Ok, I'm not new to speaker building or electronics in general, but I realize I am over my head at the moment and I need some help going in the right direction regarding crossovers.
One day a few weeks ago I decided to build some new speakers for my PC stereo system. I found a simple full range Tang Band 4" bamboo design here and decided to build it. I built it just fine, hooked it up and it made music...but it sounded so lifeless.....
So I began reading a ton of stuff about crossovers, theory, BSC, etc and I feel confident about what the problem is but I don't know how to fix it. Upon testing first and reading later, I figured out that the crossover (or eq + impedance comp) circuit that was created by the designer was meant for stand mounted speakers in a room with the listener seated several feet back....I am using this design on a desktop in a very nearfield position. When I disconnected the crossovers and ran straight wire, the speakers sounded alive again, though it lost some bass.
So, I humbly request help! I want to know how to design the proper BSC and impedance circuits myself for this application. How does one go about figuring out what circuit is required for a desktop full range speaker in the first place?
One day a few weeks ago I decided to build some new speakers for my PC stereo system. I found a simple full range Tang Band 4" bamboo design here and decided to build it. I built it just fine, hooked it up and it made music...but it sounded so lifeless.....
So I began reading a ton of stuff about crossovers, theory, BSC, etc and I feel confident about what the problem is but I don't know how to fix it. Upon testing first and reading later, I figured out that the crossover (or eq + impedance comp) circuit that was created by the designer was meant for stand mounted speakers in a room with the listener seated several feet back....I am using this design on a desktop in a very nearfield position. When I disconnected the crossovers and ran straight wire, the speakers sounded alive again, though it lost some bass.
So, I humbly request help! I want to know how to design the proper BSC and impedance circuits myself for this application. How does one go about figuring out what circuit is required for a desktop full range speaker in the first place?
ah yes, baffle step is critical, isn't it ?
Here is an article using the fx120, notice the different values based on how much attenuation is needed.
Fostex FX120 DIY Bass Reflex Bookshelf Speakers
So long story short, the resistor in parallel with the inductor will need to be smaller, and the inductor a little smaller.
Adjust for your taste.
Here is an article using the fx120, notice the different values based on how much attenuation is needed.
Fostex FX120 DIY Bass Reflex Bookshelf Speakers
So long story short, the resistor in parallel with the inductor will need to be smaller, and the inductor a little smaller.
Adjust for your taste.
Another link to check out:
General Speaker Related Articles
Lots of good info on MJK's site.
Cheers, Jim
General Speaker Related Articles
Lots of good info on MJK's site.
Cheers, Jim
Thanks guys, lots to read there....
Two questions:
1. I was reading about the zobel circuits and I've got two different ways of doing it and I don't see how how MJK's version is accurate. MJK gives a formula based on the measured DC resistance of the voice coil. The other plays test tones through the speaker and measures resistance at each freq. Once you know exactly where the impedance is the highest, you can correct for it. How does MJK's version correct for impedance when you don't know what freq it's highest at?
2. It seems to me I need to correct for reflections from my desktop surface too don't I? What about my monitor? Since I know where they are and they aren't going anywhere, I can design them into my crossover can't I? Or am I being too complicated here?
Thanks again!
Two questions:
1. I was reading about the zobel circuits and I've got two different ways of doing it and I don't see how how MJK's version is accurate. MJK gives a formula based on the measured DC resistance of the voice coil. The other plays test tones through the speaker and measures resistance at each freq. Once you know exactly where the impedance is the highest, you can correct for it. How does MJK's version correct for impedance when you don't know what freq it's highest at?
2. It seems to me I need to correct for reflections from my desktop surface too don't I? What about my monitor? Since I know where they are and they aren't going anywhere, I can design them into my crossover can't I? Or am I being too complicated here?
Thanks again!
Unless you actually chart the T/S curves you will not have exact measurements. With the simple PE kit you'll be measuring on another part of the curve from where the factory measures, You will have the measurements for that driver on that part of the curve (there is a margin of error and the measures you take can change with the weather)
dave
Simulatio BSC
Hi soundwavesteve
I have done some simulations with MJK´s models, and the filter as suggested by Emil Attlid is perfect. I have simulated a close listening position, and there is no doubt Emil know what he is doing
I have enclosed some of the plots, which are pretty close to the actual measurements.
Please note: Add 18 gram of stuffing in the cabinet (not the port) to smoothen the SPL.
Hi from
Bjorn
Hi soundwavesteve
I have done some simulations with MJK´s models, and the filter as suggested by Emil Attlid is perfect. I have simulated a close listening position, and there is no doubt Emil know what he is doing
I have enclosed some of the plots, which are pretty close to the actual measurements.
Please note: Add 18 gram of stuffing in the cabinet (not the port) to smoothen the SPL.
Hi from
Bjorn
Hey Bjorn and everybody,
I finally got my Dayton DATS speaker tester in today and began doing my own measurements. Plus I took some time today to check my speakers with test tones in the listening position.
First: the mfg specs list R(e) as 6.3 ohms, and my speakers measure 6.291 and 6.285. When I plug those numbers into a zobel calc or even when i do the calc myself, I get 7.8 ohms for the resister. Not 6.8 as Emil suggested.
So, I would assume that I should use the number that I got since it is accurate to my specific speakers. The rest of the Zobel calculation asked for L(e) [Voice Coil inductance], and my speakers measured .000974 mH at 1KHz. That works out to a cap of .015uF. Emil didn't even have a cap, and it almost seems like the number is so small as to be insignificant. What do you guys think?
Two: With no impedance comp or baffle step, just amp wired directly to speakers in the box, I measured the SPL with test tones and Audiotools on my phone, plus a radio shack meter to double check. From 100Hz to 700Hz, my SPLs measured between 67 and 74 db, then a big drop to 55 at 1Khz. From 1400 to 5000 Hz, its between 59 and 64db. With these measurements, it makes me wonder if I even need to build a baffle step - if I did it would bring down the 1400+ range even farther then it already is correct?
*note, these measurements where taken with a 23" LCD between the speakers (no gap) and with the speakers on stands 6" above a glass desk*
If I put back in the baffle step circuit suggested by Emil, everything above 700Hz or so drops even further by 5 db, which is why it sounded so dark and closed in....
Again I am very new to this type of in-depth experimentation, and perhaps it is very likely there is no simple or easy answer, but why does my own calculations differ so much from Emil's and the sim created by Bjorn? For example, the only way to get a suggested zobel circuit of 6.8 ohms is to have a speaker's R(e) be 5.45, a long way off from the published spec of 6.3 and my own measured 6.29....
I finally got my Dayton DATS speaker tester in today and began doing my own measurements. Plus I took some time today to check my speakers with test tones in the listening position.
First: the mfg specs list R(e) as 6.3 ohms, and my speakers measure 6.291 and 6.285. When I plug those numbers into a zobel calc or even when i do the calc myself, I get 7.8 ohms for the resister. Not 6.8 as Emil suggested.
So, I would assume that I should use the number that I got since it is accurate to my specific speakers. The rest of the Zobel calculation asked for L(e) [Voice Coil inductance], and my speakers measured .000974 mH at 1KHz. That works out to a cap of .015uF. Emil didn't even have a cap, and it almost seems like the number is so small as to be insignificant. What do you guys think?
Two: With no impedance comp or baffle step, just amp wired directly to speakers in the box, I measured the SPL with test tones and Audiotools on my phone, plus a radio shack meter to double check. From 100Hz to 700Hz, my SPLs measured between 67 and 74 db, then a big drop to 55 at 1Khz. From 1400 to 5000 Hz, its between 59 and 64db. With these measurements, it makes me wonder if I even need to build a baffle step - if I did it would bring down the 1400+ range even farther then it already is correct?
*note, these measurements where taken with a 23" LCD between the speakers (no gap) and with the speakers on stands 6" above a glass desk*
If I put back in the baffle step circuit suggested by Emil, everything above 700Hz or so drops even further by 5 db, which is why it sounded so dark and closed in....
Again I am very new to this type of in-depth experimentation, and perhaps it is very likely there is no simple or easy answer, but why does my own calculations differ so much from Emil's and the sim created by Bjorn? For example, the only way to get a suggested zobel circuit of 6.8 ohms is to have a speaker's R(e) be 5.45, a long way off from the published spec of 6.3 and my own measured 6.29....
PKI, I recommend D'Appolito's Measuring Loudspeakers. It will be tough for a beginner, but if you study, think and measure, it will speed you along better than any book I have.
Steve, yes sims and measures differ. One reason is measurement is not that easy to do correctly and two, sims are only so good. Both are informative. As an example, I learned some time ago how important a large radius on the baffle is for managing diffraction. No sim I know of takes that into account.
Steve, yes sims and measures differ. One reason is measurement is not that easy to do correctly and two, sims are only so good. Both are informative. As an example, I learned some time ago how important a large radius on the baffle is for managing diffraction. No sim I know of takes that into account.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.