"Fundamentals" for a new stand mounted speaker project -what goes around comes around

[JonMarsh]

I tried to order a pair of C18 too to kick of my own build.... but local reseller got this reply from Seas
" About the date for possible production date we do not know at this stage, but I believe it will probably be closer to 6 months than 1-2 months."

That might be the reason why they have 'hidden' this model from web for now too. So I guess I need to find a source in EU who would still have stock.

That's unfortunate; hopefully, as many driver vendors as there are in Europe, you'll find one of them with stock on hand. In North America it's a slim selection of vendors; Solen in Canada has two in stock, Madisound still does (they're pushing the kit using the 26LROY woofer, too); Meniscus and Madisound both have the C16 coax, but Meniscus doesn't have the C18.

[DS-21]

Yep, is the BC using the stock ASX2 version?

Outwardly it looks the same as the ones in my miniDSP and Sonance amps, but I'm no expert. Still, they don't look hacked up. Also, I wouldn't think Bel Canto's orders are big enough for factory customization.

[Face]

I was under the impression that Wyred 4 Sound and Bel Canto both customized their Ice Modules.

[AJINFLA]

Outwardly it looks the same as the ones in my miniDSP and Sonance amps, but I'm no expert. Still, they don't look hacked up. Also, I wouldn't think Bel Canto's orders are big enough for factory customization.

If/when you build something with it, try tracks you are familiar with that require good burst power. One of my standard test tracks is Flim & the BBs "Tricycle". I've already pre-biased you, but after triple checking that no limiter functions are active in the DSP and your gain structure is good, etc, etc., give a listen.

cheers,

AJ

[JonMarsh]

If/when you build something with it, try tracks you are familiar with that require good burst power. One of my standard test tracks is Flim & the BBs "Tricycle". I've already pre-biased you, but after triple checking that no limiter functions are active in the DSP and your gain structure is good, etc, etc., give a listen.

cheers,

AJ

Man, I've hardly ever run into any one familiar with that album! Of course, it could be argued I don't get out enough these days... but I did back then...

In a possibly similar vein for transient demo's, ALL of the album versions of Stanley Clarke's "School Days" suck compared with the original vinyl, BUT, the import version of Stanley Clarke's "Bass-ic" finally has a digital version that is pretty dang close... will definitely give a work out to the system dynamics.

As an uninvited opinion, I've personally found the ICEPOWER modules to be a bit soft in definition and transient power in the bottom end; they're not bad compared with a lot of amplifiers, but they aren't the last word in that department. Doesn't mean a whole lot of other factors are just as important or more so, like the speaker system, the room setup and acoustics, etc.

[JonMarsh]

Finally! Some days the UPS guy gets here at a pretty good time, others it's rather late- today was the latter...

But I have a nice box of parts for building the first test crossover...

[DS-21]

If/when you build something with it, try tracks you are familiar with that require good burst power. One of my standard test tracks is Flim & the BBs "Tricycle". I've already pre-biased you, but after triple checking that no limiter functions are active in the DSP and your gain structure is good, etc, etc., give a listen.

I'll have to check out that track, but I'm not building anything with the PWR-ICE amps. I've been using them under my desk, in the metal enclosures miniDSP showed at CES, for a little while. One drives the mains and the other the sub.



One cool thing about them is that you can cascade them using the AES/EBU digital bus, and I use a miniDSP DDRC-22D box as the preamp. So there's no AD-DA loop (inaudible, I know, but it helps the OCD) in the system. I have no complaints with them at all.

As for concentrics, I shudder to think what one of Jon's crossovers costs. I priced out the S-3EX unit, and even using my style of parts (poly caps but not boutique brands, air-core inductors because they're what's specified on the schematic, commodity 10W resistors) they came to over $100 each. True, mine would be a little cheaper because I could simplify the MF crossover a little (active between bass and mid) but even after adjusting the circuit for my baffle. But I've never actually built a passive that expensive before. Alas, full active isn't an option, because I only had one run of 14/4 for each main channel installed in the walls...

[JonMarsh]

Well, one channel full passive, plus inexpensive shipping, was a skosh over $500... three ways are expensive, there's never any getting round that... and there's only two caps in it that I would call expensive... and the 500W Erse SuperQ inductors in the woofer section are much less expensive than AWG 12 air core inductors!

[JonMarsh]

Phenolic sheets arrived late today- I'd be laying out the crossover arrangement this evening, but have a VERY early conference call with Europe tomorrow morning- heading to bed shortly. Still, it's a nice feeling that things are starting to come together.

[5th element]

And coming together so quickly! Got to love those ready assembled boxes.

[JonMarsh]

They're the best for prototype acoustical testing! And smaller projects, too! But though I was considering a 1cu ft MTM enclosure initially for this, as the concept evolved, I realized that just wouldn't work out.

[JonMarsh]

For anyone curious, here's how the LspCAD schematic is translating into a BOM at this time- it's also a good example of how even a mid to low spec component three way crossover always winds up being a bit pricey unless you're keeping really low order networks, in which case performance is likely to be compromised. This is especially the case when the woofer to midrange crossover is attempted at lower frequencies... in this instance, about 300 Hz. You wind up with much bigger caps and inductors all around, and no two ways about it, that costs money. In comparison, the Ardent has a midrange to woofer crossover point about an octave higher, 650 Hz. With the single woofer in this design, I wanted to try to have as much of the midrange as possible on the Seas driver, also considering that the Seas Midrange doesn't go as high before needed to really be slapped down hard.

[BobEllis]

Gee, you can build a very nice second amp and active crossover for that much money. And re-use the amp next go around. Active anyone?

[JonMarsh]

Yes, but then you wouldn't have a complex wide range test loudspeaker to evaluate as a full range load, AND, some of the crossover transfer functions would be difficult to implement in an active crossover- not straight forward LR4, even, as both the mid and the woofer carry part of the baffle step compensation.

Closest tool I'd have that would do this is my Metric Halo LIO-8 with the DSP upgrade, but then I'd be locked into using it as the DAC. Worse things could happen, but then I have an AURALiC VEGA for this project …

[Steve Manning]

For anyone curious, here's how the LspCAD schematic is translating into a BOM at this time- it's also a good example of how even a mid to low spec component three way crossover always winds up being a bit pricey unless you're keeping really low order networks, in which case performance is likely to be compromised. This is especially the case when the woofer to midrange crossover is attempted at lower frequencies... in this instance, about 300 Hz. You wind up with much bigger caps and inductors all around, and no two ways about it, that costs money. In comparison, the Ardent has a midrange to woofer crossover point about an octave higher, 650 Hz. With the single woofer in this design, I wanted to try to have as much of the midrange as possible on the Seas driver, also considering that the Seas Midrange doesn't go as high before needed to really be slapped down hard.

Jon ...... not that it matters much, but I think the total for L6 in the woofer section is off?

[JonMarsh]

You are so right, and that is so weird, as this is supposed to be a calculated cell, applied to all the right hand column. This was a recycled version of the TA5 Ardent BOM.

Oh well, only bought one batch of parts anyway this time around. I figure if I'm really lucky, I'll be able to log in to PE again next month, and just re-purchase that order...

Thanks for bringing it to my attention.

[5th element]

That is certainly a considerable amount of outlay into xover parts!

[JonMarsh]

yeah, but I could double that if I wanted to gild the lily on the caps and inductors, particularly for the midrange and tweeter combo. I'm hoping to "squeak by" with this combo. Any suggestions for different components of equal or better quality and lower cost are welcome, of course!

That list IS for a stereo pair- just compare that to an Ardent, though... recollection is we spend about $1500 on a stereo pair. If this was built to that quality level, it would likely be crowding $2K.

[5th element]

Any suggestions for different components of equal or better quality and lower cost are welcome, of course!

I could definitely make recommendations towards cheaper components that's for sure

That list IS for a stereo pair.

Ah! That does make things a lot better. I had assumed this was mono!

[---k---]

Jon,
Do you get to turn any of those receipts in on an expense report?

[JonMarsh]

Jon,
Do you get to turn any of those receipts in on an expense report?

Boy, I wish... that would make this a lot less painful. But I even personally purchase much of my engineering software at work, so that I'm not limited by what IT or a Marketing manager things I need to have, or is willing to pay for. There are signs my new manager will be more accommodating, but his hands may be tied by budgets given to him by his VP.

In our US team, I'm the highest level technical person, but I'm not management these days (was at one time- switched back to technical ladder track). So, when I feel I need to invest in my own career, well, I just do. This does raise eyebrows from time to time, but it's part of why I'm at the rank I am now (which required interviewing and presenting to our HQ management in Europe to be approved). I get things done. That's what matters, and not letting my knowledge development being limited by marketing managers. (Gee, Ryan, hope you're not in Marketing... :W )

Made some progress on these this weekend- finished the first woofer crossover build, and almost finished the midrange crossover. Hope I'll have that and the tweeter crossover done and be able to do first tests next weekend.

I'm also doing some deep black ops on the Class D side, having revived working on a design I was playing with back in 2011. A hybrid mixture of discrete and a little bit of IC technology; even more interesting because of some newer 100V and 150V (BSC190N15S3) power MOSFETs we have with excellent characteristics for Class D applications. This will really have to stay off the books, but reading about the AURALiC MERAK's and what NAD did with the M22, I think I had some good ideas back in 2011 and need to follow through, just for my piece of mind, and to have something else to test with. A lot of the pieces and components are just coming together nicely, like the LME series of TI/National OpAmps, including the LME49860, with up to 22V rails, instead of the 15V nominal of the LM4562. This allows doing a hybrid setup with instrumentation style front end carrying part of the system gain, and minimizes errors in the modulator power stage by allowing running lower gain, more feedback and reducing things like DC offset inherent in unmatched comparator parts. Of course, I could also try to translate the whole design to ICs, but old habits die hard, and I think Bruno had the right idea all along the way he did the UcD modules and NCORE mostly with discrete. When you try to combine too many parts into one IC, usually the process limitations result in some compromises, like the noise figure for the integrator/comparators on many integrated CLASS D modulator solutions. Can't come close to touching that of the LME series opamps from TI.

http://www.ti.com/product/lme49860

[JonMarsh]

well, it pays to keep up with other discussions on HT Guide. I didn't know about the CSS SDX10, but that's an interesting looking driver. Could be the ticket for even more output capability at similar sensitivity for this design. Ordered a pair, based on the impedance curves and other aspects of the design, I expect them to test out well, hopefully they don't have some of the execution issues other XBL2 designs have had in the past... a fairly flat impedance curve to 1 kHz, 18mm of Xmax, and sensitivity of 85 dB for 2.83VRMS is what's got me interested.




If it actually combines well with the Scanspeak 26W/00-0 PR, we just might have something rather interesting here... perhaps a bit brutish, but this time I think that's what I'm looking for... :B

[JonMarsh]

A couple of fairly interesting alignments are under consideration now, depending on what's tolerable size wise, and desired extension, as well as system Q.

This dual PR alignment at 40-45L gives pretty good extension and some protective loading down lower, somewhat resembling a classic ported QB3 alignment:

Image not available


Going deeper, at higher levels, is possible, but requires upping the enclosure volume to 60L, and using 3 of the Scanspeak PR's.



The Fb on the larger box is about the same, but the output is flatter, with an F3 of 24 Hz instead of 28Hz. Whether this is feasible or desirable hinges on how it impacts the overall enclosure/system design, which I'm thinking now should be in two pieces. With a modular bass unit, I could build either way on the bottom end, or both, for testing purposes.

[TEK]

Hi Jon
Looks interesting!
I was unable to find out if there were any spesific design goals for the "Fundamentals"...

[JonMarsh]

Well, this is kind of a weird project- I talked a little about it at the beginning, but there sure wasn't a nice "Executive Summary" anywhere... so, perhaps that is in order...

• Fundamentals- partly this is a word play, the focus was developing the bass portion of the design for a three way to use the Seas Concentrix C18EN001 dirver- in English, the fundamental frequences are often referred to as the low frequencies or bass.
• Originally, my concept was to make the smallest possible "mobile" speaker that had as wide a bandwidth as possible, intent being amplifier evaluaiton over a wide frequency range - this is where using the SW223BD02 with a PR came from.
• Being a difficult load is just a bonus for the design, as I want to see how load invariant amplifier modules and power supply combinations are (the Class D thing...) - so, having a 4 ohm woofer system and 8 ohm concentric with more convnetional style crossovers is "perfect" (though not my normal choice) as the impedance varies from 2.5-3 ohms in the mid bass up to 12 or more ohms in the upper crossover region
• The project is suffering from feature creep/performance creep- now looking at a bigger woofer, and multiple PR's, with a slightly larger cabinet (40-45L for the woofer instead of 30-35L)

If I were just making a "normal" system design, I'd use two woofers in series, and something closer to my normal crossover approach, to get an impedance curve more like the Wavecor Ardent.

This is really more a project for Evil Twin; I should probably have him take it over.... the reference amplifier for voicing will be my Aragon 8008 X3, or my Aragon Palladium mono blocks; as they have enough current drive for arc welding; then after dialing everything in, we'll see how "reference" class D stuff does with them (my NAD M22 and my nCore 400's) and how the IR modules and the power supply I'm developing for that work.

Oh, and I'll probably try a revival of Class D design I was working on in 2011; using some of our newer 150V OptiMOS3 and OptiMOS5 power MOSFETs; it's moved away from using some IC parts to a nearly all discrete design, but with some VERY high performance TI Op Amps for the instrumentation amplifier style front end, and with a Jensen input transformer for really keeping the CMRR and noise rejection as high as possible (yeah, I'm borrowing/stealing ideas from the AURALiC Merak mono blocks.) I'm already in simulation on this effort, with promising results, and have sourced a lot of the critical parts already. This probably won't be SOTA, but I think they will be fun, and they'll follow my past practice (and current nCore practice) of using a lot of local bypass- I've got a pretty slick setup worked out for that, and amazingly sourced most of the premium caps on eBay!

[5th element]

and amazingly sourced most of the premium caps on eBay!

Lets hope they are genuine!

[JonMarsh]

Lets hope they are genuine!

Actually, I've had very good luck with that so far; mind you, they're premium caps in terms of industrial switch mode power design, not conventional audio, though I did OK on that with some Nichicon parts from a US vendor. The parts I really like for the value proposition and low ESR and ESL in electrolytic are Rubycon ZL's. Much better in my opinion than anything similar Nichicon offers, and with very good high temperature operating life. About the only way to beat them is OSCON polymer style capacitors, and then the cap volume versus ESR has a very different set of tradeoffs. Those I get from Digikey.

The WIMA film caps came from Mouser, actually, who is an authorized dealer.

[JonMarsh]

CSS SDX10 in the "flesh"

OK, got a box today with two pretty heavy drivers... not impressive packaging, but adequate- but the driver are another matter! :B

If a picture is worth a thousand words, I don't need to write anything else in this post.

[5th element]

Well those are certainly built to last!

[JonMarsh]

California Audio Show in Millbrae this weekend



Here I am missing out again- the California Audio Show, in Millbrae, CA, just south of SFO, is ongoing- and I'm neither going to that, nor working on one of my own projects, but working on work-work the second day of this weekend.




Mothers, don't let your sons grow up to be Applications Engineers!

[csmielke]

Hey Jon,
In the model I did at 47.5 liters you mentioned it being net 46 liters which I assumed took into account the driver displacement in the cabinet. I found another website which boasted a calculator for driver displacement at DIY Audio and Video. I entered the physical dimensions on the drivers and it spit out .72 liters for each PR and 1.73 liters for the SDX10. The total was 3.17 liters so my enclosure is now at 50.67 liters. I was surprised by over 3 liters of displacement but when you look at the SDX10 it's reasonable.
Your thoughts?
Chris

[JonMarsh]

That sounds OK to me, let's keep in mind at this total volume, the difference will be small in impact on the response curve. Probably barely noticeable in a close comparison in Unibox.

Are you going to build some boxes soon? I just bought back the test boxes with the SW223BD02 and the C18EN001 from storage, hope to test this coming Labor Day weekend.

[csmielke]

It's the next step, I have the SDX10 and a pair of the PRs. I have to do some fiddling with the trapezoid area formula to get the right shape and volume. I am also pondering what material to use? MDF or a dual layer MDF/plywood combination to get the stiffness at a cost effective price. I then plan on using walnut veneer to match the mains (Tyler Acoustics D3Ms with updated crossovers via Danny Ritchie's help). I am not very good with a camera but will try and capture something. I also will be of little or no help in measuring outcomes since I don't have the gear. It will be my ears only.

[JonMarsh]

I suggest going with high grade ply if the money is OK for you. Either will require some prep work on end grain to get a good surface for veneering- epoxy is my preference, such as West Systems. (I'm actually a Baltic Birch ply fan, but in many areas that's hard to find- may require a specialty store rather than a big box store).

[flamethrower1]

The stuff you post both gives me a headache and amazes me, god it sucks to be nothing but a good cabinet maker

[JonMarsh]

Hoy se Cumplen Seis Semanas

First weekend off in forever (it feels like) - mostly working on long neglected Honey Do's and Honey Must Do's, including finishing the unfinished maple dresser bought back in February (that's how far behind I am...)

Right now, I'm listening to the post title, part of the Cuban Music CD GF brought back from Cuba... I think it's like Clint Eastwood, and like the NatalieP's, 'cause I'm feelin' lucky

First crossover tests on the test box configuration...





This was the first test sweep, direct on axis, no tweaking, just putting them on the stool as a stand, and using the kind of overall position that's worked well in the past (re height of woofer, if not on the floor, such as with the Modula Extreme experiments).



Too hot to do testing outdoors today, (94) or I would have considered that. The 50-60 Hz is a known room mode in this location I use for testing, and I expect the 175 Hz is floor bounce. Plan was to test with woofer and midrange unit in a variety of heights and relative locations. Plan was followed.

For comparison, looking in detail, here is the woofer module on the "stand" (nice maple construction, bought online) and on the floor, per Matt's suggestion and my own intent to investigate going that way, given the low crossover I'm using.




As expected, the floor bounce goes away. As not expected, the room mode is worse. Curious.

Following this, I spent about an hour with the woofer on the floor, experimenting with various heights for the midrange module and a variety of measurement distances. They all seemed to introduce various and different flavors of problems with the woofer to midrange transition, some at rather unexpected frequencies. (i.e., further away from the crossover point than I could just write them off to level and phase issues. Walking around the room and listening with music (East River Blues, Changes (Loggins and Messina), Hotel California (HD Tracks), I ended up liking both the measurements and the sound of the original setup- stool stand and relatively close coupled mid to woofer.

To go with floor mount, I need to completely re-do the crossover. And, perhaps reconsider the concept. What is making sense to me now, is two versions- a large stand mount version with the SW223BD02 as originally conceived, with 1 PR, and a floor stander with the CSS SDX10, and 2 PRs, in a 35 liter cabinet for the low end, and see how well we can optimize the crossover setup that way. Something for a larger room; I wanted this initial version to have good integration up close, and what with the concentric midrange and the spacing this lashup has, that it does.

Now, about that on axis presence bump between 2-5 kHz... do I fix the on axis, and put a dip in off axis, or just leave it like it is? This is 0-15-30 degrees.



These do sound a bit like what I would call a monitor- basically pretty neutral, wide range in bass but nothing emphasized, and they're probably doing a good job revealing what the TC Sounds audio interface sounds like, and the NAD M22 hooked up today for testing is not the least bit forgiving or euphonizing the output. Bass guitar and percussion sounds quite good up and down the scale.

Now I'm listening to "All or Nothing at All" from the "Daria" album... now I know I need to listen to a pair of these with a first line DAC...

Overall, I think a pleasing voicing would drop the mid to tweeter range about 2-3 dB, and give that a bit of a listen long term with the good electronics... once I get the real enclosures and another set of crossovers built!

Vocals sound good, really good with this setup... at least Daria's do! :B

[JonMarsh]

This looks to be pretty straight forward, checking back on the actual build and my LspCAD files-



the "as built" value of R3 wasn't 9 ohms, but 8.2 ohms; if I change that to 12 ohms, and change R1 from 4 ohms to 6 ohms, it looks to me that will be in the ball park for the voicing change I described above.

I have to say I'm really enjoying listening to Daria on these right now as is... that's a good sign for the basic driver character, I think.

Of course, one other nice thing- great vertical window! As well as horizontal. 30 degrees sounds just fine. If nothing else, I could also bring these with the crossover tweaks to the Burning Amp demo, besides the Ardent, just play one. No way I'm going to have time to build these before the event unless I took a week off after the FAE training.... hmmmm... why not? Oh yeah, the Class D SMPS project, which will be really hurting by then...

[Steve Manning]

Glad to see you get some play time in Jon ...... What's the volume on the mid cabinet? I suspect it's in here some place. I could see you doing a modified version of the TAD CR1, if you go the largish stand mount route. Going with the CSS SDX10, and 2 PRs, if you could live with the PR's on the sides, you might not have to do a floor stander. Depending on the volume requirements of the mid/tweeter, you might be able to go stand mount as well, and put the crossover in the bottom of the stand?

[JonMarsh]

With the SDX10 I need 2 to 3 PR's for the optimal tuning. For reasons of transport, (and reasons I have for being able to transport them, at least one set) going with the Wavecor is more compact and of course lighter.

Still, I think I should give the SDX woofer a shot, with floor mount, but then they would really need more integration distance.

Another possibility is a two piece cabinet approach; same top module for either design, and different LF sections.

Midrange volume is not to critical, 24-28L helps keep The response reasonable, but note the LF conjugate network to deal with the impedance peak.

Now, when am I going to find the time? :W

[5th element]

Well Jon you always know how to get a lot done with a small amount of time. This is what happens no doubt when you're used to working engineering deadlines!

It's nice to see that the floor bounce completely goes away with the boundary mounted woofer.

When I first tried the FST coax the vertical window was like a breath of fresh air. I do a fair amount of near field listening with mine too so it was especially noticeable. With wave guides/coax's though the optimum toe in is to actually have the loudspeakers toed in a lot. You need to have the on axis cross in front of your listening position, this is a very severe amount of toe in but it ends up giving you the widest sweetspot uniformity and also gets you listening to a decent axis off centre which is usually where coax/wg designs are at their best.

I think the degree to which that presence bump will need attenuating is one of those things that will become apparent from extended listening and how much baffle step/low bass emphasis/how much tilt down you build into them.

[JonMarsh]

Well Jon you always know how to get a lot done with a small amount of time. This is what happens no doubt when you're used to working engineering deadlines!


I think the degree to which that presence bump will need attenuating is one of those things that will become apparent from extended listening and how much baffle step/low bass emphasis/how much tilt down you build into them.

My current intuition is that it is the overall power response I'm more concerned with, and i think with the alterations to the mid and tweeter level that are planned, I should be in the right territory-

Have some Mundorf resistors picked out already to order...

Ya gotta hear vocals on these to really appreciate the near field coherency in that range that derives from the coax. These should be pretty nice when finished, I think. :W

But I think I will do a two pice tower version with the SDX10, to try that development path, too. I could place the woofer at one end of the enclosure, and experiment with floor loaded and "air" loaded (different cross overs, of course). That I don't see getting to until next year. Work is scheduled to be a B*tch through next April at this point, unless one particular project is killed (it's a possibility, because it turned out my immediate boss didn't have it approved with the VP. Sort of a case of "Fire, Ready, Aim".

And I haven't forgotten about that tech paper, either, but with all the other things that have been loaded into the hopper, the parallel processing isn't proceeding very quickly. Sort of like giving six tasks to a two core processor...

[Steve Manning]

Sounds like you need to clone yourself or pick up some Minions Jon . Wish I were closer, being off work at the moment I could play the role of a Minion.

[Evil Twin]

Sounds like you need to clone yourself or pick up some Minions Jon . Wish I were closer, being off work at the moment I could play the role of a Minion.

I am not completely satisfied with this new notion "minions", but anything which advances the endeavors of the Dark Side must be entertained in this day and age...

[augerpro]

Ya gotta hear vocals on these to really appreciate the near field coherency in that range that derives from the coax.

I noticed the same thing when playing with the Kef Q100 drivers. Very impressive in that regard.

[Steve Manning]

Jon, I was looking at the spec sheet on the C18EN001 and it was calling for a .2 cu ft sealed enclosure for and F3 of about 135Hz, seeing that your test box was at .75 cu ft, I was curious as to why the extra volume?