Time for some Ansonica - a new design for Jagman's theater

[cjd]

Will get a chamfer for sure - without it there is a notable dip in the tweeter response - even with, it will be there, but almost 2db less dip.

Wish eliminating the rest of the dips in my life were so easy.

[evilskillit]

Wish eliminating the rest of the dips in my life were so easy.

You could try applying the router to your other dips. Tho the authorities might not be as amused as that thought makes me.

[numberoneoppa]

Yes, it goes low.

Yes, it's easy to overdrive without a rumble filter if you feed it stuff <20Hz.

wheeeee. :B

Awwwe, all of a sudden, they look like such baby speakers. =)

[---k---]

Baby speakers!

[evilskillit]

Small speakers suck! Fortunately they're not small, they're just babies. hehehe

[---k---]

Actually, they are bigger than they appear. That is a pair of 6.5" woofers. It is just anything next to CJD's big 3-ways looks small.

[evilskillit]

Actually, they are bigger than they appear. That is a pair of 6.5" woofers. It is just anything next to CJD's big 3-ways looks small.

Yeah if they're anywhere near as big as my Zaph WG TMMs then they're a lot bigger than they look. Mine aren't huge but when I finished em I thought "Man those are a lot chunkier than I was thinking."

[cjd]

Quandary time.

I have some first-pass measurements, and have worked up a couple tentative crossover options.

One is a series-parallel 2.5 way, the other is a pure parallel 2.5-way.

Response is almost exactly the same with the series looking slightly better at dealing with the diffraction dip in the tweeter response.

Series is definitely more costly as it uses larger value components - though, only ~$25 more each, give or take... That's using all 14ga Erse inductors... The question is, which should I build and optimize?

I will not be doing both...

At the moment I'm leaning toward the series crossover, but am not sure if that's more complexity than is worth it (who knows who all will end up building these in the end...)

FWIW: Slopes ended up approximately: 2nd order Bessel on tweeter at 1800Hz, 4th order Butterworth on mid-woofers, at 1800 and 1400. Null is at ~1550Hz.

[cjd]

Update: Moving back to 16ga and 18ga in a couple spots where the impact is minimal, the cost difference is not insubstantial.

[---k---]

My opinion isn't worth much in this area. But, I would love to see the layouts, response graphs, and impedance. I'm most interested in understanding how a series crossover could deal with the diffraction dip better.

[cjd]

I could probably get the parallel network to work as well with some extra time... the difference is quite subtle, really.

I went ahead and ordered parts for the series crossover, will run ~$200 I believe - that includes an optional inductor (it will clean up the super critical range ~1.5kHz but only a very little bit). And parts to smooth impedance so it's essentially a flat 4ohm past the woofer impedance peak(s), zero phase shift. I'll play with using that and not using that.

I had another option using steeper filters and much worse impedance (both phase and overall net low) that was even flatter, but I think it won't sound as good in all probability. We'll start with this, see how it goes.

C

[bbcmp1979]

Impressive and Beautiful work!!!... ;x(

[cjd]

Black is the series summed response, yellow is the parallel summed response.

Red/blue/pink are series individual driver response(s).

The dip at 3k is diffraction induced. This is measured without the chamfer, so that should get better.



This is with a LRC to flatten impedance, otherwise it peaks up to about 12ohm in the 1200Hz-ish range, phase gets up to ~+20deg and -30deg at 400Hz and 2200Hz respectively. So rather benign either way. Solidly above 4ohm all the way.

[cjd]

Will update this image soon as I get my notes on port length.

edit:

2.5" diameter port should be ~12-13 inches long.
3" diameter port should be ~18" long
1x7.5" slot port should be ~18.5" long (equivalent to ~3.1" diameter)

That's targeting ~24Hz tuning.

[---k---]

Wow. I don't know if the difference between the series and parallel even qualifies as "slight". I had to squint to see the difference.

I saw that same diffraction dip in your Khanspire measurements. It originally had me worried. But, I don't hear anything. I'm guessing that it evens out when combined with the off-axis measurements???

[cjd]

That too, probably. I didn't do all the measurements here because the box isn't finished, this is to get me something close to work from.

[cjd]

Well... box2 is under assembly but I need to do a little trimming on braces for the next step, and I won't get there till Friday.

The first box top has started to split along glue-lines part of the original plywood - I suspect the top brace is a little wider at the back than the actual top, though I checked that a few times over so... not sure. Anyhoo, now I have to figure out how I want to solve this.

I know for sure that I do not really want to try to recompress things - if the brace is in fact wider, it will just keep pulling apart at the weakest link. I have to open it back up and verify my theory first...

I'm not sure if I want to thin down PVE and let it soak in with a final line of thicker PVE + fine sawdust, or if I want to move to a urethane glue here. I need to keep in mind that I want full dye coverage, hopefully... though I'll do a penetration test and if it just won't happen, I'll stick with plain birch and just clearcoat it.

While I'm at it I have decided I need to trim another 1mm from all the woofer recesses (diameter) so will be working out some jig for that, will probably use the little Bosch trim router for that (awesome little piece of equipment).

In other news - I had ordered some crossover parts from PE on Sunday, and realized Monday I'd forgotten to order any Sonic Barrier, so I ordered that... Well, the Monday order is evidently sitting on my doorstep today, and the Sunday order isn't slated for arrival till Friday! Go figure that one out.

Parts from Erse are also at my doorstep, and I would not be surprised if the stuff from Madisound is also there. I may be listening to the first crossover attempt this weekend! If I can get the second box together the rest of the way, that is. I think that will be do-able.

I also need to work on some other speakers someone schlepped over to my place a little bit ago...

C

[cjd]

Also - I will probably continue to tweak both the series and the parallel crossover variants. Their tracking is SO close it pretty much means I'm hitting target transfer functions with both quite closely, and should be able to fiddle with one and then transfer the learnings to the other and have a completely reasonable crossover that reflects my tweaks without actually building both options.

Though there sure is a part of me that wants to build both, put 'em on some nice ladder switch, and blind A/B test to see just what difference there may (or may not) be.

[---k---]

I also need to work on some other speakers someone schlepped over to my place a little bit ago...

C

That sucks that you're getting some splitting. Hopefully you can get it fixed easily and save the ability to dye them.

[evilskillit]

The first box top has started to split along glue-lines part of the original plywood - I suspect the top brace is a little wider at the back than the actual top, though I checked that a few times over so... not sure. Anyhoo, now I have to figure out how I want to solve this.

Have you tried swearing at it? I'm not sure if it will actually help, but you'll feel a bit better. :B

[numberoneoppa]

Fill it with epoxy and clear coat the entire thing with epoxy.

[cjd]

No need to use epoxy, if I need a thinnable glue that will fill the void and be clear-coatable I can use any urethane glue (I prefer Elmer's over Gorilla...)

Oh yeah. Both boxes from PE showed up today, so I've got parts!

[cjd]

Breadboarded, stereo pair currently playing...

And needing work.

Bass response is superb - no surprises there, right?

Midrange is still off. Way too forward and shouty, yet seems like they're clammed up.

Tweeters appear to be bottoming out which suggests that's where the problem lies...

Wheeee.

Got a couple more pics as soon as I get my other computer fired up, but that won't be today. It's sleepy time.

[cjd]

Ok, 5th time checking the wiring is the charm... Actually, it was trying to figure out what I was hearing and looking at the sim that I finally realized what I'd forgotten to do.

OOPS! Note to self. Be sure to wire the woofer>tweeter series connection to the rest of the "middle leg" on the crossover... sigh.

There are some things simpler about using parallel crossovers.

[cjd]

Ahhhhh yes. Much more like it.

Need more listening time, some tweaking still I think.

OMG do they deliver bass. Just how much bass (and how low it goes) still seems to be the limit on what they can deliver SPL-wise now that the tweeter actually has a filter on it (gee, I wonder why sending it full range was causing it to bottom out!)

[aduljr]

Ahhhhh yes. Much more like it.

Need more listening time, some tweaking still I think.

OMG do they deliver bass. Just how much bass (and how low it goes) still seems to be the limit on what they can deliver SPL-wise now that the tweeter actually has a filter on it (gee, I wonder why sending it full range was causing it to bottom out!)

glad to see a project on this speaker. I be interested in doing the surrounds when you are done with those. I prob wont have time to build them until end of the year

[---k---]

Very cool. I'm looking forward to hearing them.

[cjd]

Progress... I posted pics sooner than I expected!

[---k---]

Looking good. How about a side profile.

[cjd]

Left them playing yesterday while I was out at a memorial service for my grandpa which had me out for most of the day (left at noon, got home ~10:30). Walked in the door late after a long drive. They sound good - especially at low volume. So I think I'm looking at needing to find some compromise for FM curve issues for higher output levels because they start to get a little harsh as the volume is cranked - on the other hand, it's easy to crank it up and not realize how loud it is so it can't be all bad. I've been working on the SCC project so this is just sitting and making music (and hopefully NOT having cats decide to play with twisted-together-crossovers that are sitting on the floor...)

[cjd]

Definitely were a little shouty, so today I swapped some parts around, things are sounding improved. Right now I'm vaguely annoying my wife by flipping through tracks testing things out.

Too many projects going right now though I'm afraid.

[---k---]

Too many projects going right now though I'm afraid.

Yeah, but you're having fun, right?

[JonMarsh]

Looks to me like Chris IS having fun, perhaps his wife is jealous? :W

Those surrounds do look pretty meaty. I should be trying the alfresco crossover construction for prototype instead of having to modify fully built boards... some day- I guess I'm a slow learner.

Of course, you know what I'm going to say- you've got the makings of a pretty nice three way if you'd just put a good midrange in there somewhere! :B

[cjd]

Yeah, these really would be superb in a 3-way. They're darned nice in a 2.5 way, really good for some grins when you put on something with low bass and it starts loading the room... And then you crank it up a little more and run out of excursion. Ahh, limitations.

What these really scream for is a 12ohm version to allow you to go with 3 Anarchy drivers and net 4ohm. Or 6 and 8ohm...

I need to re-measure now that the boxes are complete too.

[cjd]

Updated post with port dimensions in case someone wants to build the box, now you can.

Also, though not 100% final this is very very listenable. I am working on a parallel crossover version as well, but again - if someone for whatever reasons is just having to build these NOW... here you go.

I have not yet re-measured these with the chamfer either, so this may yet be tweaked some. This response data is not 100% accurate as a result.

Note that the RLC is optional - that's C6, L6, R4. Impedance with and without is included.

C5 can be pushed up to 15uF which improves modeled phase integration, will verify exactly how that measures when I get there. Bumping it up, however, introduces a mild droop centered around 1200Hz.

Again, the response dip above 2k should be helped with the chamfer, updated measurements will confirm (or prove me wrong).

R1 can be tweaked to taste as well, I'd go 1/2 ohm increments - 3ohm and 2ohm sound OK, 3 is definitely a bit soft to my ear, 2 staring to get a bit hot. We're talking a bit less than 1ohm tilt changes on the top (5k+).

[cjd]

I should add. I really want to try a 3-way with the Anarchy now. But I want a bit more power handling, may have to consider 4 per side... (unless we get a 12ohm option... heh. 3 per side seems like it would be the sweet spot)

[gregnash]

ttt? Any updates?

[cjd]

I have dropped C5 to 10uF.

I did a new set of measurements today (stymied slightly by the neighbor having his yard "done" and tons of gas engines roaring, yuck. 95+db IN MY HOUSE) and plugged things in, crossed my fingers, prayed, closed my eyes, and ran it through the sim.

Nice and flat, the dip around 3k is down to -3dB from nominal, making this practically +/-1.5dB (better, I think, than +/-2dB)

Now to polish up the parallel version of the crossover for those that prefer that, not to mention boarding this crossover up so I can mount it and not have stuff dangling around on the floor.

[---k---]

Now to polish up the parallel version of the crossover for those that prefer that

I will be interested if you can hear a difference... of course any difference could always be chalked up to the difference of quality of the networks... Do you have a part count / cost for each network yet? Any significant differences?

[gregnash]

Good to hear that things are going smoothly. CJD I know that you have not really had a chance to fully test these but how do you feel they stand up against something like the Nat P's? From what you were saying to me in another thread, these would be comparable in price, so I am curious as to your listening preference. I was originally planning on building the Nat P's in a curved cabinet but if these will put out the same listening experience AND be cheaper then I am sold!!!

[cjd]

I'm not sure they'll be cheaper than the Natalie P though I don't know - the Anarchy has gone up in price so it's notably more than the RS180. 2.5 way gains some expense over a pure 2-way as well, though not quite as much as a full 3-way.

I believe crossovers each are running ~$125 each give or take, particularly if you can source all your inductors from Erse - I'd have to add it all up, I've picked up parts here and there, not to mention over a bit of time (and some out of my parts bin). You always have the choice to go with some fancier caps (you'll be wanting to split a 15uF, 25uF for the series version, parallel has smaller values . . .)

Parallel version I think I will not be building. So far it's being a bit more ornery about that dip in response, but I'll keep working at it. I'm quite confident it will work out, but the end result is I won't be doing any A/B comparisons.

Parts count is the same for both. 5(6) inductors, 5(6) caps, 3(4) resistors.

C

[gregnash]

Thanks for the info CJD... I will wait for some listening impressions then. No matter what I am sure that I will be more than satisfied with either setup just want to be able to pick something and start creating saw dust soon!!!! Thinking of going with a Magico V2 style build should I go with the Ansonicas?!

[---k---]

Everyone has built the NatP's. While a good design, BE DIFFERENT!

[gregnash]

Everyone has built the NatP's. While a good design, BE DIFFERENT!

Oh don't worry I probably will.... On the first page you (or was it CJD) some tower designs.... How tall can these go for a sealed version do you think? Or will they be better off ported?

[cjd]

Parallel version is vexing. It works nicely, but it just doesn't want to work without about -1dB in the response dip...

Greg: Not sure what tower you're speaking of?