How would you handle this XO situation?

I am building clones of the Triad Gold/4 surrounds for my HT (less not doing them di/bipole). I will be running 6 of these total (2 each side, one in back L&R)

[IMGDEAD]http://www.coretrends.pl/static/images/previews/0/6624-0.jpg[/IMGDEAD]

Using:

Woofer - peerless 830869 8"
Mid - Peerless 830987 3"
Tweet - Tymphany BC25sc55


My issue is I am having trouble figuring out what to do for the XOs.
My original thought is that I would run active 2 way (running the 3" HP and using a cap on the tweets). The issue is that I wanted to run the 8" crossed over above ~80hz.

Now my head is swimming......Try to use pro audio XO? Car audio XO with 12v power supply? Build my own passives?

I really like the idea of running active because it would mean I would not need to construct 6 complicated individual XOs but I dont know where else to turn.

What would you recommend I do to solve the XO issue?

Thanks for any suggestions,

~JH
 

TMM

Member
2007-09-01 8:37 am
Australia
I really like the idea of running active because it would mean I would not need to construct 6 complicated individual XOs but I dont know where else to turn.
But now you need 18 channels of DSP and amplification. Sounds expensive. Six MiniDSP 2x4 boards and nine stereo amplifiers would probably be the cheapest way of accomplishing this.

Passive crossovers will be significantly cheaper, but obviously there is a learning curve in designing a passive XO. For first timers I highly recommend cutting your teeth designing a 2-way passive XO first, then attempting a 3-way. In your case you are lucky that your mid can be run fullrange (at low volume) so you could attempt to design and build a 2-way between the mid and tweeter first, then extend it to a 3-way design.
 
Last edited:

Fenalaar

Member
2011-04-25 10:49 pm
MiniDSP 2x4 kits and Hypex UcD3xMP? The amp boards come in 2,4 and 6 channel versions, has an integrated PSU, and the aux power out can be used to power the miniDSP. If you use the HD version of the miniDSP, you'll need a separate PSU for it. The 2x4 kit runs at $80 and the UcD32MP is €80.

Johan-Kr
 
Still, a midrange will emit to full pistonic range even if filtered with 22 uF :rolleyes:
Huh? :confused: I meant that his midrange driver can be run at reduced volume without a high pass filter as opposed to dedicated midrange drivers (e.g. a 2" dome midrange) which cannot. The result is that he could attempt to build a 2-way crossover between the mid and tweeter first to learn how to design a passive XO, before committing to a full passive 3-way design.
 
Last edited:
Huh? :confused: I meant that his midrange driver can be run at reduced volume without a high pass filter

Well, the "full pistonic etc." is to be intended as full power till safety margin :p:eek:
..as I was playing with some 2 or 3 way with 4 Ω speakers one month ago, and 33 uF was on the 4Ω midrange ( guess: a FR from a recent Philips radio/cd with 6 W amplifier ) and it was not enough from preventing it from bottoming.
Still, the classic 22 uF plus L (0.5-2mH :confused: ) would not prevent to modulate in the presence of LF, for a 8Ω speaker with low Xmax (1-2mm :confused: )

...same for a tweeter, ferrofluid being an extra to be considered.

:radar:

Also, the wire thickness in the inductor will work as R as needed
( talking about reduced volume - given a 5/15 W fullrange/midrange with fair efficiency ) you can go well below 0.5 dia mm / 25 AWG for the lowpass > so about 500 turns:confused: around a ferrite core > 10 grams of metal :p will work also as resistor -
Yeah, passive sucks :crazy:
 
Last edited:
But now you need 18 channels of DSP and amplification. Sounds expensive.




Just to be clear the side surrounds of a home theater are not separate channels.

-I thought I could run the 3" 'out of the back door' and not filter highs and use a cap between the 3s and tweeters.

This would mean the surrounds could be handled with 2 channels per side.

Could this be done? (*just guessed on the xo points for examples)

[IMGDEAD]https://scontent.fsjc1-3.fna.fbcdn.net/hphotos-xfp1/t31.0-8/13055225_10208316655170057_3664493412506826556_o.jpg[/IMGDEAD]

Get a MiniDSP (or a PC with multichannel soundcard) and 2 Stereo amps first.
Spend some time learning to design the crossovers with the MiniDSP (or the computer) and some measurement gear. Translate the resulting crossover to a passive design. Repeat that passive design on the rest of the speakers.


I have seen this before; the suggestion to use the MiniDSP to get it tweaked and then build passives.

Why? When you get the MiniDSP set up correctly and where you like it, why not just keep it to handle the XO duties and not build passives?
I have been looking at the mini and Marchand stuff. Cant understand why fix it with passives if it isnt broken with the MiniDSP with XO plugs.




Also is there no 'hand grenade' program that allows a guy to enter a few of the TS prams and have a relatively close recommendation for caps/inductors? I understand a perfect XO would need to be derived from real testing in enclosure, I am just talking ballpark decent.


~JH
 
Last edited:
Hmmm...many recent & old threads about wall wounting speakers - ballpark value would imply to get flat response to π/2 -half space so much less than if BSC required - talking about the LP coil ( which would be rather big for the eventual 200 Hz LP). ((The mid section would benefit from 2nd order I guess))
This is compensated by the number of mids with about 6 dB added ( notice that in this particular design the baffles are acoustically treated ) by double driver on each direction :h_ache: An average 8 " would have 90 dB efficency and each midrange should behave badly as usual---84 dB sensitivity :confused:
 
Last edited:
Hmmm...many recent & old threads about wall wounting speakers - ballpark value would imply to get flat response to π/2 -half space so much less than if BSC required - talking about the LP coil ( which would be rather big for the eventual 200 Hz LP). ((The mid section would benefit from 2nd order I guess))
This is compensated by the number of mids with about 6 dB added ( notice that in this particular design the baffles are acoustically treated ) by double driver on each direction :h_ache: An average 8 " would have 90 dB efficency and each midrange should behave badly as usual---84 dB sensitivity :confused:

Making mine monopoles. ;)

~JH
 
Why? When you get the MiniDSP set up correctly and where you like it, why not just keep it to handle the XO duties and not build passives?
I have been looking at the mini and Marchand stuff. Cant understand why fix it with passives if it isnt broken with the MiniDSP with XO plugs.

~JH

The reason why I suggested it is for rapid prototyping. Personally, I'd almost always opt for active, but doubt I'd use a MiniDSP for that. The reason I suggested it was with the active path with MiniDSP you'd need way more amplifier channels, that's all.
Using a MiniDSP would make for a nice lab setup to dial in crossovers. But we cannot guess how far you want to go with this...

Lots of info on here on the use of software to do the simulations for passives.
 

eriksquires

Member
2013-05-10 4:11 pm
As someone who used to be in the theater sound business (film, not home) I'd like to opine. Just imagine me in a smoking jacket and a big pipe, with a fake British accent for a bit.

A surround like that is best used when you have 4 rows of seating or more in a large room. Any smaller and you'll be too close to the drivers.

If you have a large listening area though you would be better served with multiple simpler surrounds. I don't mean the extra rear surrounds. I mean having 2 pairs or more on the sides, plus those. DSP becomes very helpful then because of the extra delay needed for the second pair. You could do either a simple two way, or something like the ADS design posted here recently
 
I typed out a fairly long reply to this, and then lost it due to a combination of an iPad and public WiFi that kept dropping.

Here goes...

First of all, go and buy some measurement equipment. The Dayton USB mic, or an ECM8000 and a little USB preamp. Learn to use those and REW. This is important. I've had an ECM8000 for a couple of years and its paid for itself many times over. You'll need a measurement setup for this set of speakers, and also the subs you're probably going to build next year.

Next up, get hold of a MiniDSP (basic 2x4 will be plenty), and play with it. Find out how to EQ things. Play around: feed some music into some speakers via the MiniDSP; keep the measurement system up and running at the time, and see what does what. Try using REW's Auto-EQ function to suggest some EQ settings for whatever you've just measured. Dial them into the MiniDSP, and see how that sounds.

Having a measurement setup means you won't be guessing at things. Something sounds sibilant occasionally? Could be a narrow peak anywhere from 6-12kHz. Measurements will tell you where. Guesswork might get you close, but you might end up EQing in a broad dip in that range, when actually a narrow cut at 8kHz is what you need.

Once you've learnt to measure and EQ things (might only take a day or two of messing around), you can look towards this project. I'd go for enough MiniDSP boxes and amplifiers to do seperate processing for bass drivers and mids/tweeters. ie, you're running a passive crossover between the 3" mids and tweeters.

Getting the passive crossover right will be a little tricky, but not impossible. Since you can still EQ the end result, focus more on getting the phase alignment right. That is, you want a nice, smooth phase curve (REW will measure that for you). A smooth phase response means the drivers are working together through the crossover region (as opposed to running out-of-phase, etc), which might well give a reasonable frequency response. Anyway, if you get the phase right but a 3dB rise over the crossover region, that's fine, EQ it out. If the phase is all jagged, chances are you'll get some pretty serious notches in the frequency response, and you'll never be able to EQ those back in, since the drivers are cancelling each other out.

Active crossover with the MiniDSP will be easier than the passive. My usual way of doing it is to take a few measurements, on- and off-axis, take some form of average, EQ that to be flat well past the intended crossover frequency, and then throw a LR4 (or whatever your favourite topology may be) at it. Since the drivers are effectively "flat" well past the crossover point, the acoustic crossover is the same as the electric crossover, making integration much much easier.

I hope some of this has helped. If you decide to go for a passive 3-way, still get yourself a measurement setup. Probably the most useful tool you can have, when used in conjunction with even rudimentary EQ.

Chris
 
A time-honored response to difficult problems :)

...although I had to give it up for health reasons ;)
 

Attachments

  • Ben w Bottle 09-1999.JPG
    Ben w Bottle 09-1999.JPG
    70.7 KB · Views: 49