Is crossover design really this easy?

[leesonic]

I'm making an MTM design speaker, using some Bravox woofers WFR-06BV03B PE#299-061 (or, the blue 6.5" 12 ohms to make it easier), and some Dayton tweeters DC28FT-8 PE#275-076 (or truncated 1-1/8" silkies).

I entered the specs of these drivers in my trusty old X-over 2.0 program that came with BassBox 5.1 (yes, it still runs under XP, still does what I want it to, so why upgrade?) For 2-way, 2nd order with a crossover frequency of 3kHz, it spits out 4.69uF and 0.60mH for the tweeter components, and 6.25uF and 0.45mH for the woofers (2 x 12 ohm in parallel).

Is that it? It that what crossover design is all about? Please forgive my ignorance. There are other options such as Butterworth, Bessel, Chebychev, and Linkwitz-Reily which give slightly different numbers. Are any of these better or worse? Am I doing a good thing here trying to design my own, or should I just try a set of generic crossovers I have lying around?

Thanks,

Lee.

[johnnyrichards]

Short answer: no, it isn't.

FWIW, I used that tweeter in a Bravox project (the 12 ohm 5.5" silver cone). It is a miserable excuse for a tweeter (sorry PE). The project was named "The Ocho" and I went through eight or nine difference crossover topologies and a variety of crossover points. No go for me with that tweeter.

[Mark65]

The short answer to the headline question is: No.

Longer answer: Hell, NO!


:D


Xover Pro only really takes the drivers TS Params into account, and while better than pure textbook, it's not by a whole lot.

That Bravox driver is nowhere near that easy to work with, in particular. I've not actually heard it, but using a massaged in-box measurement from Johnny Richards, I had a LOT of trouble trying to come up with a decent cross to a HiVi SD1.1 tweeter. Not the same tweeter, I know, but I'd say that you need good reliable measurements to come up with a good crossover with that combo.

One thing in particular that leaps out at me is that lack of BSC in that xo. Placed in a bookshelf, the bass might be acceptable, but otherwise, it will sound thin.


HTH,


Mark

[patrickm]

well, that will give you a crossover, but it may or may not sound good, and because it's basically using textbook crossover formulas (which make assumptions, among others that the driver has a flat FR and impedance, neither of which are true) the odds are unfortunately that it won't sound very good. plus, those calculators do not usually take into account the difference in SPL between the speakers.

the bravox woofers impedance is nowhere close to constant -- the spike was over 50 ohms on some i measured... so you'll also need some additional components to compensate for that on this particular woofer.

if you're not set on that tweeter, i have an MTM designs below with the bravox woofers and either the peerless india aluminum dome buyout, or a vifa BC25SC06 (the peerless is the better of the two designs) and there's at least one other one on the boards.

[Chris Roemer]

Short answer: no, it isn't.

FWIW, I used that tweeter in a Bravox project (the 12 ohm 5.5" silver cone). It is a miserable excuse for a tweeter (sorry PE). The project was named "The Ocho" and I went through eight or nine difference crossover topologies and a variety of crossover points. No go for me with that tweeter.

Johnny might not like your tweeter, but a LOT of DIYers do. That tweeter (or a varient of it) has been used in MANY good sounding projects, including Curt's TriTrix. Your method of XO design is not a good one, however. You have to use a program that can simulate using actual frequency response curves and impedance profiles for your drivers.

Chris

[r-carpenter]

I think that what X-over pro does not take in to the account is the actual phase of the drivers, mounted in the box. It can't. So, no HBT and hello 5db peaks and deeps at the Fcrossover. It's an ok starting point but that's about it.
Crossover design isn't easy and very time consuming. It can also be frustrating, confusing, insulting, irritating, tiring, mind boggling, humiliating, suicidal, drinking inducing, sleep depraving and so on. :)

[mike price]

I made the mistake of designing my first crossover with X-over pro , it wasn't very good. All the guys here on the forum told me it wouldn't work , but I went ahead and did it anyway figuring the software couldn't be that far off. It was.

[johnnyrichards]

Johnny might not like your tweeter, but a LOT of DIYers do. That tweeter (or a varient of it) has been used in MANY good sounding projects, including Curt's TriTrix. Your method of XO design is not a good one, however. You have to use a program that can simulate using actual frequency response curves and impedance profiles for your drivers.

Chris

I have heard the version used in the BR-1 and Tritrix, and still didn't love it - but it was pretty much mostly an indifference at that point. The shielded, truncated version is just a horrible tweeter :( I tried, I really did... but at some point, a designer has to ask himself "Is it worth it to pursue this, or not?" With the Dayton "silkie" in truncated form, my answer was "no, let us look at something else instead".

The above is why ~ 1/2 of my designs never ever make it to the point of a TT "Look what I am working on!" thread.

In any event, yes - using impedance files only, the best result you can expect is something that does not start your amp on fire :)

[johnnyrichards]

Crossover design isn't easy and very time consuming. It can also be frustrating, confusing, insulting, irritating, tiring, mind boggling, humiliating, suicidal, drinking inducing, sleep depraving and so on. :)

+1,000,000 unless you luck out and mate a couple drivers that just work (http://gonecatfishin.net/Rompicollo.pdf), yes - all that you said and then some. Particularly the drink inducing part ;)

[leesonic]

FWIW, I used that tweeter in a Bravox project (the 12 ohm 5.5" silver cone). It is a miserable excuse for a tweeter (sorry PE).


Johnny might not like your tweeter, but a LOT of DIYers do. That tweeter (or a varient of it) has been used in MANY good sounding projects, including Curt's TriTrix. Your method of XO design is not a good one, however. You have to use a program that can simulate using actual frequency response curves and impedance profiles for your drivers.


One thing in particular that leaps out at me is that lack of BSC in that xo. Placed in a bookshelf, the bass might be acceptable, but otherwise, it will sound thin.

WOW!

Thanks for all your replies. If the jury is out on that tweeter, I might as well go ahead and use the Mission/Vifa tweeter that was in the box anyway. Maybe I'll start with the generic crossovers and experiment from there. Can anyone recommend a program that will simulate the actual response curves? Ideally it would need to be freeware, as I'm on a limited budget here.

I built these cabs a few years ago using some closeout treated paper cone drivers. They didn't sound that great, mostly due to the woefully inefficient woofers coming in at 82dB/W, the Bravox woofers have an 89dB/W efficiency rating.

As for the bass response, the box is 1.220 cubic feet, tuned to 53.7Hz, which gives a nice little hump in the bottom end.

I've attached a couple of pictures of how they used to look. After all the blood, sweat and splinters it took to make them, it's a shame to have them sitting in a closet doing nothing.

[tom_s]

Those cabinets are very nice. So nice, they deserve a real crossover, as opposed to an ear tweaked stock network. Free modeling programs - Passive Crossover Designer is what many people use. It runs as an Excel spreadsheet and gives great results if the data you load is accurate. Speaker Workshop is a standalone program and is a little harder to set up initially, but does a great job too.

You will also need to account for the enclosure. Response Modeler is also an Excel spreadsheet that lets you model diffraction & baffle-step losses, so you can account for them in your main design program. It also allows you to extract minimum acoustical phase for the drivers, so you can make sure they sum properly as you design your networks. Jeff B, who wrote the program (and PCD), makes this step easy - Open file, one click, watch the progress, and Save! You don't need to know the theory behind all of this stuff when you are starting out, but as you absorb more information it will start to make more sense.

Wolf & Paul C (Undefinition) both have instructions on using these great tools. I'm pretty sure they are in the sticky at the top of this forum.

Have at it!

Warning - If you are the type who enjoys puzzles and mazes, speaker design is fun but very addicting;) It's not easy, but things that are this rewarding seldom are.

[marscoast]

It can also be frustrating, confusing, insulting, irritating, tiring, mind boggling, humiliating, suicidal, drinking inducing, sleep depraving and so on. :)

Which says something about those of us who keep coming back for more! :p

[Chris Roemer]

WOW!

Thanks for all your replies. If the jury is out on that tweeter, I might as well go ahead and use the Mission/Vifa tweeter that was in the box anyway. Maybe I'll start with the generic crossovers and experiment from there. Can anyone recommend a program that will simulate the actual response curves? Ideally it would need to be freeware, as I'm on a limited budget here.

I built these cabs a few years ago using some closeout treated paper cone drivers. They didn't sound that great, mostly due to the woefully inefficient woofers coming in at 82dB/W, the Bravox woofers have an 89dB/W efficiency rating.

As for the bass response, the box is 1.220 cubic feet, tuned to 53.7Hz, which gives a nice little hump in the bottom end.

I've attached a couple of pictures of how they used to look. After all the blood, sweat and splinters it took to make them, it's a shame to have them sitting in a closet doing nothing.

I know buying cheap drivers is somewhat fun, but I don't get a very good box model with those in your 1.2 cf boxes. The "DIII" ( from speakerbuilder.net ) would use a pair of Dayton 6-1/2" "Classic" woofers in that box, uses your Silkie, and would have awesome bass.

What Qts, Vas, and Fs parms are you using for those Bravox woofers? If they ARE 89dB, then a pair (parallelled) would effectively be 95dB, which is a LOT, but would give 6dB of baffle step with that unattenuated tweeter. If your OLD pair of woofers were parallelled, then I doubt that their "dismal" sound was due to their sensitiviy, probably just a bad crossover design (or box design). It looks like you've got 4 vents. What are their lengths and IDs? Do you have an impedance plot and freq. response data for your Bravox woofers?

Chris

[patrickm]

I know buying cheap drivers is somewhat fun, but I don't get a very good box model with those in your 1.2 cf boxes.

What Qts, Vas, and Fs parms are you using for those Bravox woofers? If they ARE 89dB, then a pair (parallelled) would effectively be 95dB, which is a LOT, but would give 6dB of baffle step with that unattenuated tweeter.


i think they're 89db at 1w/1m with Z=12 ohm, so 2.83v/1m sensitivity would be more like 87.5db. i'd agree that 1.2 cf is a little on the small side for two, but shouldn't be terrible if he can live with an f3 near 50 hz.

a quick model in winisd pro gave a 1.2 cf box with a 52hz tuning as 4x 1"d ports with 2" length (just guessing from the photo on the diameter), and that was a modeled f3 of 49hz. (if they're 1.25"d, then 3.25" length, 1.5" would be 4.85" long)

[leesonic]

Got the driver holes cut out today, but haven't done the ports yet. Reason being, to cut the holes for those 1" ports (PE #260-470) correctly, I found that it works better to make a stepped hole. I wanted to see what some other ports I have lying around would do.

I've attached a graph. The cyan curve shows the four 1" ports, trimmed to 1" in length. The red curve shows a 5.5cm port trimmed to 5cm in length.

[johnastockman]

I'm making an MTM design speaker, using some Bravox woofers WFR-06BV03B PE#299-061 (or, the blue 6.5" 12 ohms to make it easier), and some Dayton tweeters DC28FT-8 PE#275-076 (or truncated 1-1/8" silkies).

I entered the specs of these drivers in my trusty old X-over 2.0 program that came with BassBox 5.1 (yes, it still runs under XP, still does what I want it to, so why upgrade?) For 2-way, 2nd order with a crossover frequency of 3kHz, it spits out 4.69uF and 0.60mH for the tweeter components, and 6.25uF and 0.45mH for the woofers (2 x 12 ohm in parallel).

Is that it? It that what crossover design is all about? Please forgive my ignorance. There are other options such as Butterworth, Bessel, Chebychev, and Linkwitz-Reily which give slightly different numbers. Are any of these better or worse? Am I doing a good thing here trying to design my own, or should I just try a set of generic crossovers I have lying around?

Thanks,

Lee.


Well Lee, to echo what the other more-knowledgeable-than-me gents have said: NO on the Xover Pro, NO on any textbook formulas and an even bigger no to the pre-built/generic XO's. Those cabinets look very nice and deserve a measured XO. Jeff Bagby's PCD is the way to go. Free and with great tutorials. There's a learning curve and it's definitely not as easy as the Harris program. Bassbox is a good program, but XO Pro is not. XO Pro assumes the drivers are a fixed impedance, like 4, 6, or 8 ohms. A driver's impedance varies across the freq. range, so using a fixed number won't work. I used a XO Pro-generated config for a small 2-way I was making...not even remotely pleasant to listen to. It was only when Shawn A. worked up a XO using actual measurements and a program to interpret all the measurements/variables, did I get something that was light-years (in terms of SQ) beyond the XO Pro crossover I had tried. And the impedance is only ONE of the many variables you have to deal with regarding driver specs/parameters. Good luck and let us know how you're doing.

John A.

[leesonic]

Hmmmm.... maybe I'll just return the Bravox woofers, and get a set of Dayton DC160s and the matching tweeters, and build the D3s complete with crossovers. I guess you get what you pay for, I just figured I'd try going with the Bravox's as they appeared to work well. Funny thing, I've been thinking of the Daytons all along, just buying 4 of them all at once kinda adds up. Also, there are at least five pairs of speakers lying around the house right now, so why make another set to just sit around unused?

Maybe the cabs had better go back into hibernation...

[Chris Roemer]

Hmmmm.... maybe I'll just return the Bravox woofers, and get a set of Dayton DC160s and the matching tweeters, and build the D3s complete with crossovers. I guess you get what you pay for, I just figured I'd try going with the Bravox's as they appeared to work well. Funny thing, I've been thinking of the Daytons all along, just buying 4 of them all at once kinda adds up. Also, there are at least five pairs of speakers lying around the house right now, so why make another set to just sit around unused?

Maybe the cabs had better go back into hibernation...

I think you should use THOSE cabs for a DIII. I've used the 160 (-305) in cabs up to 0.8 cf (vented) with good results. You should be able to retune a pair in your 1.22 cf box with good results. Also, your truncated tweeter is a drop-in for the standard (-070) Silky (AFAIK, only the faceplate was changed). Also, Johnny said he wasn't a fan of the SHIELDED Silky, I'm not either, but you apparently don't have that one.

You're almost there!

Chris

[Paul Carmody]

You can use your cabinet and re-tune or build the D3 to spec. Either way, chances are you'll be very happy with the results. The D3 is a real gateway drug.

[PWR RYD]

Warning - If you are the type who enjoys puzzles and mazes, speaker design is fun but very addicting;) It's not easy, but things that are this rewarding seldom are.

Tom, can I use your quote in my signature line? :)

[tom_s]

Tom, can I use your quote in my signature line? :)

It's all yours! :D