I want to build a uber-clarity 3(4?) way. First attempt=fail. What now?

[jackies]

Hi there fellow enthusiasts!
As I mentioned eariler, I'm after a super-puper-ultra nice clean sound. Just like many here, I would believe. I currently have a pair of Snell D speakers, which sound rather nice, but I imagine better speakers can be built. Thinking this way, I decided to build some. I started, naturally, reading internet for a number of years. Then I bought a ESS Heil tweeter - clear airy sound, right? Then I got Jordans JXR6 HD widebands. Being a last word in high definition - I thought they supposed to sound rather clearer than other hi-end midranges. But alas, they sound about as good as their $20 multimedia counterparts. Precision camera feel my a$$! I also got Visaton AL200 hi-end woofers. Coz they're hi-end, see?
Then I spent some time building a phase-linear active crossover, and after a while decided to just get a DCX2496. So finally, I got this lovely looking setup.

I braced myself, and BAM! The weakest sound I ever heard. Honestly, 2 way boombox speakers have more impact and more resolution! The Heils sound alright, I'd say, but Jordans mid are smeary and lispy and AL200s in around 50liter closed box produce no bass at all. I guess I could supplement that with a subwoofer, but even the mid-bass they produce I don't like the sound. So an utter, ultimate speaker bulding failure!
Now, I see my choices as follows:
1) Forget about it and be content with Snells, coz they are good.
2) Build some real hi-end speakers, using some existing and proven design.
3) Get a better mid (tangbang?) and a woofer for my existing construction.
Please give your advice now!
Is there a real super hi-end design that provides unforgettable, unrivalled clarity in sound? I really don't want to buy accutons for the "humble statement", but I guess I could splurge on some marginally less expensive project. Maybe Linkwitz Orion for dipole sound?
I don't want no MTMs, because I don't care for off axis response - I listen from a fixed position - my chair. Also high SPLs are not mandatory, I listen mostly at moderate levels. Clarity, airness, imaging, precision, resolution - is what I'm after. Fast, well defined bass also.
Please help!

[chris661]

Not surprised about the bass - what did you model it in?

For the crossover - what values did you use? how did you model it?

Using the boxes you have, you can add a bass reflex port, tuned to 40Hz and you'll get something half decent at least.

When building the cabinet, how did you brace it? Did you make sure it was completely sealed? One of the worst problems for bass speakers is a leaky box...

Chris

PS - what I'm trying to say is that you don't just give up, you find out what's wrong and fix it, then you sit there, listening to your favourite track, with a big grin on your face knowing that you fixed it.

[jackies]

Thanks for your reply, Chris!
Crossover, seeing how it's the DCX2496, I didn't model after anything. I knew that I can have my crossover points all over the place at pretty much any slopes, because of rather wideband drivers.. I had low/mid crossed anywhere between 250-700 Hz, liked around 300 the most, and mid/high crossed at 2k-10k, liked around 7k the most... Box I didn't model either, really don't see the point, except as math/computer excercise. Like, you will be getting some frequencies from a given box/driver, that's for sure. What exactly you will be getting you can hear once you finished. I was prepared to have a reduced bass response, which I was gonna correct with EQ, and I didn't want the port because I think ports sound boomy. I used a box from some old speakers, coz I was too lazy to build one from scratch, I just bought some speakers that were the size I wanted and had walnut veneer..
I added a box for the mid driver, with a system of braces for backwave dispersion/reflection/damping. For the woofer, I added a couple braces that also hold the magnet tightly. Everything is additionally sealed with epoxy, and the drivers are seated on a thin layer of silicone sealant for additional measure of airtightness. Although, when i was building subwoofers eariler, I noticed that leaking air doesn't affect the bass performance per se, it gives some extra whistling or some sort of sound of the air going through the holes...

[chris661]

In my experience, it's only the ported designs which feature a too-small box and a port tuned in the wrong place that lead to boomy bass. For example, PC sub/sat systems.

Did some modelling and a port tuned to 40-50Hz will give you a lot flatter bass response. Alternatively, a Linkwitz Transform will take you as low as you like.

When spending the money you did, you must make sure you model the speakers - otherwise you just won't know what will happen. I'm not saying the simulations are perfect, but they give you a good idea about what will happen. You can download various pieces of software which will help you model the frequency response of your system.

For now, disconnect the tweeter, put a 90uF cap in series with the FR driver, and a 3mH inductor in series with the woofer. This gives an XO of 300Hz, 1st order roll off. See how that sounds. You are hearing more what the drivers can do how they are. You still won't get as much bass as you'd like because you have a sealed box which isn't necessarily the right size.

I'd strongly recommend adding a port, tuned to 40Hz, as this will give you significant bass output to 40Hz (-3dB point), whereas the box you have gives an f3 of 80Hz. The worst that could happen is that you need to block the hole you made. But, you don't necessarily like them as they are, so adding a port may be the answer.

Chris

PS - leaky box affect the bass more than most think - you get the effect of "port below tuning", with a steep rolloff, but, because it's a leak, the "port" could be tuned anywhere, meaning the rolloff could also be anywhere.

[Henckel]

You need to measure the response of each individual driver in order to dial in a starting point for the DCX and then tweak from there. When you do this you should easily get better quality than the Snell´s how ever the DCX unmodified is not the last word in high end sound ;-)

[jackies]

Well, that was exactly my thinking.
Measurements of AL200 were done by Zaph himself, I didn't feel I had to re-measure, I did measure the Jordan, however different software gave different results, also I was getting different results with different mic placements... From my measurement results, I figured the best for Jordan was 500-5k. That's what I used as a starting point with DCX. I also used 24db/oct slope as a start too. Then, I spent a couple days listening, and the settings I liked the most, were 300 and 7k with 6db/oct slope - I liked the most, meaning I was getting the clearest sound. And even with 6db/oct slope, it's well within the capacity of the drivers, because AL200 should easily go to like 2k, Jordan 180-30k, and Heil should be able to start with 1.6k... And as I already said, I was prepared for lacking bass, I didn't expect much extension from a sealed design like that, 80Hz sounds alright. What I didn't like is the lack of clarity. Snell still sounds much clearer...

[chris661]

How does the Jordan sound on it's own?

How long did you give them to bed in?

I'd try a steeper slope for the woofer - it may be colouring the midrange (aly cones aren't the best for midrange use)

Chris

[wintermute]

I think Chris' PS summed it up If you don't work out what is wrong with the current design, you really can't expect to do better next time

I think the first thing you need to lose is the opinion that modeling is a waste of time . Yeah sure if all you do is model and never actually build something then it is, but if you just buy a driver because it is a good one and then simply chuck it in a box and hope for the best then well, you might get lucky, but you probably won't

I just modelled that Jordan Mid... what size enclosure did you make for it? I modelled it with a two litre enclosure (about the middle of the suggested sizes). It starts to roll off at around 260 Hz, but probably more telling is that it goes into non-linear operation at around 170 Hz at only 10W power input (10W should still be pretty loud depending on the overall efficiency)... This could explain to a degree your preference for crossing it at 300Hz... edit: just saw you changed to 6db slope on the crossover, the mid would be getting significant bass energy if crossed at 300Hz with a 1st order crossover, that could easily be causing quite a bit of distress to the freq's you want it to be clear in, especially if you are using a reasonable amount of power.

I also modelled the AL200 and if in a sealed box the recommended volume was 20L. I've attached graphs for both your 50L sealed and 50L vented with a tuning freq of 35Hz (40Hz seemed to have a hump when modeled with unibox). I think you should see the difference straight away

For completeness there is a plot for 20L sealed as well.. this woofer with its low Qts is definitely more suited to a vented enclosure IMO.

Modelling is pretty easy, and can save you a lot of heartache If you want a closed box then you really should buy a driver suited to one! the fact that the only suggested boxes on Visatons page for this driver are vented should be a hint

Sorry if this sounds a bit harsh.... It isn't meant to be more than maybe a reality check. If you want to succeed there is more to this hobby than simply buying "the best" drivers.

Tony.

[gainphile]

Clarity?

Try to remove the box

[jackies]

Nothing is too harsh, and yeah, this project is a BIG reality check for me. In this case I think the drivers choice was wrong. Jordan is probably meant for like, rear speakers in a HT setup or a line array in the same HT setup or something. Not for critical music listening. Box I built for it is around 2-3 liters. I had it playing radio around the clock for a few days to break it in.. It sounds pretty good, for what it is, except for the lack of lows/highs. Again, it gots lots of bass for the driver it's size. It just doesnt sound pristinely clean, kinda little smeared and (very slightly) lisping.
I'm not trying to diminish the importance of modelling in loudspeaker construction. The Snell I have uses modest Vifa driver and doesn't lisp even slightly. I think I'm talking more about listening impressions with different drivers. Unfortunately, there's really no way to listen to each and every driver and thus make an informed decision. I was looking for a clean sounding mid driver, naturally if I had a few drivers and listen to them it'd be easy for me to pick the one I like the most. But basing my choice purely on reading is of course difficult and prone to mistakes. So that's the mistake I made right there. If I wanted to make a good enclosure with certain volume, using certain driver, producing certain frequences, then sure, modeling is the way to go. Measuring, adjusting, all fine with me, I love engineering/mathematical stuff and love tinkering. But as to choosing drivers basing on their capacity of producing clean uncolored high-mids?
Suggestions?
Lose the box? I'm ready now. So Orion, or is there any other well documented projects like that?
Also I'm thinking maybe http://www.troelsgravesen.dk/cyclop.htm