Can we talk active x-overs?

[zimjj]

I'm seriously considering designing and building a 3-way (specifics to be discussed later) and am considering an active x-over. Why? Because to a neophyte like me it seems the easy way out, I have an unused 7-channel amp and all I need to do is buy the crossover. Or so it seems!

You get an instant 4th order L-W, just dial in the x-over frequencies. It's automatically phase aligned, easy to do time correction and you have a built in equalizer to even things out. Piece of cake...or is it?

There's the Behringer digital x-over - relatively inexpensive; the Rane analog- it looks very high quality - more expensive and they suggest you buy their scope and RTA. Then there's the Danish Ground Sound digital - very expensive- even as a kit. EVS will make over your Behringer for 600 plus bucks which will still come in under the Ground Sound.

So I'm looking to all you experienced guys to tell what I'm getting myself into.

Thanks,

JohnZ

[kimbo]

Going active certainly saves you the pain and stress of developing a passive crossover. As is stated many times here, crossover design is no trivial mater and always involves a lot of tweaking once the actual circuit has been developed. It is for this reason, and being a relative novice that I went active. I settled on using the Marchand XM1 http://www.marchandelec.com/ftp/xm1man.pdf with the stock LF353 chips and it was OK at first. However I finally upgraded to OPA 2134 as recommended by Phil Marchand and what a difference!!….better everything especially the sound stage. ( I am also considering upgrading the OPAs one day.) I was also able to very simply change the xover freq until I found what actually sounded best just by swapping around the little resistor plug. Nice.

I have not used a Behringer….the ability to change a number many variables on the fly is appealing….but I'd rather keep a simple analogue signal path and just use a handful of OPA's rather than digital, unless I got my hands on a really high end digital x-over……but I really don't know if there is a difference

You may find Rod Elliott's article interesting too if you haven't already seen it … http://sound.westhost.com/biamp-vs-passive.htm …. although I don't necessarily agree with all his conclusions.

Anyway, the long and short of it was that the XM1's are one route you should investigate. They sound just fine in my very modest vifa system and Phil was a pleasure to deal with. I'm sure you'll get a lot of responses from other members much more knowledgeable than me.

Good luck!

[Dr.EM]

I am also going active with my system, basically to avoid the complications with passive and also it just seems a more refined of doing it to me.

I also agree with kimbo in that I'd prefer just a simple analogue signal chain to a digital system, but realistically I doubt the ADC/DAC
conversions make much difference.

I will be using the ESP (Rod Elliot) PCBs for my crossover, along with additional circuitry (notch filter, time correction) on daughterboards. Perhaps not ideal, but it should work fine. I'll also employ OPA2134.

I am using a breadboarded crossover to determine my final values before I commit to PCB:

[gainphile]

Active XO are:

- Cheaper, no need to buy expensive inductor and exotic caps

- Sound better, no need to deal with impedance variation, thermal variation, passive loss etc. etc.

- So flexible. 24db easy... notch correction simple!... time/phase alignment, linkwitz transform... oooh so many to choose from.

- Best coupling between amp and loudspeaker is a wire.


It used to be that the single disadvantage of active xo was the cost of amps. No longer with the rise of Gainclones.

This 3-way 24db/oct XO for example.. can be had for about $25 including the transformer/psu. Try that with passive!

You get an instant 4th order L-W, just dial in the x-over frequencies. It's automatically phase aligned, easy to do time correction and you have a built in equalizer to even things out. Piece of cake...or is it?

The difficult part is not to build the xo itself but knowing where to cross, frequency response, etc. This is where you'd need a measurement tool. Or just go to Zaph's site.

[kimbo]

Here is a view of the XM1 boards in a 1U case.
I used the Marchand powersupply board but as I wanted to squeeze it into the 1U, I had to use a torroid as you can see.

Also on the rear pannel I have a pair of dual pots for adjusting the relative levels between high and low to compensate for the different gains of the pair of amps. Once I knew which is lowest I bypassed that ......and once you have lived with the crossover for a while you can bypass the other with a pad of appropriate value metal film resistors.

[5th element]

This always seems to crop up every now and again, people talking about the panacea of active crossovers.

This is simply not true.

Designing a loudspeaker around an active crossover does give you more flexibility then around a passive one. However the final goal requires a similar design approach, it's just easier to achieve with the active version.

You get an instant 4th order L-W, just dial in the x-over frequencies.

You get an instant electrical 4th order L-W, but what of the drivers frequency response? That still has be to added to the filters transfer function.

You'll see Zaph using 2nd/3rd order electrical filters to arrive at a 4th order acoustic target, this doesn't magically disappear when you go active.

It's automatically phase aligned, easy to do time correction and you have a built in equalizer to even things out. Piece of cake...or is it?

Nothing is 'automatic', unless you've got some DSP box of tricks that will measure everything for you and apply a suitable correction filter.

But once again, the active filter will have it's own phase characteristics, as will the driver, and you have to combine the two.

Time correction is easy to implement, but the same goes for any active filter.

It is true that with an active crossover the results are more predictable and guesswork will often end up closer to a desired target. But the bottom line is still the same, you need to take measurements of the drive units in the cabinets they will be used in to make the most out of them.

[col]

Not to mention dealing with the ground loops and amp hiss that are inherent to plugging all those additional amps, power cables and signal cables in. Watch out!

I run a tri-amp Tripath system with ESP P09b crossover, Allen & Heath 4 zone mixer/preamp, Citypulse DAC, poppulse digital interface and a debian linux box powering a pair of "cousin of" Earl Geddes wave guides mid-high and a pair of 6th order bandpass subs, no way was it an easy solution.


col.

[blank527]

http://www.hypex.nl/

Just introduced the AS2.100 DSP Active Loudspeaker Controller.

http://www.bmm-electronics.com/Produ...roduct_ID=7738

For the price

Best regards,
Remco

[kimbo]

The XM1 always has the high and low pass filters in phase with each other.....easy to see using CRO. It can also be implemented as a lower order filter with some parts substitution.

But yes, if I had the skills and most importantly time to try and sort out a passive solution, I'm sure it could be as good or better solution...........but in my case this was and is, the best solution for me as I had built & designed the speakers as a total novice and needed an easy, relatively fast and easy way to put away the iron and start listening. I'm happy!

[Conrad Hoffman]

Second the recommendation of Marchand. IMO, the maximum performance achievable with an active crossover is often greater than could be done passive no matter how much time you spend on it. Hooking a low impedance amplifier directly to each driver is a different situation than interposing something between them and can never be equivalent. After living with a tri-amped system for a long time, I could never go back to passive. The Marchand uses header blocks for the resistors, so you can change the tuning easily. Combined with some gain pots, it makes optimization pretty easy, at least by comparison with tweaking a passive design. Yes, you have to consider the driver curves, but with fourth order it's less of an issue than with common lower-order passives. You should also appreciate what a lousy component an inductor is in terms of being a pure reactance. With an active crossover you eliminate that entire issue.