Anyone built the Ekta Grande by Troels?

[JonMarsh]

You can use NP electrolytics- Tony does for the LF Zobels. I'm looking around to see if I can find something better than garden variety ones.

Keep in mind this is a "proof of concept" first pass done with test box data. Component values can and probably will change. In comparison, 200 uF NP Electrolytics are $3-4 each. You can get 1000 uF NP, but I wouldn't use a single cap because of the ESR and dissipation factor. Paralleling is better.

Hate to say it, but you know, sometimes there's a reason some commercial speakers cost $20K a pair. Though I will do the design for the RS drivers, I don't really think anyone's likely to go there- considering what you need to spend elsewhere. As Steen Duelund said, you should expect to spend as much or more on the crossover filter as for the drivers. Most DIY people don't think that way, obviously- often its about saving cost. For a simple two way, that can work out pretty well. For a three way with a wide band midrange, things may not be so simple.

And of course, for a tube amp, this would be far more benign to drive. :W

[JonMarsh]

If you're going to be use all film caps and air core inductors, I take it the crossover will be external?

Who said I'm using all air core inductors? :W I didn't.

At this time I'm not planning on putting the crossover external- I've reserved a lot of space in the bottom of the cabinet. Of course, by the time it's electrically designed, and I have the parts in hand, I may feel otherwise!


BTW, here's the impedance for the C79 that the crossover network sees after the affect of the zobel LCR and RC:






The actual midrange crossover is quite simple because of the combination of acoustical and electrical filters, AND the fact that there's no big impedance bump at the LF resonance of the driver.

The Q of the resonance one is compensating determines the spread of L and C needed to tune correctly, and their values.

Now this was just sealed test box data- not the ported Ardent enclosure- so values will change for this network. Just sit tight and relax... with any luck, they won't get any worse. :B

One other reminder- in an LCR zobel, DCR of the inductor is unimportant and can be quite high- so the coils are not as big as you might expect.

[Face]

Who said I'm using all air core inductors? :W I didn't.

I thought you mentioned Goertz inductors a while back due to their low DCR...

At this time I'm not planning on putting the crossover external- I've reserved a lot of space in the bottom of the cabinet. Of course, by the time it's electrically designed, and I have the parts in hand, I may feel otherwise!

Image not available

Goertz and Jantzen foil inductors are both likely, but they're not the be-all for low DCR in values at 2 mH or above. I'll explain further as we go along, and in the proper thread- but a lot of that I was planning on leaving to the web site that is in construction.

[Curly Woods]

You can use NP electrolytics- Tony does for the LF Zobels. I'm looking around to see if I can find something better than garden variety ones.

Keep in mind this is a "proof of concept" first pass done with test box data. Component values can and probably will change. In comparison, 200 uF NP Electrolytics are $3-4 each. You can get 1000 uF NP, but I wouldn't use a single cap because of the ESR and dissipation factor. Paralleling is better.

Hate to say it, but you know, sometimes there's a reason some commercial speakers cost $20K a pair. Though I will do the design for the RS drivers, I don't really think anyone's likely to go there- considering what you need to spend elsewhere. As Steen Duelund said, you should expect to spend as much or more on the crossover filter as for the drivers. Most DIY people don't think that way, obviously- often its about saving cost. For a simple two way, that can work out pretty well. For a three way with a wide band midrange, things may not be so simple.

And of course, for a tube amp, this would be far more benign to drive. :W

John,

You haven't lost me yet ;-) It was just that I am not used to the high values of capacitors in the XO. I trust your knowledge and understanding about what it takes to make great music, which in the end is what I am after.
Hi-fi I can buy at the local audio store.

I have to be honest though. The Avalon that I have dealt with in the past (I sold them when I was selling audio) were always a bit too sterile for my taste. Incredibly open, detailed, all of the audiophile adjectives, but a lot like a well refined Thiel (which we also sold). With great records or discs, the sound was impressive, but I did not feel the soul of the music. This is part of the reason that I love tubes, as they bring back much of this feeling for me.

I do not want bloat, or tizzy, etc. Just revealing to the max, but not sterile. I maybe looking for nirvana, but every now and then I come up a system that speaks to my soul.

You like one of the few SS electronics firms that I feel has a real clue (Ayre). So I have very high hopes for these speakers :-) Anyone that shoots for the moon, has my full support!

[JonMarsh]

What kind of electronics did you listen to Avalon's with? And which models? There have been some changes over time in the "house" voicing. I can easily imagine system configurations with, say, Eidolon Diamonds, that would include well regarded components and wind up the way you describe. Source material is quite critical- or should I say source units? Even when you're going "budget" digital, there's a huge difference between the presentation of, say, a Benchmark DAC and a PS Audio DL III. The latter sounds much more like live music, even with just adaptive USB input from a Mac. Sterile and "studio" is how I would describe the Benchmark and many other digital sources.

In general, for Avalon's, the best sound I heard has been with Ayre or BAT electronics, Cardas cables (Golden reference interconnect and speaker), and Ayre digital or the Cary SACD player.

I may well be full of it, but I do expect this crossover to improve on what Avalon did with the Indra. Talk is cheap, but I'm not. :W

Hey, what are you doing during RMAF in October? Come up to Denver then or a few days before and you should be able to listen to these- I'll be posting more about that as it gets close. I'll be bringing up a complete system to demo at ThomasW's. It should be interesting, and fun, too!

In the meantime, ThomasW is probably going to spank my butt for taking this thread so far off topic- it's not as if I shouldn't no better, being the other moderator! :roll:

[Curly Woods]

Levinson, Krell and Classe w/ full Cardas and Wadia front ends mostly John. That was over 8 years ago now that I think about it. Things do evolve.

I have about played out on my available vacation this year due to family sickness. I really want to make it to one of the RMAF sometime. I really miss going to Vegas for CES every year. I think that RMAF will over take Vegas as it is definitely more high end oriented.

Oh I forgot that we sold BAT too :-) I really liked these the best (tubes again :-)) My memory is fading as the years go by.

[JonMarsh]

Well, that's well regarded gear, but the Levinson and Krell I wouldn't particularly recommend with Avalon- Wadia front ends were pretty good for their time, in comparison to what was out then, but fortunately things have improved. Otherwise I'd probably be going back to vinyl, too. And Ayre is even importing a turntable from Germany these days, and making a phono preamp!

I bet you'd enjoy RMAF if you can get a chance some year to come out- I didn't make it last year, was laid up on a business trip earlier in September while I was in Denver, but this year it looks like we have an all clear for lift off from Houston.

I didn't recognize McKinney as a suburb of Dallas, but checking Google reveals all- funny, I was just in Dallas this week on business travel! It's hot enough there, but then it's been in the high 90's here this week, too!

[dlneubec]

In those zobels, low cost films such as Dayton or Cross cap could be used; or even NP electrolytics, though I tend to not favor the latter because of the dissipation factor, and because electrolytics dry out and change value over time.

These LCR zobel networks are quite necessary if you want to wind up with a reasonable impedance; otherwise you'll find the reactances necessary to get the required corner frequencies will result in 2 ohm speakers in much of the mid bass and bass. Hmmm, something about that is familiar! Like the Avalon Indra, for example! Or that other project, the Mark 21!

The DCR for inductors is important; it impacts roll off behavior when used as a shunt and insertion loss when used in series; as in other designs I've posted, I'll post recommendations. For example, L1 will probably be a Jantzen C-Core toroidal inductor, as the DCR is quite low, and this is important for the woofer damping and loss.

Jon,

I notice that you seem to favor more zobels than most designers in some of your crossover designs. I've only been at this for a couple years and looking to learn, so I'm wondering if you would expound on your philisophy a bit here. What are your main goals in doing so and where do you draw the line at what is overkill and what is just enough? Are you mainly trying to raise the impedance or is it an effort to flatten the impedance/impedance-phase as well?

Also, can you explain the "dissipation factor" in regards to NP electrolytics?

[JonMarsh]

Jon,

I notice that you seem to favor more zobels than most designers in some of your crossover designs. I've only been at this for a couple years and looking to learn, so I'm wondering if you would expound on your philisophy a bit here. What are your main goals in doing so and where do you draw the line at what is overkill and what is just enough? Are you mainly trying to raise the impedance or is it an effort to flatten the impedance/impedance-phase as well?

Also, can you explain the "dissipation factor" in regards to NP electrolytics?

Sometimes zobels are a "nice to have" (like putting in a parallel LCR network to kill the impedance rise in the crossover region between the midrange and tweeter, so that the amplitude response with tube amps is flatter); other times, an RC inductance rise compensation or LCR resonance combination is nearly mandatory to get the target acoustical response with reasonable values in the crossover, or any value!

Imagine rolling off the bottom of a midrange unit, wanting it to track accurately down to -20 dB, but the bump in the impedance curve at resonance will put a big bump in the amplitude curve of the midrange also. The same thing will happen with tweeters that have their LF resonance near the crossover frequency. That is why you see me using LCR zobels more than most- this eliminates that as a factor.

Of course, in many commercial speakers, they avoid this because it adds cost. And many designers look only at the overall on axis SPL curve, and if it seems smooth enough (especially with 1/3 octave smoothing on the data) don't sweat the small stuff or how the curve looks in the transition band. But with lower order crossovers, I think this is short sighted. Just my 0.02. I'm a bit of a perfectionist in many areas.

Other designers may tell you it's just not necessary- in some cases, it may not be. Your call, of course, when you're doing your own designs. But particularly for the Duelund style crossover being used in the Ardent and used previously in the NeoD CC, this attention to detail is necessary to achieve the overall amplitude and phase tracking targets. Steen always said to plan to spend more money on the crossover than the drivers. That runs contrary to what many DIY'ers would like to do, in order to maximize bang for the buck- I'm just out to maximize bang.

[cjd]

Play with flattening the impedance - especially on a tweeter you're crossing low, it's audible (IMHO). I got a lot of grief for doing this in the big 3-ways...

[dlneubec]

Jon,

You know what I think would be a very interesting, valuable and educational thread? If you were to take us step by step through your crossover design process as you create it for one of your projects. What your goals are, what crossover points and slope targets you are shooting for and why, what each addition to the design does, why you used it and what other options might accomplish a similar result, etc. You might illustrate before and after simulated response, Z, phase, etc. with each step.

Obviously it would be just one designers approach, but certainly a very experienced one. :T

[JonMarsh]

Well, when I put up the Ardent project web site, that should more or less be included- in case it's useful for folks. Thing is, there's a lot of ways to get from point A to B, and I would never suggest that I've figured out the best one or one that others should emulate- maybe not even an easy one! But I will include that discussion. The web site creation began a couple of months ago, but I've been too busy making sawdust instead of pushing pixels to make much progress- it may not be up until November sometime, between RMAF and business travel to Europe in October.

[JonMarsh]

Jon,


Also, can you explain the "dissipation factor" in regards to NP electrolytics?

Think I forgot this part for you- "Dissipation factor" in regards to NP electrolytics (or any cap) is how "lossy" the capacitor is due to internal resistance, specified in percentage. Basically, it's 1/Q of the capacitor, where "Q" is the quality factor; a "high Q" capacitor is relatively losses, because the equivalent series resistance (ESR) which creates losses) is quite low. Good film caps have DF of 0.1% or 0.2%. NP Electrolytics at Erse, for example, are rated at 3%, 6%, and 10%. These are all fairly unacceptable to me, especially the higher values. It means also that with significant current flowing through the cap, there will be voltage drops and losses. It also means that to model the cap correctly in a crossover, you need to include an explicit R component; this can be determined by measurement in a good LCR meter or network analyzer. Also, part of the element degrading Q may be series inductance, so this should be determined, too.

[JonMarsh]

Electronics and voicing and zobels...

Today I did a lot of swapping electronics out back and forth, particularly with regards to higher output impedance gear and more conventional amplifiers.

It reinforced my belief that zobel networks to control impedance rise in a speaker system is a good thing, especially if you want the speaker to sound similar or mostly the same with different sounding electronics. You can look at the impedance curve of the NeoD CC and predict where it will sound "warmer" or "more forward" with the Ayre AX-7e integrated amplifier, compared with more conventional solid state amplifiers I have on hand. To the degree that this interaction can be ameliorated or mitigated, I think it's a good idea- obviously constructors with strictly conventional gear and more limited budgets can leave those parts off if they choose- I expect even more conventional amplifiers may sound better with a more consistent load, but that could be "wishful thinking" on my part. Certainly the dramatic impedance swings that can occur with some crossovers in the crossover region should be avoided when possible.