Anarchy 6.5''

[LING GOWA]

Well it looks like the Anarchy 6.5 is on the shelves and again no off axis graphs. Looking at the supplied SPL graph it looks like another low cross tweeter will be needed. It looks like.

[cjd]

I have to model these up but I would probably run them with a nice little mid like the RS100... something along those lines. I think they could still deliver a TON in a small package, and that's something we don't see a lot of.

If I end up snagging a quad of the M21's for my living room IB (*cough* er... my wife only said "good luck taking it with you unless we sell the house to a single woman...") I'll have to add a few to the order - won't make a dent in the final cost...

[LING GOWA]

Zaph probably has a couple of these already.

[jkrutke]

Zaph probably has a couple of these already.

Nope, and I don't intend to test them unless Kevin Haskins specifically requested me to do so.

I saw an on-axis response curve, provided by Exodus audio I believe, that didn't look too bad. Off axis response plots for woofers are not needed to properly simulate their off axis performance.

[cjd]

I'm more curious how the suspension works re: distortion at higher excursion levels. Gobs of XMax only helps if the suspension doesn't make it unlistenable at the extremes.

[JonMarsh]

Something I'm curious about, and have run into in the past testing XBL2 drives, is that the basic linearity wasn't quite as good even at nominal levels and frequencies, even when you could push them to high levels before things "fall out of the gap" . Like having an amplifier that the first couple of watts isn't particularly clean, and it will play louder (with increasing distortion) than another amp that is truly low distortion at low to moderate power levels (not just one using lots of loop feedback).

The driver parameters are interesting, might work well in a small enclosure sealed with the flux capacitor- I mean really small. I'll have to try modeling it and see... someday.

Too much other stuff going on for now!

[cjd]

Crossing frequency requirements is about average for a driver that size I think. There's a bobble in impedance around 1k that makes me curious though.

C

[Jed]

Something I'm curious about, and have run into in the past testing XBL2 drives, is that the basic linearity wasn't quite as good even at nominal levels and frequencies, even when you could push them to high levels before things "fall out of the gap" . Like having an amplifier that the first couple of watts isn't particularly clean, and it will play louder (with increasing distortion) than another amp that is truly low distortion at low to moderate power levels (not just one using lots of loop feedback).

This has been my experience as well with XBL2 drivers. See ER18 (attachment on the right) versus the CSS SDX7 attached. While these drivers weren't tested the same day, they both were producing about 85-88db at 1 meter. Down low the ER18 seems to hold up better despite having less xmax.

I haven't tested the Anarchy driver yet, but I suspect similar trends. It would be interesting to find out for sure, however.

[cjd]

I will be working with these somewhat soon and will see what I can do to get more in-depth measurements done with them. May take me a few tries.

[Kevin Haskins]

I have all the Klippel data somewhere on my forum. I also have a pair sent off to Voice Coil and Vance will Klippel a pair in the Feb issue. The Klippel is about the most reliable metric when they are taken right. Distortion measurements are fine but I don't like publishing distortion measurements unless done in a chamber. I'm sending another set of finished loudspeakers using the Anarchy to Soundstage sometime this year and I'll have another reference measured in the NRC.

In terms of FR they are clean until that 4.2K breakup. There is a minor wobble around 800Hz but it is of the +/- 1dB variety. All of my FR measurements have been raw unsmoothed so there is nothing hidden in the smoothing of measurements. Also... I've shown some unsmoothed off-axis data in a finished design. As Zaph says though, there is no mystery in off-axis measurements for 180mm midwoofers. The dispersion narrows just like every other 180mm device and the break-up shifts slightly as you move off-axis. That is a good reason to keep breakup well damped in the mechanical design of the cone/surround.

These are raw data with the Seas DXT tweeter. I'm using a raw measurement but modeling the crossover in these simulations. The off-axis data is a measurement though with the same crossover used in each simulation.

This is the 20deg, no smoothing measured about 1.5M outdoors. I measure these on my forklift in the parking lot lifted about 8ft. The on-axis doesn't look as nice due to the baffle diffraction and I somehow lost some of the data so I only have the off-axis. I design around the slightly off-axis data because that way you don't try to correct for the baffle diffraction in the crossover.



Here is the 45deg, same particulars.



And about 70 deg.




Above axis with tweeter on top. This is the seat of the pants angle. Maybe 15-20 deg.



And below with the same seat of the pants protractor angle.




As I said, no smoothing used in any of these so what you see is what you get.

[Kevin Haskins]

This has been my experience as well with XBL2 drivers. See ER18 (attachment on the right) versus the CSS SDX7 attached. While these drivers weren't tested the same day, they both were producing about 85-88db at 1 meter. Down low the ER18 seems to hold up better despite having less xmax.

I haven't tested the Anarchy driver yet, but I suspect similar trends. It would be interesting to find out for sure, however.

The problem with a raw distortion measurement is it doesn't tell you anything about the mechanism. If the motor BL is flat then the distortion has to be coming from another mechanism. Or... there could be a problem with the build and you have non-flat BL. XBL^2 motors actually have more tolerance for being built off-center though because the BL curve is flat, and if the coil is off-center it only affects the distortion spectrum once you get out to the edge of the BL curve (high output). If you have a parabolic BL curve and the VC is assembled slightly off-center the BL is non-linear even at low output. It is very difficult to get consistent +/- 1mm placement of a coil in terms of centering it during assembly. Build a few drivers and even if you use the best jig and take extra care, they vary by a mm or so unit to unit. XBL^2 motors actually have an advantage over anything with a parabolic BL curve because of their relative intolerance to centering.

But at the end of the day XBL^2 is just a tool to achieve a goal. You cannot group drivers by one characteristic because they are a combination of suspension, Le & motor characteristics. It would be like grouping and characterizing all vehicles that use a small block Chevy engine.

[Jed]

But at the end of the day XBL^2 is just a tool to achieve a goal.

Do you have a harmonic distortion graph to share? I'm sure we'd all like to see it.

Edit: I don't mind waiting for the Voice Coil magazine--- I don't think they always test in an anechoic chamber though, and most on the board won't be able to see it, since VC magazine is for OEM. Irregardless, I look forward to seeing the results.

[Kevin Haskins]

Do you have a harmonic distortion graph to share? I'm sure we'd all like to see it.

Edit: I don't mind waiting for the Voice Coil magazine--- I don't think they always test in an anechoic chamber though, and most on the board won't be able to see it, since VC magazine is for OEM. Irregardless, I look forward to seeing the results.

No... I don't. We don't use distortion measurements in the design process only the Klippel for non-linear behavior. VC doesn't do distortion measurements in the review process either, just the Klippel. The Klippel tells us more about what is going on and allows us to make changes in the design to affect the behavior. A distortion measurement wouldn't tell us where and what is causing our problem so it is something done more for marketing than use in design of the driver.

[Jed]

VC doesn't do distortion measurements in the review process either, just the Klippel.

I don't mean to disagree with you but I'm looking at VC right now and they have non linear distortion plots listed via SoundCheck software for every driver tested, in addition to the Klippel results (Jan 2010).

By the way, Jon Marsh, if you are reading this they tested the SS6640 and the HD plots look really clean--- virtually no 3rd order distortion from 2k on up, pretty much what you got in your tests recently.

[Kevin Haskins]

I don't mean to disagree with you but I'm looking at VC right now and they have non linear distortion plots listed via SoundCheck software for every driver tested, in addition to the Klippel results (Jan 2010).

By the way, Jon Marsh, if you are reading this they tested the SS6640 and the HD plots look really clean--- virtually no 3rd order distortion from 2k on up, pretty much what you got in your tests recently.

Hmmm... I guess I shouldn't assume. Well... if they do distortion testing.... they will be available in another month or so along with the Klippel, FR, Z etc.. that they measure for the review. My subscription ran out six months ago and I could have swore at the time they only had Klippels, Z & FR in the driver evaluations. I'll clarify with Vance what they will or will not include in the evaluation so there isn't any ambiguity.

[cjd]

Spent a little time this morning sipping coffee and abusing these drivers.

To preface, this is 100% subjective observation with an hour of listening to them fresh out of the box (though they got a little play time while I fixed said coffee)

I'm running them propped up on blocks, sitting on the floor. Very scientific. A nice little chip-amp (probably ~40W into these?), sourced by a Squeezebox 3, passive attenuator. SPL was pushed as high as ~105dB peaks at 1M for the pair.

First things first: it does appear you can bottom these out (at least, I heard the sound of something that sure seemed like mechanical impact, and I didn't wait around to give it more careful attention but immediately backed off on the volume). It takes some doing. Say, some nice 16Hz pedal tones on a lovely organ recording and a wide-open volume control. A few folks here know the piece as they heard it at Ryan's place a couple weekends ago. By the time you're pushing them that hard they've lost almost all of their midrange control and you're left with bass and breakup. Back off on the output level a tad and as soon as things are within control it all gels.

I was, in fact, impressed by the level of control they maintain in the midrange while being fed heavy bass. Also very impressed by the complete lack of audible noise from the suspension even when pushed rediculously hard (though at that point I couldn't keep my ear RIGHT on them, too loud).

The breakup is audible and annoying, but far more listenable than the RS180 (the only other driver in this size I've given much listening time in this fashion). For those that don't find they need the breakup so suppressed, you can probably cross these very comfortably @2k 4th order, 1600 2nd order, with a gentle notch. Having no distortion profiles on these yet, I have no idea at this point if my subjective observation will be backed up, or my ears will be proven soft. :E I think my preference will be ~1600 4th order-ish.

I think these would be superb lower midrange drivers for dipole.

On the low low end, things sound clean and controlled, but it is notably more difficult to judge such things by ear - I suspect that suspension linearity is going to be the gotcha as these get pushed hard.

I think the biggest risk is that they do reach so very low, and it's just temping to crank the heck out of them. It's fun to watch them dance! That and their relative sensitivity is on the lower side.

Looking forward to working with these quite a bit.

C

[Saurav]

Are you planning on taking any measurements like that - free air, no baffle, or on a minimalist dipole baffle?

[cjd]

Possibly, though I do not think that the current setup is quite the right way to do such measurements. I might. I need a few cables still to get the new setup all together and will then need to work out the kinks if there are any (I'm moving to a laptop + M-Audio FastTrack Pro for this, I got tired of trying to use a desktop and needing loooong cables or a lot of misery moving monitors and computer boxes...)

[cjd]

Does this look right? 1/12 octave steps. First time doing this, there was a little environment noise (furnace, etc.) so I'm not sure if this is right or not.

Note that the graph top is -40dB.

[cjd]

Another shot, furnace shut off, 1/48 octave steps.



I think my setup has a couple glitches in it still.

[HareBrained]

I think, except for wanting the D3 to be decreasing, that's fairly impressive. +0/-3db covers from ~60 to ~1600Hz. And nothing nasty to deal with. Kevin may prefer to post that to his own. Was this IB or in the box?

[JonMarsh]

Your microphone playback level could be higher- that might help with clarity and dynamic range. In the -50 to -60 range is what I would hope for at 2.83 VRMS.

[cjd]

I couldn't find my spare attenuator so wasn't quite getting things where I wanted since I had to run the amp wide open. Still fiddling, not 100% sure I have things right. But I think the attenuator is the missing piece. I may need to adjust the voltage divider after the amp, or switch to pulling reference before the amp rather than after. Or maybe I need a different setting on the FastTrack input. I was close to clipping it. I'm not sure this was all the way up to 2.83V but it may have been. That was one thing that made me not quite sure I got this right.

Had I thought of it then I suppose I could have just gone without loopback.

This was in a leaky "sealed" box. Nearfield.

I did a couple sweeps from 10Hz up, it was fun to watch the driver.

C

[Dennis H]

I suspect the level thing is just an ARTA setup issue rather than actual low levels. Assuming you calibrated everything per the manual, that looks like you're doing a single channel measurement and didn't tell it you're using the mic so it's recording the mic output in dBV. If you do a single channel, tell it you're using the mic, and give it the mic's sensitivity in mV/Pa, it will display dB SPL. If you do a two channel with the mic, it will normalize to dB SPL at either 1V or 2.83V.

Setup - Audio devices
View - Sound pressure units

Edit: okay we were typing at the same time. If you were doing a loopback but didn't tell it you were using a mic (so it just measures voltage) it's saying the mic output is 40dB below the amp output.

[cjd]

I suspect the level thing is just an ARTA setup issue rather than actual low levels. Assuming you calibrated everything per the manual, that looks like you're doing a single channel measurement and didn't tell it you're using the mic so it's recording the mic output in dBV. If you do a single channel, tell it you're using the mic, and give it the mic's sensitivity in mV/Pa, it will display dB SPL. If you do a two channel with the mic, it will normalize to dB SPL at either 1V or 2.83V.

Setup - Audio devices
View - Sound pressure units

I was doing two channels but missed reading anything about what you've mentioned here, which setting to use so yeah, I'd bet that is the issue. I read through this on the fly, will have to spend more time reading up.

[cjd]

Hah. more typing at the same time. I'll go through the manual and redo setup again.

[Dennis H]

Yeah just start at the front of the STEPS manual. It's pretty easy with a DVM and just enter the mic's sensitivity (~8mV/Pa for the Behringers) rather than doing the cal with an SPL meter thing.

[cjd]

I had it at 5mV/Pa - must be the default?

When I ran live data through it was balanced, more gain on the mic and the sound card said it was clipping. Hmm.

[Dennis H]

I dunno. One problem with the M-audio USB boxes is finding a gain setting on the inputs that doesn't clip and that you can find again every time you use it. I think the pot goes from big cut to big gain. The EMU boxes are easier -- all the way to the left is unity gain.

[cjd]

The good(ish) news: I think the output on the driver is about where it needs to be - much louder and I would not have been able to go as low as 20Hz without picking up mechanical noise from the driver.

I think I had the actual separation much much lower at one point but the visuals (not the number readouts) were nowhere near balanced. I also think I may have had the loopback set to "instrument" - I'm not sure if it should be or not, but it looks like not. I also have NO IDEA whether "PAD" (-20dB) was set for either input (oops).

[Rick Craig]

Looks to be better than the Adire woofer which I found to be over-hyped as well as other XBL woofers that followed it.

[cjd]

I'm trying to be skeptical and judge that perhaps I've made too many mistakes here for this info to be right, but my ears are supportive of this being a pretty darned nice little driver. I'm still skeptical. There are a couple other things I now know I need to do slightly different, with the specific impact of those changes unknown till I get there.

I'll keep plugging away at this, as there are definite things not-yet-right in the measurement setup too. The danger is that I'm now thinking about how best to get an IB as well as starting a collection of boxes with removable baffles strictly for driver testing...

[Dennis H]

The danger is that I'm now thinking about how best to get an IB as well as starting a collection of boxes with removable baffles strictly for driver testing...

Who are you, stranger, and what have you done with C(I hate measurements)JD?

Edit: PS, if you haven't noticed, going to View - Percentage distortions gives a nice-looking pic.

[cjd]

I still hate measurements. But I know how very variable they can be, and doing things like providing distortion data can get those pretty pictures taken out of context *really* fast. Right now, I'm sure people could discount this data completely based on the test setup not being perfect.

Then too, getting T/S data needs boxes of known volume to do right.

The setup I have now will, I hope, be a LOT more stable to use since it's an external sound "card" with phantom power, and it should also be much much easier to set up and tear down, move to different places, etc. Much of what I hated was the need to set up and get through ALL the measurements in one go, and if it was being picky, it just added to the stress.

We'll see how ARTA is on FR measurements, but the software switch is also part of the equation, and it's in the price range I'm willing to buck for if it works right (Speaker Workshop, on the other hand... though it's amazingly similar in its needs and I can probably use it with the same cables I'm using for ARTA).

I also realized I failed to get parts to make an impedance cable - I might have everything already (I hope).

[---k---]

diVine Sound

Sounds like I missed a press release.

You aren't going to go get to expensive for me now, are you?

[cjd]

Sounds like I missed a press release.

You aren't going to go get to expensive for me now, are you?

I just thought it was punny and appropriate since I have decided to name most of my new speakers after wine grapes when a name is not otherwise forthcoming (I'll probably give the older projects a proper name now too)... some horrible bashing of bad Italian with English...

[Saurav]

You could make it one step worse and call it di(e-on-the-)Vine. That would match your 'I never get any projects finished' lament

[JonMarsh]

I just thought it was punny and appropriate since I have decided to name most of my new speakers after wine grapes when a name is not otherwise forthcoming (I'll probably give the older projects a proper name now too)... some horrible bashing of bad Italian with English...

I think this is just fabulous, Chris- a keen sense of marketing leavened with a little humor is something sorely lacking in much of the DIY market...

[---k---]

Uh-oh! http://divinesound.com/

Divine Sound and Entertainment is a highly professional Disc Jockey service with over 15 years experience in the Southern California area. All of the DSE Disc Jockey/Emcees deliver a very polished and professional presentation, using state-of-the-art CD and MiniDisc systems, providing the highest quality sound available.

I was getting attached to the names "Sietecerocero Project" "Ochocincos MTM" and the "Seisceros TM". Are we going to have to rename them to conform to this new fad you're undergoing?

[cjd]

I was getting attached to the names "Sietecerocero Project" "Ochocincos MTM" and the "Seisceros TM". Are we going to have to rename them to conform to this new fad you're undergoing?

I don't know what you're talking about... :uhoh:

[cjd]

Oh yeah...

Plus... all my designs sound soooo grape...

[JonMarsh]

I see a potential conflict coming up with the high class tone you're trying to set... let's say this Anarchy 6.5 works out, and they come out with some subsequent drivers based on the design concept- which you want to use, too- but a Maremino 2 using a "Son of Anarchy" driver, I dunno, there could be a bit of a culture clash...

[Saurav]

Sonarchiono... sounds like a grape varietal to me.

More to the point - how low do you think this would play on an OB, and how loud would it go? I can probably calculate the SPL from its Sd/Xmax/baffle width.

Edit: If I believe the 12mm Xmax, then SPL won't be a problem. So that really just leaves distortion?

[Rick Craig]

Sonarchiono... sounds like a grape varietal to me.

More to the point - how low do you think this would play on an OB, and how loud would it go? I can probably calculate the SPL from its Sd/Xmax/baffle width.

Edit: If I believe the 12mm Xmax, then SPL won't be a problem. So that really just leaves distortion?

The Klippel tests are really important for a driver like this because they'll show if the rated x-max is realistic. If it's suspension or motor limited then it may not be suited as well for an open baffle application.

[Saurav]

I should clarify that I'd use this as a midrange, not a woofer, so I only need it to go down to say 250-350Hz.

I remember reading something about a 3rd party Klippel test in an industry trade magazine - has that been published yet?