New speaker: SS6600, W15? and W26?

[harryharry]

I am planning to build a new speaker based on Seas W15CY, W26FX001 and Scan Speak AirCirC 6600. I have been using the SS 6600 and an old W17 together with a W26 for a long time but have never settled with this speaker. The main problem is the 3rd harmonic distortion of the W17 at about 1700 Hz.

But before I begin to design this speaker I have to consider alternative drivers.
I am very pleased about the SS6600 so this I think I will keep in the design.

For the midrange there are other alternatives. I have designs with Scan Speak Revelator driver but I prefer the Seas Excel for critical listening. The Seas Excel has a very analytic sound, and some would say a not so engaging sound. I have been considering using a PRO driver like the PHL 1130/1120. It does not measure as well but I have heard that it is still more engaging than Excel.

For the woofer the W26 measure well but I have never gotten it to work well below 200 Hz without eq. (This is probably room related) Below 50 Hz I use a infinite baffle consisting of 2 18" FI IB3 in this room. I have considered using maybe a 15" pro drive from Beyma or Ciare something in the line of Mezzo Galactica

[5th element]

You should be able to push the 6600 down to around 1200hz with a steep xover as long as you're not expecting to hit high spls. By the sound of things you've got a large room so this might be a problem, but it would let you hear the W17 with the 3rd order peak somewhat attenuated.

Maybe try this first before spending lots of money on drivers?

The W15 does sound nice, I use them as mids and I'm very pleased with them. But there is the Zaph Audio ZA14 to consider, perhaps the tiny Scan speak 10F-4424G00 and maybe the 18Sound 6ND430-16.

The Aurum Cantus AC130/50CK also measures extremely well in the harmonic distortion department, but might need a low crossover. If you can control the FR irregularities the only thing to really watch out for is a tiny spike in 2nd harmonic at a tad over 2khz.

The ZA driver is ridiculously cheap for the measured performance that it brings though, so it might be worth buying a pair just to experiment with. This is of course if you're in the US.

[harryharry]

The speaker (SS6600, W17, W26) does not sound bad at all. It is today my favorite speaker (i own several DIY speaker models). I have changed and optimized the cross over probably more than ten times during the last years. However even with steep slopes and low cross over point I miss something (or maybe just want to make something better :T ). I have realy never liked to push the tweeter to much. I feel the tweeter then quickly starts to "shout"

I will use the top cabinets of this old speaker with the W17 and a new Seas DXT to make it into a near field monitor so I will use the W17 and the cabinet in another design. (No bad feelings!)

I would realy much like to try the ZA14 but I live in Norway and it is quite expensive to get the drivers sent from Madisound.

If no other suggestion I will test the W15.

I am thinking of a crossover point at about 2 kHz between the W15 and the SS6600 and a crossover point at 300 Hz 2nd order between w15 and w26.

Has anyone any built speakers with both the w26 and w15. I know Jed "Tombstones" used these drivers. However this was in a OB MTMW design. I am thinking of a TMW design with closed/ported cabinets.

[5th element]

I am using the W15 at 2khz also with 4th order slopes and I don't have anything to complain about. I do want to try out lower xovers and possibly 2nd order at 1500hz.

I hear you about pushing tweeters too far, it's not something I am too fond of either, but I think many people are somewhat over cautious with this regard.

I am also using the DXT with the W15s and absolutely love the way it sounds, nothing whatsoever to complain about. And with its ability to handle low xover with ease, it should be a fine match with the W17s. Although I can't help but think you might prefer it over the 6600 with the W15s

[Face]

Here's some additional info on the 6600: http://www.htguide.com/forum/showthr...highlight=6600

As for below 200hz, it's definitely your room.

[harryharry]

Yes, I am quite sure it is the room. I have solved it with computer based room correction. I am running uncompressed music through J.River Mediacenter with convolver VST plugin. The Fir filters have been made with Audiolense. This solves the problem and also cleans up the soundstage. I use a Lynx Two B sound card and a TwistedPear Buffalo II DAC.

Face, looking at your signature your have designed a speaker (The Regals) where you use w12 instead of w15. Why did you choose this driver? I use the cheaper sibling L12 in another design and like this driver a lot. Have you made the crossover drawings available for this speaker? I see that you also use the expensive c79 in another design. How does these speakers compare and which one do you like best?

[5th element]

I think the main trouble with the W12 is that the break-up node isn't much higher then the W15. In a three way both can be used as low as each other (under normal circumstances) and the W15 has higher sensitivity which is always a good thing.

Naturally the W12 will beam less, but at the xover frequencies you'd use with the metal cone driver, this doesn't really help in anyway.

Then if you ever wanted to make a 2 way the W15 would win again.

I guess this is why the W12 isn't used that often. BUT if you're after a dedicated midrange, in a passive three way system, where the added sensitivity is only going to waste - then there's no real reason to go for the W15 either. Of course I don't think I've ever seen third party measurements of the W12?

The accuton drivers are interesting, I'd love to try the C90-6-079 and the C173-6-090, but the prices are a little bit out of the this world! I mean the performance is nice, but I can't seem to justify the cost in any way.

[Face]

Face, looking at your signature your have designed a speaker (The Regals) where you use w12 instead of w15. Why did you choose this driver? I use the cheaper sibling L12 in another design and like this driver a lot. Have you made the crossover drawings available for this speaker? I see that you also use the expensive c79 in another design. How does these speakers compare and which one do you like best?

I originally planned on using the Nextel version, but it had a dip from 1-2k that I had a hard time correcting. The Mag version has a peak there which measured flat in my cabinets. As for the W12 vs W15, I figured the W12 would make more of a point source. I haven't made the crossover available for it yet as I'd like to revisit one more time when I'm done with my 6600/C79/W22NY project. I'd say it's 99% done though at twenty something pieces per cabinet. 8O If you decide to go forward with these drivers I can share notes though.

As for the C79 vs W12, I can't make a comparison yet since I'm still in the cabinet building stage of the C79 speaker.

What 5th element mentioned above is correct about the added sensitivity of the W15 not being needed as with full baffle step even the W12 is still padded down.

FYI, I'm crossing at 325 and 2500. 3rd at 325, much steeper at 2500.

For some feedback on the design: http://techtalk.parts-express.com/sh...ghlight=regals They were referred to as the Seas/Scan speakers, Excel Monsters, and Regals.

[harryharry]

Face, it seems like the people visiting DIY NY 2010 really liked your speaker. I have more or less desided to go for a W15. Then I have the option of using this driver in a small book shelf speaker if I do not like it. I did some simulations in LSPCAD with the W26 today. It is now sitting in a cabinet with a volume of about 98 l with a port tuned to about 22 Hz if I remembered correct. With a 2nd order slope at about 300 Hz the breakup at 4 kHz would be about -35 dB down. I think you are doing right going 3rd. Another alternative would be to make 2nd order with a notch filter. Since I use an active crossover run through a computer it is quick and easy to test and listen to different topologies. I will do some test before I deside on the ideal crossover point and steepness but 325 Hz and 3rd order sounds reasonable. (I hope I will manage to get away with 2nd order.)

Here are some pictures of the speaker I am replacing. The cabinets were built 15 years ago and I have used several drivers in the same cabinets only modifying port lengths.


and the temporary crossover:

[Face]

FYI, even with a 3rd order slope on the W26, I still used a notch.

[harryharry]

I am very interested in looking at your crossover design when it is tweaked ready. These drivers are time consuming to design crossovers for and the drivers are not for beginners.

Do you use a LCR circut for the SS6600?

[Jed]

Face's crossover won't do you much good unless you use the exact same box and drivers.

It's a good idea to use an LCR impedance circuit for the SS6600. It has a fairly pronounced Fs hump that shows up in the acoustic transfer function when crossing low (below 2k). I've also found the phase integration is much better with the LCR impedance correction in place.

[harryharry]

I agree that it is not possible to copy the crossover directly. However using an excisting crossover which works nicely can be an usefull reference when I am planning to use similar drivers (W26, W15 and SS6600).

[Face]

I am very interested in looking at your crossover design when it is tweaked ready. These drivers are time consuming to design crossovers for and the drivers are not for beginners.

Not a problem.

Do you use a LCR circut for the SS6600?

No, just a zobel.

Face's crossover won't do you much good unless you use the exact same box and drivers.

It's a good idea to use an LCR impedance circuit for the SS6600. It has a fairly pronounced Fs hump that shows up in the acoustic transfer function when crossing low (below 2k). I've also found the phase integration is much better with the LCR impedance correction in place.

Even crossing above 2K I noticed the same thing, but as able to correct it with only a zobel.

[harryharry]

Face, what volume size have you chosen for the cabinets and to which frequency is the port tuned to (if any for the W26)?

[Jed]

Not a problem.


No, just a zobel.

Even crossing above 2K I noticed the same thing, but as able to correct it with only a zobel.

LCR circuit impedance correction at Fs is a "zobel."

[5th element]

I thought a zobel was traditionally an RC circuit designed to combat rising impedance, not the peak in impedance due to resonance.

[Johnloudb]

I thought a zobel was traditionally an RC circuit designed to combat rising impedance, not the peak in impedance due to resonance.

Speakers have inductance, and this is taken into account in calculating the R, C values. So the circuit as a whole is an RLC network and maybe as a whole is called a "zobel."

So, I think Jed's correct, and I've heard the term zobel used for RLC circuits used for correcting response peaks in speakers.

That's my novice input ...

[5th element]

The R - L - Cs are describing what the electrical filter is built up of and has nothing to do with the parameters of the loudspeaker itself.

The loudspeaker voice coil has a natural R and L parameter. R is modified (increased) as frequency rises due the the reactive component L.

The zobel network again has a value R that is modified by the reactive component C, such that it's impedance rises as frequency decreases.

The component values are chosen such that the two balance each other out and thus, it removes the impedance rise created by the reactive voice coil.

This is what I've always read that a zobel network was.

Of course this doesn't mean the term cannot be applied to an RLC resonance trap filter, but I cannot remember ever seeing it done.

[Johnloudb]

Okay, well like I said, that was my novice input.

Of course this doesn't mean the term cannot be applied to an RLC resonance trap filter, but I cannot remember ever seeing it done.

Yeah, I guess one could say, "Zobel is, as zobel does" or not. That is the question.

[Jed]

Of course this doesn't mean the term cannot be applied to an RLC resonance trap filter, but I cannot remember ever seeing it done.

It's a term used quite often here at htguide.com. Here's an example:

It all depends on the span and nature of the zobel network. A zobel for tweeter Fs resonance to take the bump off the roll of curve is no big deal- little power will ever be sent there. A zobel for controlling woofer impedance at a the upper port impedance peak in order to prevent interaction with an LF crossover is a whole different critter! LspCAD can give some guidance on current in components, as I recall- you'll note that in the Ardents, the LF zobel uses a 50 watt resistor. That's not just because I like gold anodized aluminum....

~Jon

I guess I can see both sides for using the term to describe a resonance trap circuit that flattens the impedance Fs of a tweeter or bump in impedance from transmission lines or bass reflex loading, versus the typical application of the term for impedance leveling the rise at high frequencies.

Anyway, we digress, and for that I apologize. The 6600 has a rather pronounced Fs peak that I've found can be tamed with the RLC trap, "zobel", etc. Maybe not too necessary if crossing above 2k, but I wouldn't take the W15CY that high given you are using a tweeter that can cross relatively low.

Jed

[benchtester]

...The 6600 has a rather pronounced Fs peak that I've found can be tamed with the RLC trap, "zobel", etc. ....

Jed

For those wishing to learn more about this practice, an interesting description is in Zaph's ZD5 project page:

http://www.zaphaudio.com/ZD5.html

The XT25 is has a very similar impedance peak compared to the 6600 (partly because they both don't use ferrofluid). I made a pair of speakers which were like the ZD5, but I used the 6600 instead of the XT25. I don't think I changed any component values in the LRC across the tweeter; my drivers were that close in Fs.

Summary edit: With Jed and Zaph using an LRC to flatten/dampen the tweeter at Fs, it's pretty compelling. Now back to the regularly scheduled programming...