarily focused on linearity. How does XBL2 play in that regards?

Let's look at Kippel BL force plots for a couple of different drivers.

Here's a Kippel plot for a Satori midwoofer, same size range... using a good conventional motor.



This is not untypical for the degree of aymmetry, based on rest position, and the Bl force drop off as you move across the gap.


Let's look at the Anarchy 708.





This is rather a different beastie, isn't it? the dual gap structure results in peak Bl force being realized in the 5mm excursion area, and, OK, just to be kind, we won't mention that the actual Bl force is about double... It has centering issues also, as do many drivers, but you can make the case that you have a very linear Bl product over a +/- 5mm range, pretty healthy for a 6-7" mid woofer. By pretty linear, I mean within 5% of nominal. So, I'd say using it at up to 5-6mm Xmax is going to be in the "sweet zone", whereas that is not remotely the case for a conventional mid woofer, where the Bl force is dropping to 50%.

So, what about enclosure tunings? The two you showed from Unibox would be described (for that total volume) as a maximally flat (like a Butterworth filter) alignment, and an EBS alignment (Extended Bass System), when you are tuning the Helmholtz system (whether ported 4th order or PR 6th order) below resonance.

There are always options to play around with, including the enclosure volume, but in the end you can't fool Father Physics. But sometimes you can come close...

Here's an alignment for the 708 that I like, using a Dayton RSS265PR with only a single disk (the minimum cone mass). (that should sound familiar, to those who follow Evil Twin's Ardent D work).




Those of you familiar with Unibox may be scratching your head a bit at the graph formatting, but those adjustments to text size and position were automatic when importing the Excel worksheet into Apple Numbers.

With typical boundary reinforcement in the area below 100Hz, this looks to be a pretty useful alignment, apart from the idea of putting four 10" PR's somewhere on an 80L enclosure. But then, having a port large enough to keep the air velocity down in that size enclosure is also a problem, and so would be the midrange resonances for that large a diameter port.

Just sayin'...

Those are important considerations for a woofer. For a subwoofer, the port issues don't loom as large, but for 80L there are probably better options for a sub than 4 708's.

Like two CSS SDX10's with PR's. XBL2, too. VERY low impedance rise with frequency, so minimal inductivity modulation and low IM. Not at all your typical subwoofer driver.

[technodanvan]

I should have been more clear - this is not intended for subwoofer duty, rather I may use it as such (initially) to simply test out its bass capabilities in-room. 80L was chosen after modeling several other box sizes as this seems to provide the smoothest response. It also happens to approach the box size necessary for the towers I have in mind.

As always, thank you for your input Jon. Always something to be learned! I had considered modeling this with a PR (or multiple PRs) but for cost purposes felt that money could be more useful elsewhere in the system (especially if it takes four per tower!). The downside of using a port is how massive it needs to be to eliminate noise - something approximating 24 inches long x 6 inch diameter. It'll fit though, and the volume it takes up may actually be valuable. You might remember I had considered pairing this with a Seas Excel C18EN002/A (aka King Coax).

[Bear]

Depending upon what you want to do for bass alignment, you could look at a single RSS265-PR (~425g) in about 20L with a pair of 708s in parallel. It doesn't give the depth that Jon's model, above, does. However, the step response does give a decent impersonation of a sealed Butterworth alignment with a bit of a boost at the low end and a touch of extended ripple after the initial effect is mitigated.

The Wavecor PR223 also seems to fit the 708, but you're paying 3x for the PR as you are paying for the driver itself! (note that a single PR223 models well with the big Purifi woofer, at a fraction of the cost of the Purifi PRs)

[ergo]

Anarchy 708 seems to provide a very good value for money. Such a pity I can't find it in European stores... Ordering from US would probably make it expensive enough to loose that good value proposition.

[JonMarsh]

The Wavecor PR223 also seems to fit the 708, but you're paying 3x for the PR as you are paying for the driver itself! (note that a single PR223 models well with the big Purifi woofer, at a fraction of the cost of the Purifi PRs)

This is a point I have noted, too, and intend to follow up on- though all of the PR223’s I have on hand have a planned home already… but a quick investigation seems like a worthwhile idea, given the number of the PuriFi woofers we have…

[Bear]

This is a point I have noted, too, and intend to follow up on- though all of the PR223’s I have on hand have a planned home already… but a quick investigation seems like a worthwhile idea, given the number of the PuriFi woofers we have…

My doodles have a 14L box with a 362g (+200g) PR223 as providing an interesting combination up to about 100 watts. The PTT6 is beyond Vance's measured Xmax at that point, but still around the factory spec. A 20L box with 262g models better, but the power handling is PR-limited to about 50W (depending upon how important going below 27Hz is). All of those scenarios are modeled using 0.3R of series resistance.

[technodanvan]

Well after playing around some more I think I'm just going to leave the PR alone. I can't seem to model it correctly within Unibox or WinISD, at least based on the measured impedance, which means I probably have incorrect parameters for the PR. There was no getting around a need for a subwoofer for this system anyway and as far as I can tell the PR doesn't appear to be hurting anything (i.e. causing weird/notable peaks in response).

If I get ambitious I might try to make a cover for the PR hole and measure the response of the sealed cabinet, but for now I'm going to move on and rig up something overly fancy for some off-axis measurements of the existing cabinet. Lazy Susan bearing is on order.

[Bear]

Well after playing around some more I think I'm just going to leave the PR alone. I can't seem to model it correctly within Unibox or WinISD, at least based on the measured impedance, which means I probably have incorrect parameters for the PR. There was no getting around a need for a subwoofer for this system anyway and as far as I can tell the PR doesn't appear to be hurting anything (i.e. causing weird/notable peaks in response).

If I get ambitious I might try to make a cover for the PR hole and measure the response of the sealed cabinet, but for now I'm going to move on and rig up something overly fancy for some off-axis measurements of the existing cabinet. Lazy Susan bearing is on order.

Let me know if you want some pointers on Unibox. The main issue happens when the manufacturer doesn't disclose the data that Unibox needs (e.g., SEAS). Otherwise, it tends to stay fairly close to the manufacturer published specs using the up and down arrows.

[technodanvan]

I appreciate the offer, but I'm not sure what else can be done at this point. I'd need the PR to mass virtually nothing to get the impedance peak to match with what I'm seeing in DATS.