Is there a reason more of you aren't running active? - Page 3

The majority of crossover discussion on the web is about ideal crossover types. These are exactly what active crossovers allow you to implement more directly. The tricky part that most discussion likes to skim over is getting the driver response to actually fit the theoretical crossover types. This can be especially difficult in the passive realm for those not very familiar with passive circuit analysis. It is much easier in the active realm, where you only have to understand the acoustic targets and much less so how to achieve it.

The choice of what frequency and what type of crossovers to use is not an exact science, but rather an applied science. Ask 10 professionals, and you should get 10 varied answers, and most of those answers will start with "it depends...". The real key is understanding how to take measurements and what your measurements do and don't tell you. Crossover design without real measurement is a shot in the dark. You will know you are starting to get somewhere when you start fretting about how to get anything close to a full range measurement without an anechoic chamber. In general the answer is usually by combining a few measurement types along with a lot of head scratching and experimentation.

I fully agree, with the main limitation of available power. In fact this is already a reality, and I met with an interesting technology company who has already done a surround preamp with what is effectively HDMI inputs and an 8 port Ethernet switch on its back panel. They make the receiving chips that can be added to the input of an amplifier and can even house the DSP curves for a bi-amped speaker in the main processor sent out to each amp. So at that point you don't even need any more than the receiving and 2 amplifier channels in a speaker. You can easily move to 3+ way systems by executing the crossover in the speaker, which could be controlled over the same Ethernet connection. Look at Crown's HiQ Net system for a simple example of how something like this could be adapted for home theater with the right software/hardware combo. Thiel briefly offered a product using this technology but has gone back to mostly passive systems due to dealers not warming up to powered speakers as they don't get to sell expensive amplifiers. I'm hopeful this sort of possibility will be something I can consider offering in the next 3-5 years as a fully packed system. Fingers crossed.

Marty, it's not something easily answered in a single post, which is why there's a lot of confusion about the issue I think. Mark pretty much summed it up about as good as one can in a single post. If you hunt around the web, you'll find discussions about where to cross over a tweeter or a woofer. Steep versus shallow filters. Etc. Delay is like phase aligment, it's applicable to discussions about phase.

As Mark mentioned, the single most important part (learned this the hard way, don't believe the guys who "simulate" their designs) about getting an active or passive cross over done right, is measurements. When I stopped screwing around and learned how to take proper measurements, all the theory banging around in my head just clicked into place and my speaker designs actually started to sound good. The flip side is knowing what those measurements mean.

On pro-sound forums. This may be still in the realm of black magic for consumers, but it's been SOP in pro-sound for over 20 years.

Going back to the original post by Bass Addict, he wanted to know why more of the guys on AVS didn't design their own active crossovers. The reason (beyond cost, which is a big issue) is that these guys are not interested in learning to design a crossover properly. Of course, you could get a DCX and tune it by ear but that will sound like crap. Or you could take measurements and tweak until it is flat. That is better than the "ear tweaker" but there are still many more principles and subtelties that, IMO, if not understood will wipe out any gain from going active DSP.

Now, I'm not saying it is some black art that takes 10,000 hours to master. But you can't just jump in, tweak some knobs and end up with a better result than a well-designed passive design which is already available.

If it was as easy as buy DSP and click buttons, than Mark Seaton really has nothing on any of us. Buy a pro coax, add two pro woofers, tweak Digimoda and become a zillionaire!!!

The fact is that most AVSers are amateur woodworkers willing to do some basic soldering. Not many are interested in passive OR active DSP design. A few express interest in active DSP because it seems like a shortcut but it is not a replacement for understanding the principles and methods of speaker design. Greg Begland, Tux, Mark Seaton etc understand this stuff and enjoy the design process.

IMO, this is what it boils down to:

If you want to build a box and get a great value on a DIY speaker, passive is likely the best option.
If you want to design your own crossovers, active DSP is an excellent option.
If you are looking for a shortcut to building a speaker that doesn't presently have a crossover design, there is none. DSP, poorly executed is a waste of time.

---

Gbegland, you don't have to swap components 40-50x. In fact, you rarely have to swap more than a few components to tweak. I typically use LspCAD and can simulate enough to not need to do more than tweak a few components on final listen. I simulate when I design actives too. Tweaking the DSP and measuring over and over is a waste of time.

---

IMO, the best use is a hybrid approach which I think Mr. Seaton mentioned. Take a 2-way passive and make it a 3-way with active DSP. This way you don't have the huge inductor cost of crossovers in the 200-300hz range, you can tweak room issues in the midbass range, do whatever smoothing you can't do with a passive, and tweak the HF tilt depending on the room and speaker interaction.

This can also be done with a 2-way and minimal passive components. It saves you an amp channel and gives you nearly all of the benefits. This might be the best route for many. Maybe some SEOS designs with a minimal passive and some DSP settings and guidance on how to tweak for room and taste.

I'm working on a 3-way SEOS where I plan to offer an all passive and an active/passive cross overs. I'm pretty excited about it. I don't need another speaker. I'm just addicted to SEOS

There's a few of us around. It's worse than meth

Just to be clear... I'm not advocating for the average DIYer to design it. I'm just speaking from the standpoint of someone designing a loudspeaker that I'd much rather design an active one than to live with the limitations of passive design. Also...I'm using all analog filter design so I'm not even talking about DSPs. I have nothing against a DSP and the resulting ADC-DAC conversions. I just don't have any experience using them at the board level and I'm very comfortable designing with opamps.

Also... in my mind the DSP should be in the HT processor. The consumer needs it primarily for room issues, not to design crossovers between transducers. The amount of knowledge and measurement skill needed to design proper crossovers is always going to be well outside of the range of the average or even advanced enthusiast. That is something that should remain fixed by the manufacture of the device. Give the consumer some tone controls if they want to play around with how it sounds but the crossover transfer function is not something where that ability to tinker adds any value to the final solution.

Yes, I see no reason to do completely passive 3-ways that have a fairly low cross on the woofer. There are also other speaker types that don't make sense to ever do with a passive simply due to them needing excessive components, but that doesn't include typical dome tweet mid two ways and horn CD pro woofer designs. I'd also still take higher end drivers over active DSP if the budget dictates one of the other.

+1. Unfortunately for Tux I've decided to pursue the art of crossover design and I've probably asked at least 1,338 questions in the last week - all of which he has answered thoroughly and quickly, to which I'm very appreciative! I'll soon (tomorrow sometime) have the ability to measure impedance of my drivers and will have my own inputs to PCD. I'll probably get a build thread started as well shortly using the Selenium woofs and DNA350/SEOS12.

Apart from sub builds which are far and away the majority of DIY HT builds, most people effectively wantt a kit. It may not all come in one box, but selected drivers, an already designeed xover and a BOM and an enclosure spec. So design aspects are entirely irrelevant for the majority of builders whether active or passive because most will take no part in any aspect of the design.

There is no reason whatsover that it could not be done as an active option, where a SEOS designer also could noot do a design in active and offer a setting file for MiniDSP and they could use an older AVR for 6 channels of power. Or the designer specs it for 2 Emotiva XPA3 (OTTOMH example) for a 3 x 2 way. Even if a builder chose to use different amps with a pre obyained setting file, at worst you'd need to do some channel balancing to adjust for different gains. Not difficult to do even with an RS meter. If they already have a decent mic and REW, it's a doddle.

The Nao Note is offered this way from memory and I see tuxedo is offering similar for a 3 way.

As for amps popping at turn on (from earlier post I forgot to reply to), I have 20 different ones in my rack between 20 and 2 years old; not one pops on turn on/off. For the trivial power draw of a MiniDSP or two, leave them on all the time, so neither will they.

Wow. Ready to drop any hints as to what shape this will take? I've been months agonizing over which of the current designs to build, all the while knowing I'd be most interested in a 3-way that I didn't expect anyone to ever want to design.

If nothing else can you say whether or not this is expected to be a super-premium "crazy" design along the lines of what ChopShop seems to be cooking up? Because if it is, I'm sadly out and better commit to the nicest 2-way available that fits my needs.

Ya, nothing to hide. Don't have all the details worked out yet. But it'll be a SEOSXX / DNA-205 (maybe 350 or 360) over a pair of side by side 5" high sens mids and 2 woofers. I'd like to make the woofer section very flexible. Passive, active, 10", 12", one woofer, two woofers, flanking for centers, WTMW, TMWW. All kinds of variations. It's why I told Chop Shop to make his build with modular cabs. Because that's what I'm doing and it's really paying off. In the end I'd like someone to be able to choose the orientation that fits their room, screen, and setup. It's going to be a lot of fun. I think in terms of budget it'll hang around $400 - 500/channel. No cheap way to do a 3 way like this. But there may be a cheap woofer option that would get it down below $400. We'll see.

It'll target the person with a healthy budget, needs flexibility, and wants a 3 way because a 12" midrange bothers them. It'll even likely end up using the SEOS10 to get it narrow enough for those builders also.

Cabs are pretty much built. Just experimenting with ideas and getting parts in and measuring. More solid details will probably pop up on diysoundgroup in a couple weeks or so. It's a beast to get my head around. Not quite as simple as a regular 2-way. Maybe the whole thing wil be a total fail.

Thanks for sharing, tux. That's a very reasonable price point you're targeting for such an ambitious speaker.

The reason I'm in your target market isn't that a 12" mid bothers me, it's because I worry that big 2-way monitors are apt to suffer floor bounce issues when stuck up on stands or converted into towers. I think it may be one of the underestimated benefits of a 3-way design, that it's possible to keep the woofer near the floor and your design axis up near ear level at the same time.

Looking forward to see what you come up with!

Anti, are you looking at active or passive? If active the choice of drivers makes it considerably easier to select a workable combination. My surrounds are like this, but all AE and B&C drivers, so above you intended price bracket. A 15" sealed or ported, an 8 or 10" mid, SEOS12 and CD will come together very well with well chosen drivers. I've built a few in this format (and it was a large part of my modular PA set up) and there was little difference with the basic xover settings between them.

If you're going passive, I can't help, only mentioned it as an 'in case you were considering'.

Thanks for the tips Alpha Niner, but I'm looking for a passive solution - and lacking measurement equipment or skills or experience I'll almost certainly be building someone else's fully developed design handed to me on a silver platter.

No worries. Best of luck.

Just an alternative opinion, albeit out of context at this point. I wrote the late, great Zilch about 2 years ago asking about going active with my JBL's. His response was, "unless you're trying to fix a known problem, it's not worth it." Again, I'm not sure of his context at that point. I will say, since going active, the palpable impact is much greater!

I hear ya A9X, except there is also the cost issue. I'm not saying active DSP designs are a bad idea. I'm just saying that the market is far smaller. Most of the SEOS kits have around $30 worth of crossover parts per speaker. So $60. Add in either an entire amp or half an amp if the person is already using separates and it nearly doubles the cost of most kits. The SEOS kits are ~$300-400 for all parts including a CNC'd baffle. The MiniDSP is $150 and a single quality amp channel is at least $150.

For a stereo setup, the passive SEOS kit would cost around $700. The same kit with active DSP would be around $1100-1500 depending on if you are assuming the person already had 2 channels of separate amplification or if you comparing to someone using an AVR. That is a pretty big jump in price. I bet the passive version would outsell the DSP version 10:1.

The Nao Note is a far more expensive project. He also offers a version that is passive.

For something like ChopShop's BMS uber 3-way build with the BMS coax I think going passive on the coax mainly for dividing and time alignment and DSP to divide the coax and woofers and EQ the response.

I do think you have a point that kits (likely high end driver versions) could be offered with active DSP spec's and instructions. I know I sound like I'm anti-active, but I'm all for active DSP in the right situation and done properly. It is just not a quick and easy shortcut around knowing how to design a speaker. I've received so many PMs where the person assumes it will be a shortcut to easy speaker design that I should put together a form letter response.

The SEOS designs do come with active settings. At least all the ones I've done do. I only know of two people who has used my settings. So I'd say 10:1 is pretty close. And one of those users eventually implemented the passive cross over to simplify his setup.

So passives + decent amplifier = $210 for a stereo 2-way configuration with shallow slopes and no time-alignment. I would argue that a more advanced xover with 4th order slopes and EQ would easily be in the $250 price range...include the cost of the amplifier and you're looking at a ~$400 solution.

I do think it would be possible to come up with an analog active xover + 4 channels of amplification in the $300 price range. It's basically two nice $150 amplifiers with some extra front-end circuitry that does the same thing as your passives. In kit form, I think that could be achieved at the same $200 price point of the passive solution.

A good digital active solution brings steep slopes, lower distortion, more EQ, and time-alignment to the table. I don't think it's fair to compare that to the simple passive xover because there will be a very real difference in performance. Is that difference worth the price of entry? That's where this gets very subjective...



I don't know if it will ever come to fruition, but as a thought exercise I'm designing a 2in/4out "active amplifier" that could be sold for $500 (think four channel Crown XTi, but way better GUI, audiophile amplifiers, and consumer compatible gain structure). Using your numbers, that's a $300 upgrade to get steeper slopes, better voicing (more EQ), lower distortion, and time-alignment. For me, those improvements are totally worth the $300. But compared to the same transfer function with a best possible passive, it's more of a $100 difference. I would gladly postpone a project a few weeks to raise that extra cash...and the distortion difference gets bigger as you add more components to the passive xover.

Any chance you've run a TD12X and SEOS-12 in your designs actively? I would give it a shot if settings were available.

No I haven't. But I think Bill did. He probably has settings available.

"a more advanced xover with 4th order slopes and EQ would easily be in the $250 price range"

Except I'm not sure any of these speakers would be better off with 4th order electrical slopes and they do include some "eq". The crossovers weren't designed to be cheap. Two-way constant directivity horns with 12" woofer passives have no use for steeper slopes. Also, the "time alignment" is done via the passives and the SEOS kits I've seen have phase under control. Sure you could be slightly more precise with a DSP, but that benefit is tiny in this case. There are certainly other kinds of speakers which demand more extensive crossovers, but not this type of speaker.

IOW, I would use very similar transfer functions with a DSP as these passives use therefore the comparison is fair. A $700/pr kit would inflate to at least $1100-1500 depending on if you are adding 2 amp channels or 4 amp channels for guys using AVRs.

Steep slopes are not always useful.
Time alignment can be done with a passive in most cases.
DSP does have the EQ advantage...but the main advantage is in cleaning up the room mode issues in the midbass area and requires each user to apply their own EQ.

Yes, DSP has greater potential, but the minimum 50-100% price increase is the reason why it is much less viable for the kits most AVS guys want. That extra cost is a game changer.

I don't follow what you are saying about an active analog crossover with 4 channels of amplification for $300. Can you even get 4 channels of decent amplification for $300 anywhere? Would this active analog crossover have any eq and delay? Would it be designed for a specific speaker or would it be some filter and PEQ knobs (like a CX3400)?

A 2in/4out DSP with 4 channels of amplification for $500? That sounds great. Are you selling this? That would undercut Behringer, Crown and Peavey by about 30-50%. It would undercut MiniDSP's plate amps by 75%. It would certainly be worth it on some of the stuff I do because I'm using higher end drivers for some projects and then the cost of DSP starts to make sense.

Look, I'm not saying DSP is always a waste, I'm simply answering the OP's question of why people who are buying $700/pr passive kits (that includes all drivers, horns, passive components and router'd baffle...so another $100 of wood and misc) don't go active DSP. If active DSP settings were included, most would still go passive because cost would increase 50-100%.

Do you have any details on a setup like this? I for one would definitely be interested in this. How much power can this amp of your put out?

Nope, just the passive crossover schematic. I asked the other day.

Not sure if the TD12X is much different than the TD12M, but try these if you'd like.

http://www.diysoundgroup.com/forum/index.php?topic=74.msg1923#msg1923

I'll be running active.

So there!

Like I mentioned, it's more of a design experiment to see if it can even be done at a low price point. A straightforward approach would consist of +/-15V rails, which should be good for 20Vrms with a bridged output stage. That's 50W into 8ohms or 100W into 4 ohms.

That's a bit weak sauce for a 15" going down to 40Hz or so, but way plenty for a compression driver tweeter. The cool thing about the amplifier topology we designed several years ago is you can just change the power supply rails to hit whatever output power you want. The original prototype would do 250W easy sauce and the idea can be scaled to 30kW if you had a power source that could deliver it (obviously that would be crazy expensive).