Hello all, I hate to burden with just a simple crossover question, but I will admit to my utter ignorance in the art of good crossover design. My plan is the following:
I like very much the looks and sound of the Sonus Faber Concerto stand mouted speaker, its a relatively simple looking 2 way using what appears to be a modified Vifa Xt woofer and some Scan Speak or Vifa Soft dome. I have decided that rather than plopping down 2K on a set, I would try my hand at building something that may outperform them and keep the asthetics that I find very appealing. The cabinet will be an almost exact clone, with some added size to accomidate the bump up to 16L internal volume. I would like to use the Scan Speak 8545K00 for the mid/woofer in place of the Vifa, and the Scan Speak 9700 in place of the tweeter that the Sonus uses. It is my hope that the better drivers will help to bring a higher degree of accuracy to the table than the original. Sonus has a very engaging sound, but it needs some work tonally in my humble... As both seem to hold quite a degree of regard, and look good on paper, I have confidence they will serve me well in this purpose.
I've run some preliminary simulations in Winisd and settled on a 16L cabinet tuned to 33Hz, the roll-off seems to be well suited for room gain and group delays and etc. are well within the acceptable limits. Any input on the cabinet size or tuning would be appreciated as well.
The crossover is where I run into the problem. I would like to run an even order network to minimize phase problems. I am unsure though in the ability of a 12db slope to alleviate the nasty peak at around 3K that the woofer has. But, on the other hand, I am concerned that a 24db slope will be too steep and will sound somehow unnatural in the end. I would like to keep the crossover point at 2K or lower to mimize beaming and to move away from the peak the woofer shows up higher.
Any and all help or input or fresh ideas would be greatly appreciated. If any more info is needed, please feel free to ask as this project is just in the first planning stages.
I like very much the looks and sound of the Sonus Faber Concerto stand mouted speaker, its a relatively simple looking 2 way using what appears to be a modified Vifa Xt woofer and some Scan Speak or Vifa Soft dome. I have decided that rather than plopping down 2K on a set, I would try my hand at building something that may outperform them and keep the asthetics that I find very appealing. The cabinet will be an almost exact clone, with some added size to accomidate the bump up to 16L internal volume. I would like to use the Scan Speak 8545K00 for the mid/woofer in place of the Vifa, and the Scan Speak 9700 in place of the tweeter that the Sonus uses. It is my hope that the better drivers will help to bring a higher degree of accuracy to the table than the original. Sonus has a very engaging sound, but it needs some work tonally in my humble... As both seem to hold quite a degree of regard, and look good on paper, I have confidence they will serve me well in this purpose.
I've run some preliminary simulations in Winisd and settled on a 16L cabinet tuned to 33Hz, the roll-off seems to be well suited for room gain and group delays and etc. are well within the acceptable limits. Any input on the cabinet size or tuning would be appreciated as well.
The crossover is where I run into the problem. I would like to run an even order network to minimize phase problems. I am unsure though in the ability of a 12db slope to alleviate the nasty peak at around 3K that the woofer has. But, on the other hand, I am concerned that a 24db slope will be too steep and will sound somehow unnatural in the end. I would like to keep the crossover point at 2K or lower to mimize beaming and to move away from the peak the woofer shows up higher.
Any and all help or input or fresh ideas would be greatly appreciated. If any more info is needed, please feel free to ask as this project is just in the first planning stages.
Well, I hoped somebody would have at least something to say. I've been searching through many different crossover tutorial pages but I still am about as clear as mud on what I need in this project. I've seen second order on the woofer fourth on the tweeter, second order on both, fourth on both, all possiblities, but I don't understand how to make that decision. Please guys, anything would help.
Another thing I don't get is the L-pad attenuation and how it ineracts with the crossover design. Do I need to adjust the values I plug into the calculator to accomidate the 2db's of attenuation I need on the tweeter or is that a seperate circuit?
Another thing I don't get is the L-pad attenuation and how it ineracts with the crossover design. Do I need to adjust the values I plug into the calculator to accomidate the 2db's of attenuation I need on the tweeter or is that a seperate circuit?
I've used 24db/octave slopes (that's 24 dB/octave acoustic, not electrical!) for my last 4 or 5 projects, after spending many years before that trying to overcome the disadvantages of 6dB/octave xovers. IMO, it's one of the best ways to go for power handling, insensitivity to driver offsets, and low distortion. It seems to give better polar patterns as well. The downside is that you don't get nice-looking square waves.
The keys are understanding the difference between acoustic and electric slopes and going into battle with the right weapons- frequency response and impedance measurement capability and software that lets you model crossovers with the experimental data you get from the drivers.
The keys are understanding the difference between acoustic and electric slopes and going into battle with the right weapons- frequency response and impedance measurement capability and software that lets you model crossovers with the experimental data you get from the drivers.
Thats my problem SY, my slopes will be practically entirely electrical as the response of the 8545 hit a rise at 3200 and the inductance limit at a little over 4K from the looks of the graph, 4K is a full octave above the crossover point of 2k +/-. The tweeter doesn't begin rolling off until down in the 500-600hz region. My real concern is the peak at 3200 from the woofer, how can I keep that from making my speaker sound nasal or too forward? I was thinking 4th order on the woofer and 2nd order on the tweeter. From my understanding of phase relationships, this should put me 180 out of phase allowing me to just reverse the polarity of one of the drivers and I am back at 0. Is this correct?
Does anyone have any recommendations for alternate crossover points or slopes that may yield a better result? My emphasis here is mainly on good horizontal dispersion as imaging and depth are really what I'm looking for in this project. I can live with a little bit of compromise on the frequency response as long as the speaker is engaging in the soundstage.
Does anyone have any recommendations for alternate crossover points or slopes that may yield a better result? My emphasis here is mainly on good horizontal dispersion as imaging and depth are really what I'm looking for in this project. I can live with a little bit of compromise on the frequency response as long as the speaker is engaging in the soundstage.
Seth Smith said:
I was wondering if you have any measurement equipment, and the ability to export those measurements to a crossover design program? If you don't then all of your planning regarding phase and slopes of the crossover point may be wasted. Any formula derived crossover will have errors from the ideal due to baffle diffraction, driver response variations, etc.
I think your driver choice is excellent, but I wouldn't be so quick to rule out the Vifa XT series woofers as serious contendors. They are quite linear, do very well on IMD tests and have a great sound. I like them a lot! Both the XT wood pulp and the Scan Speak woofers have cone breakup modes in the 3K-5K region so crossovers often need a little extra attention by ear.
Scott Hinson
I don't have any measurement soft or hardware as I'm completely new to the serious DIY hobby. I would like this to be a one time project, so, an investment of 300+ dollars in computational aid for one pair of speakers is a bit much I think. What is the least I can get by with? I figure $70.00 for a mic, and another $70 for software at the bare minimum. Is this about accurate?
The XT isn't a bad driver and is in fact my second choice if the Scan Speak's large 3K peak can't be dealt with. It trades a bit of off axis dispersion for its smoothness though and I am concerned this may collapse the soundfield. Any thoughts on that possiblity?
The XT isn't a bad driver and is in fact my second choice if the Scan Speak's large 3K peak can't be dealt with. It trades a bit of off axis dispersion for its smoothness though and I am concerned this may collapse the soundfield. Any thoughts on that possiblity?
There are a lot of issues here. Not the least of which is your use of a terms "serious" and "one time". SoundEasy can be had for $225, a mic can be built for about $5, a mic preamp is around $50. Or you can go with Speaker Workshop and save the cost of SoundEasy.
Puting a cap and an inductor on a driver does not result in a 2nd order crossover in most all cases. It's the total electrical and acoustic slope that determines the order.
Russ
Puting a cap and an inductor on a driver does not result in a 2nd order crossover in most all cases. It's the total electrical and acoustic slope that determines the order.
Russ
russbryant said:
I totally agree. If you think about the amount of money you're planning on spending on the drivers you've selected, and the amount of time you'll invest a SoundEasy puchase may be the way to go. It's not an easy program to learn (at least last time I used it it was't...I haven't upgraded in a while so I don't know what it's like right now.) but it will do the job.
Scott
russbryant said:
Not exactly sure what that was supposed to mean
Considering the Sonus Concerto can be had new for 1400, the project's cost is already at roughly 1100 in materials alone. The addition of another 300-350 for software puts me dead even in dollars spent and my time and frustration hasn't even been considered yet. While I'm confident that the DIY result will outperform the Concerto, it makes the decision more difficult in determining what is the most cost efficient route to take when all things are considered. I still need to buy a larger 2 channel amp, stereo pre with high resolution DAC, and CD Transport. That being said, the budget alotted for the speakers is not to exceed $1600.
As this system upgrade is going to cost close to 5K after all is said and done, another DIY project will not be undertaken for quite some time to come, especially if these monitors turn out as well as I hope they will. Now do we understand each other on my hesitance to pay for soft and hardware that will be essentially a one time use? If it is necessary, then there is no avoidinig it, but if there is an effective cheaper way out, I would much prefer it.
I understand the concept of summed slopes, but at the 2K crossover point, why is there an acoustical roll-off in the equation, the woofer does not hit its inductance limit for another octave, and the tweeter does not roll off until below Fs (500Hz). Do these figures somehow change when put into an enclosure? I can understand low end limit on the woofer, but not high. Please elighten if you will.
Seth Smith said:
Not exactly sure what that was supposed to mean
Ohhh! Please don't be upset in any way! I can't speak for Russ, but I'm trying to help out a local guy who asked for some advice. If I've offened you in any way I apologize, I'm just trying to clear up some confusion that I see!
Seth Smith said:
Keep in mind that the least expensive commercial speakers using the 8545 SS woofer listed in the LDSG are the Meadolark Shearwater for $2400 a pair. (There is a Norh kit that's cheaper but I'm only looking at completed speakers.) So yes, it's getting close parts wise to the Sonus Concerto, but they are more expensive parts than are what in the Sonus speaker.
Seth Smith said:
If you factor in the time you'll spend on the project (which should be fun) then the DIY route is vastly more expensive.
Seth Smith said:
The frequency response of the drivers will change drastically when placed into an enclosure. Vifa tests their speakers on an essentially infinite baffle in 2-pi space. (According to their web site)
The plot below shows what can happen with a driver in a speaker. The red line is the frequency response of a Seas 27TDC with a 3kHz textbook 2'nd order crossover. The grey line is the ideal Butterworth 2'nd order 3kHz crossover. There is a pretty big difference, mostly due to response variations caused by cabinet diffraction. In fact for this speaker I spent quite a bit of time minimizing the effects of diffraction, so the error for most applications might be even more.
Scott Hinson
Attachments
The next plot shows what 30 seconds of optimization can do. Same driver, same enclosure, same 2'nd order 3kHz Butterworth target crossover shape, but with completely different crossover values.
The capacitor value changed by a factor of 3, the inductor by ~5. I also had to add a series resistor. The last parallel resistor of ~700 ohms could easily be left out with no adverse effects..I had it there in case I needed it.
Scott
The capacitor value changed by a factor of 3, the inductor by ~5. I also had to add a series resistor. The last parallel resistor of ~700 ohms could easily be left out with no adverse effects..I had it there in case I needed it.
Scott
Attachments
Thanks for the info Scott, this is definately what I was looking for!
I went to half price books today and tried to get a copy of the loudspeaker design cookbook, but I had no luck. Some interesting looking texts on op amps but thats another topic...
Those plots look very nice, is that the soundworks program? Can you describe a bit the setup you used to measure and process your results?
I knew the cabinet was going to give me issues, thus my loss for working out this crossover. The leather wrap I'm going to use and the double roundover I intend to use is going to throw some wierd things in the mix too.
As far as the cost, I'm just trying to make budget, I know the speakers are better, thats why I want to do this, I'm very confident that if I can get this to work out, this project will destroy a Concerto for the same cash. My labor is no biggie, I like to do this stuff, I enjoy the challenge, its just the time expense that hurts, freetime is not a commodity I have a surplus of.
I went to half price books today and tried to get a copy of the loudspeaker design cookbook, but I had no luck. Some interesting looking texts on op amps but thats another topic...
Those plots look very nice, is that the soundworks program? Can you describe a bit the setup you used to measure and process your results?
I knew the cabinet was going to give me issues, thus my loss for working out this crossover. The leather wrap I'm going to use and the double roundover I intend to use is going to throw some wierd things in the mix too.
As far as the cost, I'm just trying to make budget, I know the speakers are better, thats why I want to do this, I'm very confident that if I can get this to work out, this project will destroy a Concerto for the same cash. My labor is no biggie, I like to do this stuff, I enjoy the challenge, its just the time expense that hurts, freetime is not a commodity I have a surplus of.
Seth,
I just meant that I don't see how this can be a serious hobby for you if you plan on making one set of speakers and then calling it quits. You might be better off buying a kit from one of the better designers. You could get a seriously good set of speakers that way.
If cost is that much of an issue, you can still use Speaker Workshop. It's free. It will make the measurements and last I heard the crossover optimizer is working too. It might not be as easy to use as some other programs but for free it's really good. You just have to have a sound card that works with it (most do).
If your concerned with that little peak at 3.2khz wait till you see it when the driver is mounted in the baffle with all the baffle step.
Scott: You should upgrade to the new version of SoundEasy. The user interface is getting better all the time. Actually the next version is being talked about now. Not sure when it will be released though. Join the User Group if your not on already.
Russ
I just meant that I don't see how this can be a serious hobby for you if you plan on making one set of speakers and then calling it quits. You might be better off buying a kit from one of the better designers. You could get a seriously good set of speakers that way.
If cost is that much of an issue, you can still use Speaker Workshop. It's free. It will make the measurements and last I heard the crossover optimizer is working too. It might not be as easy to use as some other programs but for free it's really good. You just have to have a sound card that works with it (most do).
If your concerned with that little peak at 3.2khz wait till you see it when the driver is mounted in the baffle with all the baffle step.
Scott: You should upgrade to the new version of SoundEasy. The user interface is getting better all the time. Actually the next version is being talked about now. Not sure when it will be released though. Join the User Group if your not on already.
Russ
Like I said, cost is the major issue here, I simply don't have the funds to do this two channel upgrade and build a new set of high resolution monitors every month. I'm not calling it quits per se, I'm just realistic in my timeline, the money simply won't be there for at least another year after this project is done. In the meantime, I can troll these boards and keep learning.
I've looked at countless kits and none seem to offer what I'm looking for in a single package, there are either acoustic or asthetic shortcomings in just about everything I've seen thus far. If there was a kit that really appealed to me, I would still be asking the above questions just to see what can be done in the truly custom realm.
I know BSC is going to be necessary, but I didn't even consider this as it is a total unknown until the drivers actually get in the cabinet. My line of thinking was establish a baseline design, build it with cheaper components, and when it is all tweaked, then invest in the Theta Caps, Alpha Core inductors, and Mills resistors to add the finishing touches. The problem is determining where that starting point is. That was really the heart of the original question, what's the crossover point, what slopes, and what topology.
I'll look into speaker workshop right now and see if it is going to work on my laptop.
I really do appreciate this guys, its helping tremendously, keep it going!
I've looked at countless kits and none seem to offer what I'm looking for in a single package, there are either acoustic or asthetic shortcomings in just about everything I've seen thus far. If there was a kit that really appealed to me, I would still be asking the above questions just to see what can be done in the truly custom realm.
I know BSC is going to be necessary, but I didn't even consider this as it is a total unknown until the drivers actually get in the cabinet. My line of thinking was establish a baseline design, build it with cheaper components, and when it is all tweaked, then invest in the Theta Caps, Alpha Core inductors, and Mills resistors to add the finishing touches. The problem is determining where that starting point is. That was really the heart of the original question, what's the crossover point, what slopes, and what topology.
I'll look into speaker workshop right now and see if it is going to work on my laptop.
I really do appreciate this guys, its helping tremendously, keep it going!
Alright, I've got speaker workshop downloaded and I gave the help section a quick once over. It looks like it can do everything I will need to design the crossover and really tweak the system. Now, I just need to make my supply purchases. Maybe by the time I get the cabinets done, I will have the program down enough to really put it to use.
Seth Smith said:
Seth,
Are you an Engineering student at U.T.? Have you taken a look at my member profile?
I've never found the LDC to be particularly useful, not for the level of speaker engineering that I do. What I have found incredibly useful is:
Testing Loudspeakers by Dr. Joseph D'Appolito.
http://www.madisound.com/books.html
When I met Dr. D'Appolito two years ago at CES he and I chatted a few minutes about the book. I thanked him for putting together such a well founded science based approach to loudspeaker measurement, and frankly development. This is the book to buy to understand what is needed to accurately measure and model speakers.
As far as "destroying" the Concerto...well....I can tell you this. The Sonus Faber folks are VERY competent speaker designers, who have relatively unlimited design resources (for a speaker company anyway) at their hands.
I use LAUD to measure, CALSOD and LSPCad Pro for modeling. Measurements for smaller speakers are made in my office with 8' ceilings, window time allowing good resolution to around 300Hz or so. Nearfield measurements are spliced and scaled using the techniques presented in Dr. D'Appolito's book. Everything is run through a Hilbert transform to get the minimum phase response from the frequency response data. I'll then measure the acoustic centers of all drivers to get relative acoustic offset data to enter into the simulation packages. Typically I'll take 15 and 30 degree off axis measurements as well.
Larger speakers require a trip outside to the driveway, and no wind. Typically late at night or very early morning. Measurements are made with the ground plane method at 2 meters to scale to 1 meter sensitivity. Sometimes for very large speakers I'll measure at 4 meters since most people will be listening that far away anyway.
Scott Hinson
Seth,
If you can get Speaker Workshop to where you can do a frequency measurement, then you will at least be able to see what the effects of tweeking the crossover are. This will help you get to a reasonable crossover in a shorter amount of time.
I'll second the "Testing Loudspeakers" book by Joe D'Appolitto.
Scott,
Espresso?? I'm trying to sell the wife on a new e61 group machine. There are some serious machines on the market now.
Do you frequent the alt.coffee newsgroup at all? I check it out every once in awhile. I'm doing some home roasting with a popcorn popper now.
Regards,
Russ
If you can get Speaker Workshop to where you can do a frequency measurement, then you will at least be able to see what the effects of tweeking the crossover are. This will help you get to a reasonable crossover in a shorter amount of time.
I'll second the "Testing Loudspeakers" book by Joe D'Appolitto.
Scott,
Espresso?? I'm trying to sell the wife on a new e61 group machine. There are some serious machines on the market now.
Do you frequent the alt.coffee newsgroup at all? I check it out every once in awhile. I'm doing some home roasting with a popcorn popper now.
Regards,
Russ
russbryant said:
Russ, I sent you an email about the espresso, didn't want to de-rail the thread tooo bad.
Seth,
Can speakerworkshop do impedance as well? If so, that's definately something to check as you tune a crossover as well. You'd be surprised to find out how quickly you can present a 3 ohm -60 degree load to an amplifier even when using an 8 ohm woofer and a 6 ohm tweeter.
Regards,
Scott
Nope, no engineering, I'm a bio guy, never was too fond of advanced mathematics. I just got in from a super busy day and I'll check up tomorrow afternoon when I've had some sleep on the impedance capability and resolution of speaker workshop. As far as impedance angles, I know shallow is good, but exactly why is that?
As I'm sure you all have noticed, I'm the type of person that has to talk everything out to really get a grasp of whats happening and make all the fine connections.
As I'm sure you all have noticed, I'm the type of person that has to talk everything out to really get a grasp of whats happening and make all the fine connections.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.