Active crossovers

Hi,

Has anyone here used a Rane SAC 22 active crossover? My Ruark Prologue One R's are currently biamped through 1 integrated amp and one power amp, I'm moving over to setup without a preamp, and 2 power amps, and at the same time I'm going to go active. Any other active crossovers I should consider?

Cheers.

Paul

 

behringer probably do a stupidly cheap one that has loads of facilities and costs 1/2 as much as an 'audiophile' one - probably has about 75 - 90% of the performance and could be fiddled with to make up the difference for pennies.

 

http://www.behringer.com/CX3400/index.cfm?lang=ENG

http://www.behringer.com/DCX2496/index.cfm?lang=ENG

http://www.samsontech.com/products/productpage.cfm?prodID=100

http://www.bryston.ca/crossel.html

http://www.bss.co.uk/includes/product_sheet_include.aspx?product_id=28

http://www.dbxpro.com/products.htm (they do many different models of 'loudspeaker management systems')


My personal choice if you wanted to stay analogue would be the Bryston for quality or the Samson for cheap and still extremely good.

If you can go digital (benefits can outweigh the fact its digital, especially if you use digital mostly) I'd choose the Behringer DCX2496 or the BSS if you can afford. The Driverack 4800 is a bit OTT.

Also consider that the TacT RCS stuff does crossovers as well with the add-in cards.

 

Thanks for the replies,

Tenson - please may I bombard you with a barrage of questions? Sorry for this, but I'm a bit lost here, you seem like you know a fair bit about the Behringer boxes and the more I look at them the more they seem like they'd be a good solution for me.

The only source in the system is digital (Sonos box). I plan on having no preamp in the system, and just feeding the power amps straight from the Sonos which will be doing digital attenuation. I know some people say this loses dynamic range, but I've heard others say they can't tell the difference and I've never heard of it being ABX'd.

Sorry for being a complete luddite here but does a digital active crossover like the DCX2496 perform the crossover job in a fully digital way - so it takes a digital input from the source and I'd get 2 digital outputs from it one the high pass the other one low pass? Then I'd need 2 DACs right? I'm guessing that I've got totally confused and it's nothing like that, would appreciate being filled in on the details.

I've been searching around and found a few threads about the DEQ2496, I like the idea of room compensation, and I've also read it has a great DAC - I was going to buy a Zhaolou with the AD1842 and BlackGate caps, but if I can do it all in one box and for less money all the better. Can the DEQ be used in a mode where the input is digital, it does the room compensation, then outputs in analogue?

I found a thread about modding the output stage and PSU in the DEQ - It mentioned using S&B transformers - what is a transformer? A relacement for the stock opamps? And what options are there for upgrading the PSU? Ideally if I could get these mods done commercially in the UK for a sensible price I'd go that way.

Thanks...

Paul

 

If you are using the Sonos, the DCX(2496) is ideal. You give it a digital input, it performs the crossover in full digital mode then converts it to multiple analogue outputs that go to your power amps.

The DEQ(2496) is another box of tricks. This is for room correction rather than crossover duties. You can use both if you want. Digital in to the DEQ - digital out - DCX - analogue out - power amps.

There is another option though.

The DCX has some EQ facilities as well. They are less comprehensive than the DEQ's and it doesn't have the automatic set-up for room correction but if you know what EQ is needed to compensate for the room then you can do it here. I use a separate system to measure my room so I just tap the required settings in to whatever EQ I want.

If I were you I would get a DEQ and a DCX. Use the automatic set-up for the room correction on the DEQ and then send its corrected digital output to the DCX. I can show you some mods for the DCX's analogue outputs to give its sound quality a major shot in the arm. If you don't fancy that then send it to me and I'll do it for a small fee.

The DCX does not have digital outputs for using 2 or 3 separate DAC’s, but you won’t need it with the quality available from the modded version.

Digital attenuation is not something I am fond of, but I would presume the Sonos does it in 24bit rather than 16bit (thats what Wadia and people do so you don't loos dynamics). If that is the case then great, the DCX and DEQ can handle 24bit signals. If not, then it will be no worse than it is now.

 

Oh, the S&B transformers in this case are used as a volume control of very high quality. You shouldn't worry about them for use with the DCX as it will cost nearly £1k just for those. I'd stick with your digital attenuation.

If you find it does effect the quality doing it that way though, have a look here at the '6 channel volume control' at the bottom of the page. It can go after the DCX so you can use them with full level digital signal then attenuate in analogue before the power-amps. - http://www.selectronic.fr/international.asp

 

STOP!

Your are putting the cart before the horse. You are considering solutions to a problem, but it's clear from your questions that you don't really understand the problem.

Ripping the passive crossover out of a set of speakers and sliding in an active crossover can make matters worse rather than better. To understand why this is the case, you need to know a lot more about speaker design.

Before you spend quite a lot of money, and potentially screw up a functioning pair of speaker, I suggest you read the introductory text on the topic.:

Loudspeaker Design Cookbook
Vance Dickason
ISBN: 1882580338

[Tenson, you should know better than to simply encourage him..]

 

I think Tenson's recomendations are good enough to replace any budget speaker analog xover, even too good...

 

I guess that you are right about needing to know a little more. The man asked for equipment recommendations and I gave some.

However, the speaker is not a complex one. Its a two way that Xovers at 2KHz. I would make a guess that they use pretty wide bandwidth driers (the Talisman does, a Seas bass driver, looks like the CA18RNX). That means that with a bit of info from Ruark, or by looking at the current passive Xover it will be very easy to duplicate the required settings and then even play around to improve on them. I doubt it has any notch filters, maybe a Zobel but that isn’t really needed in an active system. Basically the crossover should be very straightforward.

Edit, maybe it would just be easier to buy some active speakers though!! I think the Tannoy Ellipse 8's or 10's are IDEAL for this type of thing. You can sell you power amps and all that. these even take a digital input and I think have some level of room correction built in as well (oops no it doesn't do room correction).

http://www.tannoy.com/Ellipse8iDP

 

Oedipus,

Thanks for the words of caution - I'll look a little deeper before I start making changes, and I intend to do it in a reversible way anyway, so the speakers can be sold in future if desired.

Tenson,

I'm going to go with a DEQ and DCX I think, thanks for the offer of doing the mods on the DCX, very kind of you. I'm not entirely incapable with a soldering iron but on the whole I think it's probably better left to someone who's done it before. Can you tell me roughly how much it'll cost me all in for parts and your time?

I have posted on the Sonos forums and hopefully will be able to find out if the attenuation is done in 24 bit, otherwise I may look at some form of analogue attenuation.

I don't really want to change speakers - I rather like the Ruarks for both sound and looks. I'll start researching crossover design a bit, but I'd appreciate a pointer or two on what I need to ask Ruark to find out how to replicate the passive crossovers settings, I know the cutoff is 2KHz, I presume I'll need to know the rolloff rate on both sides of this.

Will the low bass and high treble also be rolled off in the crossover (eg. low pass at 20ish KHz and high pass at 20ish Hz)? What else is relevant?

I also use a REL Storm III sub, will I be able to make any improvements by running this from one of the XLR outputs on the DCX, or is the filtering in the sub likely to be good enough for it's purpose? (It currently runs in parallel with the mids in the main speakers, not sure how, some form of attenuation back to line level before using the subs internal amp presumably).

Thanks again, when I drop the DCX off to you I think I already owe you a beer!

Cheers

Paul

 

Hi Paul,

What you need to know from Ruark is the frequency of the high pass filter and its 'slope' (the rate of roll-off) and the frequency of the low-pass and its rate of roll-off. You also need to know if there are any 'pads' (matches the level of the drivers) and any 'notch filters' (a bit of EQ if there is a peak in the response). You should also ask if there is a ‘Zobel network’ some manufactures attach the caps (capacitors) for these directly across the back of the driver.

There is an advantage to running the sub from one of the outputs of the DCX, you can stop the main speakers having to reproduce frequencies below where the sub kicks in and you can also time-align the sub to the main speakers using the delay options (done by the difference in distance of the main speakers and the sub to the listener).

I'll send you a PM about the mods.

 

Let's be honest here, if bitmonkey wants to fiddle around a bit and have a bit of "fun" diddling with someone elses speaker design then that is fair enough.

However, the idea that it is possible to "play around to improve on them." is somewhat preposterous, because whoever did the design at Ruark probably:

1) had done this sort of thing before and understood the compromises he (or she) made in the design of the speaker as a whole

2) had access to several thosand $$$ worth of test equipment

3) had plenty of time to do this because they were being paid to do it.

Now, if you've not designed a speaker, you're losing on (1). On (2), the tools are prohibitively expensive [to rent], because for a halfway decent design you need an anechoic chamber to design a speaker - if you don't have access to such, then you can fall into the trap of design a speaker/room combo because the room is convolved into your measurements. On (3), you're competing with someone with 40 hours a week to kill..

Coming back to point (1), if you start out with an active design in mind, then you make other choices in the design (like drivers and baffle placement and on and on...)

An excellent suggestion

The again, you could just buy some passive speakers where the issues have been nailed by a really competent design team (eg Paradigm, Revel..). It's unlikely you could improve upon their efforts at home in any meaningful way - no matter how you (or I, or bitmonkey) went about it...

 

It depends what you consider improvement (subjective or objective). Bitmonkey would have a starting point of where Ruark left off, which is probably a very good place to be at. He can then tune to suit his personal taste and even do things that were not possible in the passive design. For example he could use the tweeter to a lower frequency by using an 8th order slope. He can time align the drivers. He can cut low frequencies the sub handles.

Whatever happens, it will be no worse than they are now with the added bonus of being a digital Xover.

If he didn’t know what the crossover settings are at the moment then it may be more of a problem but I’m sure Ruark can tell him. Personally I would get a pair of those Tannoy's in the first place but if Paul likes the Ruarks and doesn't want to change thats his decision.

Paul you should be very careful not to blow up tweeters though You don't want to send them the subwoofer output by mistake!

 

Ruark might have all those resources, but they also have a budget that doesnt include versatile active filters in that particular model...

 

Thanks again everyone - I'm persuaded enough by the various arguments to be trying to arrange a demo on some PMC active speakers (AML1s and TB2S-As), and will go from there.

I've been told it's difficult to get demos on AML1s, anyone here within say 50 miles of Hastings got a pair and wouldn't mind giving me a demo if I bring the beer?

cheers.

Paul

 

The TB2S-A are not 'active' they are 'activated' which is basically a passive speaker with am amp glued to the back.

You can demo the AML1 at KMR Audio in London. http://www.kmraudio.com/catalogue/index.php Also have a listen to the MEG RL901K if you go along for an ear opener (go round the back of them, notice how the sound changes).

The room really is not suited to the AML1 though so get a home demo, they are very kind folks I bet they would let you. Kub3 (also named Paul) on here has a pair of AML1's, I think he is pretty close to you.

I'd love to hear a pair of the Tannoy Ellipse as I bet they are smashing but I don't know where to demo them.

 

Things could easily be worse. It is quite likely that the transfer function of the current crossover ranges doesn't map well onto the Behringer (a simple example is the baffle step). The point is that the Behringer isn't a convolver for arbitrary transfer functions, but allows the user to apply a fixed set of curves. In combination with the Behringer DEQ, more can be accomplished, but even then it can be fiddly to replicate the transfer fuctions needed by a real speaker.

If you change the crossover point and slopes, the power response of the speaker changes in 3 dimensional space. Now, you are suggesting using really steep slopes, but that leads to the issue of a "step change" in the directivity index (DI) - which happens if the drivers have different polar responses. This kind of change in the DI has been shown in subjective testing to be a bad thing.. (Read Floyd Toole's papers..)

The nail in the coffin for redoing the (active) crossover at home - and doing a half decent job of it - is measuring and controlling the power response of the speaker.

After all, at home you're not in the free field, but are listening in the reverberant field.

 

Why would the baffle step be any different when using a digital crossover? If the current crossover has baffle step compensation that can easily be duplicated on the digital platform.

The only possible problem I can see is if there is no impedance correction and they just use the slope of the crossover to counteract the change in roll-off due to that. You use the same slope in the digital crossover and it is too steep or shallow. The chances are though, if they don't have impedance correction they are just no crossing in that area.

Crossovers can be swapped from active to passive very easily, I have done it swapping from a digital crossover I used to design the system then implementing that same crossover in passive. It worked perfectly. It should go the other way well enough as well.

I think you will find that it is far easier to get the required crossover settings for a speaker on the DCX than it is in a passive crossover.

 

Following on from the initial discussion of using the Sonos digital volume attenuation, I have confirmed with Sonos that this is done in 24 bits, and that the digital output is 24 bit.

I am curious though - can someone give me a fairly basic explanation of why attenuating in 24 bits doesn't kill the dynamic range, but doing it in 16 bits does. Also, I'd appreciate it if someone could point me to an introductory text which would give me a bit more understanding of digital encoding and processing.

Paul

 

CD is 16bit, 24bit has a much greater dynamic range. So if you convert the 16bit signal to 24bit you can attenuate the signal and still have a full 16bit or more dynamic range. You can only do it to a certain point though and you start loosing dynamics still, but its a lot better.