4 way speaker help high end

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hey guys I was originally going to buy a speaker but there's nothing on the market that really matches exactly what I want, plus I enjoyed building subwoofers in the past

It will be an entirely digital source, as I want to actively split the signal for each component.

I want to use 2 inch tc epic 10's in each speaker, each in sealed cabinets. I will be powering them with 500w each.

Here's where I run into problems...how high exactly can these drivers play? I was hoping to cross them over at 150hz but I know this may not be possible, it may need to be more like 80-120.

Next I was going to use a ribbon tweeter from 20 khz ish to 4-6 depending on which model. Open to recommendations.

For midrange I was thinking of using the Bohlender Graebener Neo10 or a high quality driver, not sure which, would like recommendations.

Next is the part I'm really stuck on...I need something that can really produce massive mid bass (protentially 80 hz-300 ish) slam. I was thinking maybe a pro driver? 10 inch/2x8 inch with decent xmax? Really not clue, something with really low distortion at high volumes, which still sounds full at low volumes.

Anyway if any of you guys could offer some help I'd be really grateful, I don't know a lot about the frequencies above 80 hz....

Design goals are speakers that are able to go really loud with very low distortion, which are roughly 11-12 inches wide.

So to sum up .... 25/30 - 80/150 ...tc epic 10 x2 each side
80/150- 250/400 mid bass driver??recommendations wanted
250/400 - 5k-7k neo 10
5/7k + ribbon tweeter/other recommendation


If you could also help with amplification/ a digital crossover to split/adjust as required the signal that would be great. Thanks in advance.

Budget is basically unlimited


If there's a better driver than the tc epic 10 then I'd be interested in hearing about it...
 
As far as the specific drivers i don't know much.

In terms of the electronics MiniDSP 8x8 and UMIK-1 measurement mic would be my suggestion. Then you can pretty much throw any drivers together and integrate them with full freedom over crossovers points, levels, EQ etc etc.

If you want a better sound quality then suggest nanoDIGI taking the spdif outputs to your own DACs.

Alternatively, if you want to have the power of FIR filtering, then mininSHARC module with I2S interface to your own DAC modules will do nicely.

When you're talking about 8 channels of amplification with budget, i'd be looking at the Hypex modules.

Lower power for the tweeter and higher power modules for the bass.
 
Last edited:
"I hope you can afford to loose/waste it too"

I'll set a ROUGH budget, less than £12000

I won't be building this yet, will do it in two years, broke at the moment (student) but guaranteed position in a lucrative field.

will spend a LONG time researching before I do anything.
 
I would like to think this is possible for less than £5000...wouldn't imagine it would cost too much more than that...

tc epic 10 = $150, x4= $600
Bohlender Graebener Neo10 Planar Transducer = $200, x2 = $400
Ribbon tweeter = eg Aurum Cantus G2 $200 (this unit may change) = $200 x2 = $400

then there's the mid bass... say $600? for 4 8inch/2 10?

subtotal $2000

mini dsp $400

Speaker cabinet $200??

that's $2600, then I need amplifiers, lets say $2500

then finally leave $400 for random expenses.

so $5500 = £3540, plus import taxes if \I use mainly american products...still under £5000.

I say cost no object but diminishing returns above this price when doing it all diy surely?

I definitely want to build it myself/be able to tweak everything...that's at least half the fun :)

Still having a problem with the mid bass driver... thinking dayton reference/Aurum Cantus AC-250MKII 10" Carbon Fiber Sandwich Woofer/something by seas... really no idea what would sound cleanest loud and have good sound at low volume in a small enclosure.

Thanks for the help so far, the suggestions are making me feel like this is possible!!
 
Problem with the beyma speaker is the xmax...4.5mm, at 100 hz I don't think it will be enough.

I want the woofer above the epic 10's to be able to equal the output at the crossover point if possible, efficiency of the epic 10's are low (85db) but they genuinely take 500rms
 
Problem with the beyma speaker is the xmax...4.5mm, at 100 hz I don't think it will be enough.

I want the woofer above the epic 10's to be able to equal the output at the crossover point if possible, efficiency of the epic 10's are low (85db) but they genuinely take 500rms

Efficiency is relative when running a system active. Adjusting the level either digitally or preferably input level adjustment on the amp. to match. Looks like your overall upper power limit would be around 112-113dB.
 
Maybe try simming the Beyma in a TL, ( Leonard Audio - Coming Soon , great tool from a fellow forum member)

How loud do you want your system to be? I am in the process of designing a 115dB SPL system with 2 of these mids and can hit that target with a .5 configuration and LR8 filters. They will cover 150 - 800Hz in this application, crossing over to an 1.4'' CD Horn on top and double 10'' woofers in a TL on the bottom.
 
Last edited:
Well I don't have a lot of time at the moment to do a real sim, but modelling the peerless xls 10 with about 6 filters I got a completely flat response and as it has roughly the same cone area roughly as the tc epic 10 I was able to see the cone excursion/spl.

I'm at 21mm excursion (which isn't disastrous for the tc epic 10), xmech is 37 I think, xmax 18.1 (for each driver) at 30 hz (roughly my target for low frequencies) to get a (combined) spl of 117 db (assuming one driver with 84mm of excursion)

Then power limit on the driver (total 2000w for 4 drivers) at 85 db sensitivity gives a limit of 118db. Then there are power compression losses to account for. However, then there is room gain to add.

SO take home message is because there will be room gain and I want to keep everything working not at 100 percent I want the mid bases to be able to do 120 db (1 meter, both speakers). For a 10 inch driver this means an xmax of 7.5mm (assuming 8.5 inches actual radiating area). For 2x 8 inch drivers (radiating area 6.5 inches) this means an xmax of 6.5 mm.

Ideally I'd like the mid bass drivers sealed, to make the design more compact and improve transient response. Is there a driver which can do this? If not I'll change my goal slightly.

For the neo 10's I get a max spl of 111db each (92db sens, 75w AES) for for 2 it's 114db....Seems like to reach my goal I'll have to get 2 per side to get 117db?

^Maybe I should look for another option for midrange? I saw a thread a while ago with the b&w 800d midrange which wasn't that expensive as a replacement part which had 96db sensitivity or something crazy as well as good power handling. Not sure how easy they would be to get hold of though... open to other suggestions.


Ribbon tweeters should do 118db in a pair.

Would like to be able to do 115db (1m for both speakers, so 112db each without the speakers playing to their limit).
 
Last edited:
TBH I don't know, are there distortion measurements available?

The goal you set there is pretty hard to reach, power compression will happily take 4dB of output from the theoretical max SPL even with pro-drivers at those drive levels. Also take into account that most ribbons have a very narrow dispersion pattern in the vertical plane, but I do like the way they "sound", take a look at what I did with a big ribbon (and learned what not to do, my next design doesn't look very similar). http://www.diyaudio.com/forums/multi-way/207159-trinium-my-latest-build.html
 
Last edited:
Well I get a "theoretical" max of 121, assuming 4 drivers, 200w each, 92db spl @1w

Assuming 4db losses it's still 117db.

then at 100w it's probably going to be 115 because of less power compression, but crucially assuming the cabinets are build like tanks should sound incredibly clean at this level (if the drivers are as good as im hoping).

What I'm aiming for is something I can just turn up as loud as I want in a loud room which will remain controlled at all times.

I don't know what output max will really listen to but if it does 115db without straining then 100-110 should sound amazing (I'm hoping).



An idea did strike me while I was thinking about the dsp...even if it's corrected to be completely flat at say 90db, the output will become "unflat" at different power level inpputs, all sorts of variables like power compression, nonlinear suspension/cabinet resistance etc etc.

So then I though about measuring it flat at a series of output levels, giving a three dimensional function which could be given a high order polynomial to allow interpolation to the exact power input.

Then I realised you can't judge all these effects from pre recorded output signals at one frequency because music is much more complex. So I thought about creating a stochastic model which runs in real time which analyses the expected future output of each driver to help give a more linear response no matter what the signal is or the previous decay of heat which could be modeled via the partial differential heat equation.

I may be trying too hard here...

EDIT I also don't know that much about audio engineering to be entirely honest^just an idea I'm floating.
 
Last edited:
Why not go for dual TD15H crossed active at 60-80Hz to a Scan 18W Revelator for the range 80-300Hz and then go for something like a Scan 15M or a Fostex FE168 to match the ribbon? The fullrange driver would give you more flexibility in the ribbon matching but the 15M`s a nice contender too and can accept lower order crossover. For ribbons, not sure why it has to be a ribbon when there are exceptionally good domes, but I would look into the Raal range if I was you. A four way is very hard to design properly.
If you go with low woofer cone area it will need excursion to compensate, that introduces nonlinearity and distortion especially if driver has poorly designed motor. Using the Fostex and some high sensitivity tweeter can built you a system in the 93-94db range.
 
Last edited:
Your initial driver choices made me think of this:Introducing: Echelon also pictured below.

Unless you're using these as a PA system, I would think 112dB is a little bit over the top. Peaks to 100 or 105dB should be more than adequate especially if you want to keep your hearing for the rest of your life.

Since you seem to be after very low distortion, I'd stick with the Neo10. Distortion measurements below split into 2 graphs. Very nice between 550 and 3000Hz.
Naturally the Neo3 as a tweeter is a logical suggestion. Bohlender Graebener Neo3-PDRW Planar Tweeter w/Back Cup 264-735
Maybe Rythmik for the subs? Servo subwoofers •Rythmik Audio Direct Servo subwoofers
And Audax has some nice high sensitivity woofers
The Madisound Speaker Store
The Madisound Speaker Store

I would probably fire your subs on opposing sides, cancelling their vibrations and leaving lots of baffle room for multiple drivers for the remaining 3-way.
 

Attachments

  • IMG_2692.jpg
    IMG_2692.jpg
    278.1 KB · Views: 199
  • B&G-Neo10-HD-50-1.5k.gif
    B&G-Neo10-HD-50-1.5k.gif
    19.5 KB · Views: 200
  • B&G-Neo10-HD-1.5k-20k-2Xdistance.gif
    B&G-Neo10-HD-1.5k-20k-2Xdistance.gif
    17.1 KB · Views: 197
Your initial driver choices made me think of this:Introducing: Echelon also pictured below.

Unless you're using these as a PA system, I would think 112dB is a little bit over the top. Peaks to 100 or 105dB should be more than adequate especially if you want to keep your hearing for the rest of your life.

Since you seem to be after very low distortion, I'd stick with the Neo10. Distortion measurements below split into 2 graphs. Very nice between 550 and 3000Hz.
Naturally the Neo3 as a tweeter is a logical suggestion. Bohlender Graebener Neo3-PDRW Planar Tweeter w/Back Cup 264-735
Maybe Rythmik for the subs? Servo subwoofers •Rythmik Audio Direct Servo subwoofers
And Audax has some nice high sensitivity woofers
The Madisound Speaker Store
The Madisound Speaker Store

I would probably fire your subs on opposing sides, cancelling their vibrations and leaving lots of baffle room for multiple drivers for the remaining 3-way.
I agree: +105dBC should be enough. Other issues should be of more concern than the last drops of dB-juice.

Personally I'm on a hornstray right now, but if I were on a more "domestic" route, I'd probably go for a 2,5-way + sub (80-100Hz). Beyma 150TPH + 2x10" PHLs + subs would likely be my best guess.
 
Last edited:
Well I get a "theoretical" max of 121, assuming 4 drivers, 200w each, 92db spl @1w

Assuming 4db losses it's still 117db.

then at 100w it's probably going to be 115 because of less power compression, but crucially assuming the cabinets are build like tanks should sound incredibly clean at this level (if the drivers are as good as im hoping).

What I'm aiming for is something I can just turn up as loud as I want in a loud room which will remain controlled at all times.

I don't know what output max will really listen to but if it does 115db without straining then 100-110 should sound amazing (I'm hoping).



An idea did strike me while I was thinking about the dsp...even if it's corrected to be completely flat at say 90db, the output will become "unflat" at different power level inpputs, all sorts of variables like power compression, nonlinear suspension/cabinet resistance etc etc.

So then I though about measuring it flat at a series of output levels, giving a three dimensional function which could be given a high order polynomial to allow interpolation to the exact power input.

Then I realised you can't judge all these effects from pre recorded output signals at one frequency because music is much more complex. So I thought about creating a stochastic model which runs in real time which analyses the expected future output of each driver to help give a more linear response no matter what the signal is or the previous decay of heat which could be modeled via the partial differential heat equation.

I may be trying too hard here...

EDIT I also don't know that much about audio engineering to be entirely honest^just an idea I'm floating.

Maybe take a look at MFB systems? They use feedback to battle these phenomena. Motional Feedback - Wikipedia, the free encyclopedia

IMO it's also a good idea to setup the DSP at levels that matches your "typical" listening levels after a short warm up time. That should at least minimize non-linear effects.
 
Last edited:
Maybe take a look at MFB systems? They use feedback to battle these phenomena. Motional Feedback - Wikipedia, the free encyclopedia

IMO it's also a good idea to setup the DSP at levels that matches your "typical" listening levels after a short warm up time. That should at least minimize non-linear effects.

The MFB system looks interesting but it also relies on correcting based on feedback (as far as I understand it), at least theoretically it would be better to correct before the signal goes out?


The lamba 10TX drivers look interesting, could do a three way with them covering up to 400hz ish to the neo10. They have amazing sensitivity, but I wonder if they will sound as clean as the epic 10 at 20-40 range at high output levels. Their xmax is 14mm but using three and no mid bass would certainly make up for it.

I don't want something bigger than 11-12 inches wide so 10 inch drivers are the upper limit. Although maybe I could side fire a pair of lmsr 12's? But the cabinet volume required may make them really deep then...

I want something which can, when required, fill a large room, but which isn't too big. I won't be listening at 115db near field!! Don't worry about my hearing.

Thanks for the replies so far, it's been very interesting and informative for me.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.