Almost-Newbie needs some 3-way speaker & tri-amping advice :)

[Vortex]

Better than "have to" compensate for sensitivities, it becomes trivial to do so, and becomes part of setting the crossover up. Gain controls are going to be in your amps or your crossovers or if DSP, integral in the crossover. Don't be afraid of DSP, it makes your job so much easier to get the sound you want.

Phase2. Or not. Hmm.
*considering*

 

[Vortex]

This is just in case you're not familiar with this or if you have perhaps forgotten:

- 2 drivers in series gets you the same sensitivity as a single driver but impedance is doubled
- 2 drivers in parallel gets you a 6dB increase in sensitivity but a halving in impedance. This is why so many dual mids and dual woofers are wired in parallel. And because many (most?) amps today have no problem with a 4ohm load.

So just to double check - must you limit yourself to 8ohm driver loads?

Also, driver size is not necessarily the final arbiter of how low a woofer will play. Instead, it's a driver's combination of TS parameters that determine if it is more suitable to a closed or ported box and how low (F3/F6/F10) it will play in a given size box. Each driver needs to be simmed in a competent box modeling program, like Unibox (Excel required) or WinISD pro to know what it's going to do on the low end. You may find that defining your design F3 goal and your max woofer box size may help you in your woofer selection.

Thanks for this knowledge refreshment - apparently I knew it wrong then. I thought, any kind of doubling would increase sensi by 3dB not 6. So, thanks again, very valuable info

I actually don't stick to 8 Ohms, we can build 4 Ohm capable amps too, however except electrical parameters, wattage, power need etc. I don't know which nominal impedance setup is better in general or if there's a difference in sound quality at all (between 2 very identical 4 and 8 Ohm systems).

 

[Vortex]

Finished multi-amp project with amp modules and Linkwitz Riley filter

 

[Vortex]

Thanks for this knowledge refreshment - apparently I knew it wrong then. I thought, any kind of doubling would increase sensi by 3dB not 6. So, thanks again, very valuable info

I actually don't stick to 8 Ohms, we can build 4 Ohm capable amps too, however except electrical parameters, wattage, power need etc. I don't know which nominal impedance setup is better in general or if there's a difference in sound quality at all (between 2 very identical 4 and 8 Ohm systems).

Btw I'm not trying to make my bass section sound in LFE region below 20Hz but if I take one of my favorite songs from the less beloved pop industry, it's a shame that many audiophile grade speakers simply cannot reproduce these basses like they should. (Good example is again the Audio Note AN-J I listened to, it sounded amazing with the Eric Clapton unplugged LP but this Madonna album was quite weak at the bottom. Remarkably weak). This song is btw one of the most challenging ones IMO for bass sections: the smooth, fine, continuous bass tone has to come at all frequencies with the same power and tons of badly designed open boxes fail here due to wrong reflex port tuning etc.

https://www.youtube.com/watch?v=a4ChERg__G8

 

[Vortex]

Another one: https://www.youtube.com/watch?v=MDM0yAJjlBo
At 0:56 and especially from 3:01 .. these should sound like smooth-beautiful soul touching earthshakes (cannot describe it better) ..

And here bass plays also a big role, flat response is key without peaks (and of course measurements are 1 thing and our ear's sensitivity is another) : https://www.youtube.com/watch?v=BEDvAGq2dEA

Anyway, bass is very crucial for me, right. I used WinISD several years ago, playing around with a Fostex FE208EZ driver (which I heard on the same setup like the AN-J-s, put into Jericho horn baffles - omg, sweet, loved that, except missing some highs).

I'll play around with the program again soon this weekend. Thanks for the remarks/tips.

 

[LineSource]

Vortex, I tried to meet your stated goals.
BUT.... Read Geddes paper on controlled directivity, and share a look at designs which reduce room interaction with your Dad like Fusion-12 EconoWave and the larger Tweek Geeks BMF-1

*Avalon style low diffraction vertical baffle cabinet. Google: Avalon diy speaker images
*Tweeter Ribbon = Fountek Neo X3 (Rock&Roll durable, $110)
---(LR4 electrical at 1900Hz)
---(time delay Xover circuit)
*Midrange = 6" Satori MW16P-8 in sealed 0.5cuft volume ($130)
---(LR4 electrical at 1900Hz)
---(LR2 electrical at 190Hz since sealed alignment and baffle step will also attenuate)
---(time delay Xover circuit)
*Woofer = 12" Dayton DS315-8 in sealed 2cuft volume ($80)
---(BW3 electrical at 190Hz to attenuate any midrange grunge(reverse polarity))
---(include 20Hz Qtc=0.7 LinkWitz Transform bass boost crossover circuit)
---(include baffle step compensation Xover circuit)
---(explore room compensation circuit)

*Active crossovers cannot correct lobing from large Center-to-Center driver separations or baffle diffraction effects. This Avalon style cabinet is optimized for diffraction control. There are several diy-copy drawings on the web which you can give to the cabinet shop.
*WinISD plot for sealed Mid, and sealed Woof with 20Hz LinkWitz Transform provided to show good efficiency match.
=================
http://www.gedlee.com/Papers/directivity.pdf

https://www.pinterest.com/localhifishop/tweek-geeks-bmf-speaker-in-the-making/

https://www.google.com/search?q=Ava...KB5rnSAhWEMGMKHZJXDUAQsAQIGw&biw=1280&bih=654

http://www.htguide.com/forum/showth...nt-Reference-Thread-How-we-realized-the-Dream

 

[Vortex]

@ LineSource: thanks, great stuff !!!

About baffle: the only things I kept in mind until now:
- it should have no parallel walls at all
- voice coil planes should be vertically aligned when we look at their final position from the side
- white piano finish

You provided me quite some hints now. Still processing the info

Of course I'll have to take care about room resonance and reflection effects but I'm also trying to weaken reflections: I'll have big pictures on the walls, special ones: actually putting sound absorbing material into (onto?) the empty picture frame (2-4 cm thick) and then giving this damping thing a nice 1-layer white textile coating. I'll have some of my photos printed onto the textile material but I'm also thinking about simply having something abstract weaved into the textile itself, not clear yet. Anyway, my walls will be somewhat damped, at least a big proportion of them. Carpet is given (covering almost all space between TV and furniture), ceiling is naked but will remain naked for now, for the 1st try.

Anyway, I'm not a perfectionist (I lied now) but I'm happy to see couple of similar faces here who pay special attention to a bunch of parameters involved around audio. There's plenty of them to screw up a project completely.

 

[Vortex]

Hah, those connecting aluminum rods in the second link's pictures.. very nice. I thought about something similar when building a big tall box but made of hard wood. Aluminum might be more rigid. Cool.

PS: anybody ever built any kind of separate box(es) out of hard concrete ? That must be stiffness omg

 

[Vortex]

Off a little bit: N1: Perfekter Klang aus Beton-Boxen - DETAIL.de - das Architektur- und Bau-Portal - Concrete. (I see ScanSpeak Revelators here, tweeter too).

 

[LineSource]

re: Perfekter Klang speaker......
Analog and digital electronic crossovers provide time delays which remove the need for a step-baffle. The edges on a step-baffle generate diffraction. When a sound wave reaches an edge(baffle or step-baffle) it "sprays" the energy in several different directions, often also generating odd angle wall reflections, and this confuses the soundstage location queues. Some room wall reflections might generate desirable ambience. The truncated pyramid Avalon cabinet's beveled edges generate a modest amount of randomized direction diffraction, dither, which some speaker experts believe can improve the listening experience in a home.

With "exceptional" recordings, large-radius(3"-6") edge treatments, like the Sound Physik Kronos or B&W sphere, can generate a more accurate soundstage reproduction than beveled-edge cabinets. One person's "accurate" is another person's "bland".

As expected, mitered bevels around the tweeter offer the most cost/work effective DIY cabinet construction benefits. If you need to reduce cabinet construction costs, you could explore these ideas.

 

[Vortex]

After long considerations I've decided to work with these drivers (meant for 1 side here):

- low: 1x Dayton Audio PA460-8 18" Pro Woofer
- mid: 4x FaitalPRO 6FE100 6" Professional Midrange Midbass Woofer 8 Ohm
- high: 1x Aurum Cantus AST25120 Aero Striction Tweeter 8 Ohm


Outfit would be most-probably:
- separate trapezoid vented box for the bass driver (something like here but lower and no parallel walls)
- separate thin box for the mid-high section (still considering a grouped-together design in favor of vertical sound dispersion instead of the classic thin-tall design).. put onto the bass box..
- white piano finish
- all speakers will be IEC-measured again one by one in the lab to see real data instead of manufacturer datasheet (and correct/calculate any deviations)

Further considerations and ideas:
- speakers selected with more or less the same sensitivity (mids have lower SPL but 2 in series and then this to parallel increases sensitivity by +6dB and will match with the bass and tweeter) - just for backup purposes, should the system be driven with passive 3-way crossover and 1 stereo amp temporarily

- generally good sensitivity (around 98dB all, that's okay)
- power handling when using them as home cinema fronts (30% of time)
- custom preamp (passive ? active?)
- 24 dB/Octave 2/3-Way Linkwitz-Riley Electronic Crossover, XO freqs at around 250Hz and 2.5 KHz
- tri-amping
- Bass: some kind of MOSFET monoblocks at around 500-700W 8 Ohms
- Mids: tube push-pull class AB with KT 150s ~ 150-200W 8 Ohms
- Highs: not sure yet. Pushpull class A OTL might be weak here, or even not.. (have couple of 6C33C-B NOS tubes).. I'm also open to MOSFET again too .. or else.. whatever drives the ribbons the best (sonically)..
- gains matched..
- going to be my very personal big experience, all this stuff. I hope the best.

 

[karl sartori]

re:multi way

for crossovers i suggest you try a dbx pa2 unit,plug it in and go ,it has the ability for you to adjust all parameters on the fly.check it out on youtube.i've been using it for a couple of years in aDIY triamped phono system and it is spectacular,no sound colouring whatsoever, it being a digital unit.my threepence worth.

 

[Vortex]

Thanks, I'll give it a try

 

[Vortex]

Finished multi-amp project with amp modules and Linkwitz Riley filter

Do you use same amp modules for bass-mid-highs ? (I mean: absolutely the same?)

Right now I'm thinking of healthy-required power needs to driver the sections with enough peak-reserve.. so I came across
- a 2500W@4R Class-D for the bass
- ca. 250W@8R for the mids
- ca. 100W@8R for the highs

Now how I'm going to design amps for that, is another question. 2 factors to be considered:
- I have to maintain the same amplification gain in the designs all over the chain as a basis, allowing for some fine-tuning of gains via DSP or preamp or on the amps themselves. Any good practice where to adjust these ? I think there're several points we can fine-adjust gains and achieve the same SPL at the end
- amps for mids and highs can be "smaller", they require less power so the PSU part can shrink as we go up with freq.

For the Class D I'm going to buy a Hypex UcD2k.

Amps for the mids and highs are still open, although DIY for sure. Just cannot release the challenge and the fun of building these for ourselves. All class AB for sure.. (don't want to heat the room).

 

[Vortex]

Made some calculations for the most critical part - bass. The small increase in SPL peaking at ~130Hz is the "Simulate voice coil inductance"-effect of WinISD. True or not, that's it for now.

I played around with the parameters and I'll probably go with the red line version - 600 liter box, tuned at 24Hz. It's also ideal, somewhere around an almost 1 meter long cylinder with a 18" diameter (like the speaker itself). I might correct the bass section with gain on the filter side - or even leave it as is.

An externally hosted image should be here but it was not working when we last tested it.

 

[globalplayer]

Made some calculations for the most critical part - bass. The small increase in SPL peaking at ~130Hz is the "Simulate voice coil inductance"-effect of WinISD. True or not, that's it for now.

I played around with the parameters and I'll probably go with the red line version - 600 liter box, tuned at 24Hz. It's also ideal, somewhere around an almost 1 meter long cylinder with a 18" diameter (like the speaker itself). I might correct the bass section with gain on the filter side - or even leave it as is.

An externally hosted image should be here but it was not working when we last tested it.

You want the FR of the enclosure flat. In this design you loose 3dB of bass output.
Ideally you want to look at the datasheet of the speaker where the speaker rolls off like 3 to 6 db and then tune the box so it will have a 3 db boost a little lower where the speaker rolls off and the port will make a flat curve in the end and extend bass range for free and flat.

 

[Vortex]

I'll equalize accordingly. Or even win this back with correct placement since the boxes will be put into a room. We'll see after some fine tunings regarding gains & some equalization, if needed at all. The DSP will do the trick.

 

[Vortex]

@ Steve/BlueWizard (from the other topic):

The Faital midranges are exceptional, confirmed by one of my friends who works with PA stuff since more than 10 years (stage amplification). This applies to some standalone midbass/miranges of larger size as well. This is what I consider now, using 1 bigger midrange of ~95dB, this way I avoid this 4-midrange-madness I got into , furthermore I win some vertical space to lower tweeter and midrange into head-position when sitting. Originally I planned a vertical array with the 4 midranges but then all these stacked on top of the huge woofer would have placed the tweeter too high already for a sitting listener and AMT's vertical dispersion isn't that great we all know.

Faitals are great and they also handle power well, according to some russian youtubers a guy tried to burn a Faital midrange "as usual" (crazy guy) and couldn't manage, BMW car hifi, big amp (although 2000 Watts car-amp -> usually meant for 2 Ohms, so trying to burn a PA 8 Ohm midrange with that hmm.. so-so.. might be strong enough or even not). These are fine PA stuff, apart of good linearity and great sensitivity also very well built so I just completely skipped normal home hi-fi range of speakers (except the AMT tweeter which is again, capable of 100W RMS. Even if only the half of it is true, crossed far above factory recommendation, in a safe 2-2.5kHz region, I don't have the feeling I'm going to blow them tomorrow). They're not cheap ? Well, I'm not that millionaire type but looking at how nice it would fit into this design I cannot resist. Speakers (if good) serve for years anyway right ? I like ribbons anyway.

It would take slight exception to the crossovers you have chosen, but that's a matter of personal opinion. Give the very wide response of the Woofer, I would be inclined to think 250hz and 2000hz to 2500hz. 150hz takes the Midrange down to the bottom of its workable range. The more inside the workable range, the better the Mid is going to perform as is true of just about everything in life. Pushed to the limited, it is going to be strained, well inside the limits, substantially less strained.

You're right, I'll lift lower XO a bit upwards. Original very-first design was a box with a smaller front and a side firing woofer so to avoid rising bass directivity I tried to keep first XO as low as possible (that's also the reason for the 4th order LR). Now with a front facing bass driver I can lift a bit (but don't want too much. If I look at Mark Knopfler's voice on my spectrum analyzer I'd try to keep him mostly on the midranges - sounds crazy hmm ? Yes.. I dare to drive the strong mids at their lower 'limits' although they're stated as midwoofers, in fact it's mainly box size which I can still keep small to nicely cope with the bass with a sharp LR4 filter.

Also, especially on the low end, the closer you are to the low end limit, the less power handling you have. Though in your case with 4 Midrange, that's probably not a concern. An array of 4 would give you double the power handling, half the excursion, and a +3db boost in the output.

Yeah, LR4 helps here, again. If I'm trying 2nd order filters on the DSP first, I might increase lower XO to 200-250-ish Hz-s to see what happens.
Check this one - what I'm shooting for now. I don't think power handling would be an issue for the midrange, bass is 500W RMS so 200W RMS for 1 midrange is more than adequate here (in theory but I'll listen at fractions of all these of course) Low fs (outside of my planned mid region - less trouble with rising impedance), quite good freq response and even the 45° off-axis response is pretty nice, the scissor between 0°and 45° opening significantly well over 3K where I'm on tweeter already so crossing at around 2-2.5kHz I still might achieve a nice sound dispersion (horizontally). Vertically it's another thing, basically limited by the AMT but I don't care, I listen while sitting.

Generally PA Woofers do not go as low as Hi-Fi Subwoofers. The disadvantage to a Subwoofer is that the impedance is relatively low, and that implies you have to drive it with a separate amp. Though you were planning to do that anyway.

I have absolutely NO experience with PA stuff. This will be the first PA build. I like experimenting and taking risks, but considering the specs and the fact that PA stuff follows the stated specs probably more exactly than normal drivers do (PA manufacturers don't dare to lie that big) I hope a good sound with PA now. Box will be large however, I just ignored recommended size. For the music I'm listening to and looking at my spectrum analyzer, uh, no, 40-ish f3 is a bit high for me especially if this is a vented box with sharp slope down below f3.

Keeping in mind you have to subtract the volume of the drivers from the internal volume of the cabinet. So, let's throw in a bit more for the volume of the driver.
1.75 ft (21")W x 2.75 ft H x 3ft D = 14.44 Ft³
Is that a workable cabinet for you?

Oh my.. I'm sitting in the middle of Europe and going crazy of these metric/imperial switches. Aaaahh. (Sry not your fault). So.. If my calculator doesn't cheat on me, it's about 408 liters (408 dm^3) so yes.. I'm still going a bit larger, for now my WinISD project tells me 452 liters (15.98ft^3) tuned at 28Hz (just at fs) and f3 at 25.7Hz.

Lastly, if a full Active Speaker system, what Crossovers were you planning to use?

DSP first, it will provide me some filter types but after all what I've read here and there on the internet, I might try the LR4 type. Especially reading this here. Slope-wise I'm striving for a narrow crossover range so the sharper slopes the better (if SPL-s and/or amplitudes are matched nicely), only constrain is the ringing which increases with higher order filters - but hey, I let my ear decide. Now I'm on 3-way 2nd-order Butterworth passive one so this new project will be kind of a "holy ****" experience (in either directions but I hope the better one here)

For bass driving (500W RMS stated) I'm still (or again?) struggling what to use as amp. Class D I would allow here if I'm striving for plenty of reserve power. Some say +3dB needed for peaks so 1kW would be fine just in case I turn up the volume occasionally. But considering normal room power I might omit this +3dB peak reserve 'cause I won't reach that power requirement anyway so I might go for a 500W-ish Class AB Mosfet, custom build (with help), not more.

When Class D, I have a recommendation, a quick tip was the Hypex Ncore amps. Still thinking if I need this for normal living room powers, however, a 500Watt-ish Class AB is heating like crazy. Summer is hot here so there's a good chance I'm going for the Hypex boards.

 

[Vortex]

10PR330 .. just put couple of midranges into WinSD, this seems to be the best fitting - but actually, all are OK, I just fine-tune the results.

 

[BlueWizard]

1.75 ft (21")W x 2.75 ft H x 3ft D = 14.44 Ft³ (21"H x 33"W x 36"D)

Bass box only -

53.3cm W x 83.8cm H x 91.4cm D

Using 4 Six Inch drivers in a 2x2 array, I estimate the Mid Box to be 53.3cm wide.

So to get the overall height of the system - 53.3cm + 53.3cm + (best guess) 15.24cm for an overall height of 121.84cm or 47.97 inches.

But that is internal volume allowing a bit more for wood and seams, I would take it up to a bit over 50 inches or 127cm or 1270mm.

53.3cm Wide x 127cm High x 91.4cm Deep. Though the Mid-Box would not have the 91.1cm deep.

Steve/bluewizard