Ardent Speaker Camp

[Face]

As long as the caps are of high quality, it's not an issue.

[dmalphur]

I think Jon just called Tony Gee's crossovers simplistic...

He may be the only person that can get away with that comment! :lol:

[JonMarsh]

hello jon,
ive been reading and reading.......and reading, and im very impressed to say the least!
i have been seaching for another design to build for years now and i have to say that im interested.
the problem is that im advised by many that its important to have as few components in the signal path as i can get away with. there are lots and lots of them in this design- i know there is little theory to support the sq loss of doing this - but still im uneasy with the idea of all those caps on the signal path.
i have tried to get to the bottom of this idea but i just cant. -its suported by the fact that many of the speakers i have realy been impressed by were 1st order, with the aim of as few components as possible.-can anyone offer some enlightenment on this?
im kinda between tony gee's soup sandwich and this atm.

ps- just awsome cabinet work!

I understand your concerns here- let's look at what's really happening.


First, on the tweeter, there is one cap in the signal path. The only way to get rid of that cap is if you have a very high sensitivity tweeter in proportion to the design, and you use an inductor in shunt instead, with a series resistor at the input. That cap is fairly critical for quality, and must avoid internal electro-striction resonances at all cost- that's why I'm using a Clarity MR. It's the cheapest cap I can find that meets the performance target. Upgrades are possible, of course, if you have deep enough pockets for Duelund Cast caps, for example.

Using the resistor/inductor on the input only gives you a first order roll off, though- which is a problem. I built 1st order designs that were transient perfect on axis in the 70's, but the thing folks don't take into account about first order networks is that they operate with some phase quadrature, so the peak summing is no on axis but in an off axis lobe. Kind of messes up the power response. I like the Duelund crossover because of it's mathematical elegance and because if implemented correctly, all drivers are in the same relative phase at any frequency. No phase quadrature through the transition region. In fact, pretty much the entire speaker bandpass is the transition region!

Looking at the midrange, there are just two reactive components in the direct signal path- an inductor and a capacitor. You'd have the same for a first order. THe forward Caps are Clarity ESA, the "Budget" high end caps addressing electro-striction and other manufacturing parameters around sound quality. The series inductor is a low DCR type which has no measurable skin effect below 20 kHz.

The LF shunt LCR is in a range where capacitor colorations won't impact the sound significantly, and operates to make sure the midrange is rolling off properly. The upper shunt LCR, shaping the midrange response around 2 kHz, is more critical, and uses a Clarity ESA cap also. The ESA is my favorite bang for the buck cap- not cheap, but certainly deserving of the brand name "Clarity". :W

On the bottom end, it's not caps that are the worry in the crossover, but inductors- here I'm using a very low DCR construction main inductor, to promote good control of the woofers and minimum insertion loss, and an LF zobel to control interaction of the driver impedance with the network. The acid test in my mind is a number of recordings with standup bass; does it have slap and punch, but without bloat? Is the harmonic structure of the instrument intact, to where you can practically hear the wood? That's part of why I'd been listening to Acoustic Extravaganza by KT Tunstall (all the bass work is acoustic), as well as some Jacque Loussier recordings.


Cut number 11, a live recording of Bach: Partita in B-flat Major, of this album is particularly visceral- on both bass and piano. They don't normally do "live" cuts, and haven't done a live album, but this one is quite worthwhile- makes me which they did do them.




Yes, that's 50th anniversary of the Jacque Loussier Trio- but it's his second set of sidemen, he wore the first set out. Jacque was born in 1934, his most recent recording released in 2006. I have about a dozen or so, including several SACDs.

One thing I'm pretty convinced of is the resolution from the midrange through presence range and top end this system has been delivered as desired. If I knew what I know now back this time last year, the only thing I'd do different would be to go for a gutsier higher output bottom end- but since these are going in my master bedroom when finished, it would be kind of overkill. But the same system concept, but say, with dual 8" Scanspeak 22W on each side, would be pretty killer for a main room sysetm. (and about $1,000 more expensive to build- where do you draw the line?)

[TacoD]

Very nice x-over, it is also my experience that good transfer curves are more important, i.e., brings more sound quality rather than sound degradation due to the extra components

Also if you are on a budget then you can better buy some mass produced speaker, in my view DIY allows you to do things properly rather than let the account rule the design.

Tony Gee tries to fix his x-over with expensive caps. In my view you have to start with a good x-over, then good coils/ caps will pay off even more.

If you want low DCR coils which also can take some power, I can recommend the Mundorf I or E core coils. Especially the 2mm wire ones are really heavy, and you immediately hear the difference in bass control.

[lbstyling]

Jon, I cant thank you enough for the time you have taken on that answer. :T realy appreciated.

can you provide the basic perameters for me? max spl before say 1% distortion and absolute max, over all efficiency etc?

i think i would try the basic version without L10, C14, and R10, as i have a class t amp with a great load tolerance. (must say im blown away with the higher power tripath amps, whipped my parasound halo in every way! 8O )- nice saving there!

and outboard crossovers me thinks.

i for one would be in favor of the dampened baffle area around the tweeter- the neat acoustics speakers use this around the focal t120, and having been to a demo of a before and after, there was a notisable clarity added to the trebble with it.- gotta reduce all the reflections we can- lord knows ive got no chance in my room once the sounds out in the open!!

tacod- a big coil is great, but im not sure about a iron core one, i was under the impression they wander in value under load/temp??

[JonMarsh]

To some of your questions,

Jon, I cant thank you enough for the time you have taken on that answer. :T realy appreciated.

can you provide the basic perameters for me? max spl before say 1% distortion and absolute max, over all efficiency etc?

i think i would try the basic version without L10, C14, and R10, as i have a class t amp with a great load tolerance. (must say im blown away with the higher power tripath amps, whipped my parasound halo in every way! 8O )- nice saving there!

and outboard crossovers me thinks.

i for one would be in favor of the dampened baffle area around the tweeter- the neat acoustics speakers use this around the focal t120, and having been to a demo of a before and after, there was a notisable clarity added to the trebble with it.- gotta reduce all the reflections we can- lord knows ive got no chance in my room once the sounds out in the open!!

tacod- a big coil is great, but im not sure about a iron core one, i was under the impression they wander in value under load/temp??

Max SPL before "X%" distortion is rather frequency dependent, due to the nature of the drivers. The plot for 90 dB sweep distortion gives some reasonable guide:



Keep in mind my target for this was just a speaker that's comfortable with ~100 dB peaks. I'd measured quite a few 7's for HD before selecting the ER18RNX; they were the lowest distortion, and tend to be the limiting factor for the system. Dual SS 22W8867 would change that... note the distortion I measured for 26W in the upperbass/lower mid:



For correlation, -40 dB is 1% distortion, -60 dB is 0.1%.

Sensitivity is basically set by the parameters of the Seas ER18RNX- plus crossover insertion loss. The ER18RNX is ~88 dB/2.83 VRMS on an infinite baffle; parallel them, that puts you up around 94, put in full baffle step and a dB or so crossover insertion loss, and you're back down around 87 dB/2.83 VRMS for the system.

With cored inductors, it's all dependent on the design. A laminated steel inductor such as the Mundorfs can work well and handle fairly high power if the core is sized adequately. Solid steel laminations tend to have more eddy current losses, but for a woofer that's rarely a problem.

The large LF inductor (Jantzen C-Core) used in the Ardent is a powdered core toroid which has pretty low distortion and very good permeability stability up to about 400W. 400W will toast the ER18RNX pretty thoroughly, so I think that's sufficient. The toroid also has low flux leakage, so there's minimal interaction with other inductors. One company using laminated steel inductors (Erse) notes in their online material that their inductor has lower distortion at 500W or so than the Jantzen; they don't promote that the Jantzen has lower distortion at lower output levels (a factor of that eddy current issue).

[TacoD]

The measured performance of the Mundorf is very good, I also used the jantzen (with duelund label) and these are also very nice. But I prefer the larger Mundorf for bass duty.

[augerpro]

Jon disregarding insertion loss for a moment, would you still use the C core over 14g air core if the woofer was running higher, say as in a typical two way design?

[JonMarsh]

Well, even in a typical two way, because of BSC, you may wind up using a 3 mH or so- like in my little Modula MT XE. I used a Jantzen C-Core in it for the same reasons- damping and insertion loss. Makes more of a difference with a four ohm load to have 0.4 or 0.5 ohm DCR than for an 8 ohm. Obviously, there are cost performance trade offs, but the 3.3 mH air core is about 0.6 ohms, while the C-Core is 0.07 ohms.

There might be times when for reasons of tuning the Qts differently, you'd want the extra series resistance. Other than that, and cost, there's no reason to use the higher DC parts I can think of.

[augerpro]

Interesting I may try the C core for my high effeciency speakers. The can handle much higher power (not that I'd really need it) than a single 6" woofer, do you think that would cause problems, again in comparison to an air core? I guess I'm not totatlly clear on the pros and cons of the toroid versus a good quality steel core compared to air core.

[JonMarsh]

With solid core inductors or transformers, it's all in the implementation. That means the core design, the permeability, the B-H slope, and where the permeability will start getting nonlinear as you reach the limit for that material. The C-Core appears to be a powdered iron material of some kind with an insulating binder; that means it's a distributed gap type of material, and distortion on those inductors is pretty low in the first 300W, still pretty usable up to 500W+. An air core will be more linear, true, but you should see the size of some AWG12 air core inductors I had wound at North Creek! Ginormous!!

Some core materials deservedly have a bad rep- like the bobbin ferrites, as used by many manufacturers, which have moderately high permeability (a high AL product) but won't sustain very high flux, and when they go into saturation, it all goes to hell in a handbag in a hurry. So, for example, stay away from the Jantzen P-Core except for low power applications. I've used transformer steel as well as low permeability powdered toroids in making inductors since the early nineties, but my coil winding machine is at Tom's, and I'm getting lazy- easier to buy now. For your high efficiency speakers, I'd still consider using the C-Core for a passive crossover, if you have inductances in the 2mH range or above.

[augerpro]

Thanks Jon!

[lbstyling]

thankyou jon, awsome answer...as always.

you should release a book- kinda a- 'Loudspeaker Design Cookbook -Level 2'- i dont care what it costs, i would buy it- and im sure many others would too.

when i asked about the spl at 1% ect, im trying to get a feel for how well it would take transients as a L+R in my surround setup. My current TJL3W'S are very low destortion (seas excel), but are too limited for my suround use in terms of max output and distortion at these kinda levels, as well as what seams to my ears as dynamic compression aproaching these levels (effect seams audiable at ref -15 and above, with lpf set at 100hz)
it seams more noticeable as the distortion at lower levels is so low, theres a bigger contrast in sound between spl levels.

if the limiting factor in spl is more the bottom end, this is a positive factor for my situation.

thanks all for the low down on the coils- ill pop that in the memory banks. appreciated. :T

just stunned at those distortion graphs!!!! :E

[JonMarsh]

I hear where you're coming from- when you have drivers like Excels or the Scanspeaks, or other good low distortion parts (a lot more available these days at moderate prices than was the case five to 7 years ago), then when things do finally take on an edge or develop some grunge, the contrast is quite noticeable- like listening to really good headphones but then running out of gas on the headphone amp.

Of course, unless you're in a really big room, that might be pause for thought that the SPL's are getting a skosh carried away.... there's a lot of difference between peak transients on well recorded music, and things like "Santana Supernatural" where everything is just plain loud (mastered for radio, obviously).

Well, as much as you like those distortion graphs, I'm working on something better- the Modula Xtreme, with dual SS 26W on each side, Accuton C173N-T6-90 on the mids, and the Scanspeak D2608/9130 waveguide loaded. Those should be interesting... because it's a "Modula" system, the restriction is that they have to be built with PE cabinets. Will be interesting to see how that works out.

Here's the head unit CAD model (mid and treble)




Initial tests show the waveguide solves a lot of issues with diffraction in the tweeter range, simplifying cabinet design. That T6-90 Accuton is VERY low distortion even at rather high SPL.

This is at 100 dB output.



Custom front panel from LBL for all cabinets, and a custom mount for the head unit fabricated from LBL. Using 1 cu ft PE sub cabs for each 26W SS. It's part of this weeks "Speaker Camp" work, too.

[lbstyling]

ive got about 3.5-4m between me and the speakers, so i dont know if you consider this a very large room (300m3 total) but it does make speakers have a slightly distant sound to them (room issues possibly) its not drastically bad, but still noticeable.

when the wifes out i bring them forward to 2.5m and its a nice improvement. Baised on this, im fitting lots of shelves across the wall behind the seating position to help the room accoustics (hopefully!) by increasing diffraction/absorbtion. sold to the wife as a place for all her books! he he.

as long as the baffle isnt larger than 10" wide(in a classic shape) , i can likely get away with almost anything in terms of swimbo, so im interested!

the wave guide on this modula may even help the room interactions being a little more directional, so no complaints here! :T

my musical tastes are wide (isnt everyones if theyre honest?) but i have a soft spot for LARGE scale orchestral work like John Williams, where i have the average level at 85-90 db(ish) with dynamic peaks and cresendos gathering on top of this. This may be a result of my youth spent with early car subs 10 years ago involving building a system capable of 152db @ 30hz. tut tut. :roll:

looks like the modula xtreme may be a better choice for me then, i do realy want the speaker to end all, and in my room this may be it.

[lbstyling]

mmm- more graphs!!

ive got about 3.5-4m between me and the speakers, so i dont know if you consider this a very large room (300m3 total) but it does make speakers have a slightly distant sound to them (room issues possibly) its not drastically bad, but still noticeable.

when the wifes out i bring them forward to 2.5m and its a nice improvement. Baised on this, im fitting lots of shelves across the wall behind the seating position to help the room accoustics (hopefully!) by increasing diffraction/absorbtion. sold to the wife as a place for all her books! he he.

as long as the baffle isnt much larger than 10" wide(in a classic shape) , i can likely get away with almost anything in terms of swimbo, so im interested!

the wave guide on this modula may even help the room interactions being a little more directional, so no complaints here! :T

my musical tastes are wide (isnt everyones if theyre honest?) but i have a soft spot for LARGE scale orchestral work like John Williams, where i have the average level at 85-90 db(ish) with dynamic peaks and cresendos gathering on top of this. This may be a result of my youth spent with early car subs 10 years ago involving building a system capable of 152db @ 30hz. tut tut. :roll:

looks like the modula xtreme may be a better choice for me then, i do realy want the speaker to end all, and in my room this may be it.

thanks Jon. :T

[jweb09]

Jon, I have been tracking this thread for months now. Any idea when the PDF file for the design might come together??

[JonMarsh]

Before the end of April. All of the individual CAD PDF's are done, anyone that wants that just PM me, but writing a build PDF like the one for the Modula MT XE will take a bit of time- it's started, but this week I'm focussed more on making sawdust than pushing pixels- finishing up some veneering and mounting the crossovers internally.

And the Modula Xtreme project is taking stealing some CPU cycles, too.

[PewterTA]

:bump:

Just checking in on any new updates.

[ThomasW]

Jon was kidnapped mid April by his employer. There's been no work on speaker projects since that time and none anticipated in the near future

[Ray Collins]

Thanks for the update Thomas...

Ray

[JonP]

Well, good to hear that it's merely a case of kidnapping and nothing else. I was starting to get worried with this most un-Jonlike behavior.

I was so looking forward to finally meeting this crazy guy at the NorCAL DIY event, hearing all his new designs, especially comparing my Modula MT's to the new XE version, etc... Quite a few there were wondering where he was, thruout the day.

So, glad to hear it wasn't the Singaporean Mafia, (or German, or Rebellion for that matter) and now we just have to figure out the hostage release/rescue plans.

[Bear]

Well, good to hear that it's merely a case of kidnapping and nothing else. I was starting to get worried with this most un-Jonlike behavior.

Have you read through the Neo D CC thread?

[JonP]

Been a while.... Not sure what you mean, unless your talking Jon and work. I was thinking in all the 4yrs or so I've been hanging around here, I don't think he's ever gone 2 weeks (1 week?) without posting... they must have him locked in a dungeon, shielded from all net access.

[Bear]

Been a while.... Not sure what you mean, unless your talking Jon and work. I was thinking in all the 4yrs or so I've been hanging around here, I don't think he's ever gone 2 weeks (1 week?) without posting... they must have him locked in a dungeon, shielded from all net access.

The answer is in the Neo D thread. Yep, Jon had an extended forum absence right after getting the basic design finished, but before he could complete a high-end design. Something like 9 - 10 months went by with almost no Jon sightings. :E

[JonMarsh]

You know how all those really snarky high end companies talk about new models or improved versions of existing models incorporating technology and lessons learned from their most recent flagship product or project? :twisted:

Yeah, well, coming right at you, in a few weeks. Not enough changes to be called an Ardent MkII, but still probably somewhat significant.

Besides, gotta replace that smoked Be 6640 tweeter... ouch! :E :roll:

[silvercans]

Smoked? Uh oh, I hear Be is quite rare in Evil Twin's homeworld of Tatooine, did he try to steal your Be tweeters for his experiments?

[JonMarsh]

It was carelessness on my part with crossover boards all over the floor, and two touched in circuits that the test amp didn't like, and it oscillated at rather high frequency- could barely hear it, or the undertones from it.

Ah well.... live and learn. See, learn from my mistakes, save yourself some money and time!

[Face]

Do you still believe the Be version is worth the added expense?

[JonMarsh]

Do you still believe the Be version is worth the added expense?

I do, but that's a personal judgement for myself- I'm buying another one to replace the one I smoked. It compares pretty well with the diamond Accuton in sonics (resolution without hardness), but goes lower than the $2K Accuton. OTOH, with a little care one can use the standard 6600-00 to very good effect, as well as the D2608/9130. I think the difference in the 6620-00 with the curved "mini-waveguide" (if that's what it is) is what is doing what I personally don't like sonically.

But obviously, that's a rather subjective call, and others may disagree with my feelings about the 6620-00. Just don't pick on me about it, or I'll call out my Sith Lord friend to help... :B

Evaluating them accurately is of course subjectively is dependent on a lot of factors- it's not hard to have a system that gets most of the lower frequency stuff (say below 2 kHz) fairly right, but above that there's a host of factors besides the tweeter as you know!

I'm heading out on travel this week, but will contact you and take care of the PayPal once I get back on the weekend- hope you've got them still! Thanks for being patient.

[Face]

Thanks for the response.

Yes, they're still around somewhere.

[benthe8track]

I've been playing with some more "exotic" materials like phenolic sheet, and I'm glad to find something easy to source and so nice to work with- and very stiff. It's a great "launch platform" in my opinion, and at least deserves to be used in the front panel where, IMO, the stiffest, strongest coupling of the driver to the rest of the cabinet is needed.

McMaster-Carr has an awesome selection of exotic materials in good sizes for speaker building. I've ordered carbon fiber and a few different plastics from them but the shipping is brutal up here.

[penngray]

I just read the last 10 pages to get caught up.....This is still an incredible learning experience. Thanks Jon Marsh and others who coninue to post incredible details so the rest of us continue to learn (and feel like total idiot noobs....Maybe that is just me! )

Im very interested in the 6640 discussion. Since Im looking at the expensive RAALs this is the other alternative. A waveguide with a 6640 would be a sweet design, Is that what the "Module" Design post represents?

[JonMarsh]

The Modula Xtreme isn't that Xtreme- just using a Scanspeak D2608/9130, some of the bigger SS tweeters may work, like the 66000, but with their high 2nd order HD, I'm partial to their little brother, the 9130. Might try the 66000 some day in a waveguide anyway.

So far I've found hard dome tweeter are mor difficult to get to behave well in the 10-16 kHz area; the faceplate and grille of the 6640 might make that even more difficult.

[Face]

So far I've found hard dome tweeter are mor difficult to get to behave well in the 10-16 kHz area; the faceplate and grille of the 6640 might make that even more difficult.

You could swap face plates.

[JonMarsh]

Well, that's an idea- the 66000 and 6640 have basically the same faceplate by appearances, the 6620-00 is quite a different beast. I do have a "dead" 6640 I can disassemble to look at. Wonder if I could mate the dome with a 66000 VC assembly? I mean, I'm good, but I'm not THAT good by any means! But for only $44.60, not much financial risk...

After taking apart the dead 6640, my conclusion is that it should be possible to rebuild it as a 6600, but nothing else- but that's better than chucking it out! Save it for a center channel project with my lone C79 spare?


As to the 6640 in waveguide, Well, next time I have a lot of extra moolah lying around for tweeter waveguide experiments, I'll have to give that a shot! :B :W Don't hold your breath, though!

[CBerg]

So mainly a lurker here, but seeing how bamboo seems to be coming into vogue as a "launch platform" thought I would post this link in that it seems like it would be a superior baffle. I have seen this material in person, it is quite nice.

http://www.crosscuthardwoods.com/Countertops/count_index.html#Bamboo

Not cheap, not crazy expensive either.

Lucky to live in the pacific northwest, bunch of hardwood type distributors/stores so easy access to appleply and a host of other interesting woods and woodproducts.

Regards,

[JonMarsh]

Nice link, thanks!

[JonMarsh]

Not so low budget speaker porn

Got lots to do this weekend, but will be pushing out the speaker work in spite of parts needed and received.



Instead, it's time to take care of the mod for ThomasW's Pioneer player, turn it into an SACD to LPCM over S/PDIF machine like mine.



Even bought some new 14" focal length magnifying visors...



Gonna look like some little Swiss watchmaker, I suppose! At least I'll be able to see better than last time. My clarity is just fine, the parts are just too small.

[Bear]

Well, that's an idea- the 66000 and 6640 have basically the same faceplate by appearances, the 6620-00 is quite a different beast. I do have a "dead" 6640 I can disassemble to look at. Wonder if I could mate the dome with a 66000 VC assembly? I mean, I'm good, but I'm not THAT good by any means! But for only $44.60, not much financial risk...

After taking apart the dead 6640, my conclusion is that it should be possible to rebuild it as a 6600, but nothing else- but that's better than chucking it out! Save it for a center channel project with my lone C79 spare?


As to the 6640 in waveguide, Well, next time I have a lot of extra moolah lying around for tweeter waveguide experiments, I'll have to give that a shot! :B :W Don't hold your breath, though!

Jon - Any chance the 6640 voice coil assembly will fit in the chassis of the 6600? I have a pair of 6600s I need to re-build (got them used and abused), and I'd be willing to put together a pair of hybrid 6640s if the voice coil was a drop-in replacement.

Bear

[JonMarsh]

I checked with Madisound, they (ScanSpeak) don't have the 6640 available as a VC/dome spare, and it specifically for legal reasons can't be used with a faceplate without the grille, due to the toxicity of Be. So, for now, they only sell complete replacement tweeters. Annoying, but understandable.

I will convert my left over 6640 chasis to a 66000 sometime soon. Maybe I'll try that with a waveguide...

[Only1MeBz]

Only1MeBz

I found this photo of the Black & Decker Table in search of a replacement leg...i registered just to ask you for any information you could give me on the part # or the place to find it...I am seeking to fix mine and would greatly appreciate your help. TY, Missy

[ThomasW]

Miissy,

Sorry but we're not psychic so we don't know what you're talking about when you don't post picture but refer to one somewhere in this 35 page thread......

One option since you mention Black & Decker is to use their parts website

http://www.blackanddecker.com/CustomerCenter/Parts.aspx

[numberoneoppa]

Holy shit, Jon, those capacitors are gigantic!

[JonMarsh]

I expect she's referring to the pictures of my foldup B-D workbenches with the aluminum tube legs. I expect the parts site you referred her to is the best bet, because I doubt any parts are available at retail.