Thursday, December 13, 2007

design cont....




Alright folks, I know we are like a week early, but, unless you want to do all the work yourself you must rely on and trust the work of the people that came before you. In this case, I do. We also can not thank the people enough for their tireless efforts of the past, for with out them none of this would be possible. There is a slight parts delay, you need to get as many of the parts correct the first time around so as to not lose ground later. This part happens to be the inductor which they say is a simple part but you have to understand how an inductor works in order get the right one and that's not so simple. The 24db/Oct device should be sufficient to fill a 20'x40' room and there are no forseeable problems associated with that. The commercial 48db/oct device is another story, it seems that the 48 numbers do not exist because of the degree of difficulty deriving the 24 numbers. So I shall attempt to extraplicate the 24 numbers and test that to hear if there is anything detectable to the human ear, or mabey a driver can pick up some differences at louder volumes.
Above is the enclosure that will be used for testing, the final design will be discussed later. I cut down some plywood and simply cut circles to match the drivers, well not so simply, you have to get the exact size hole the first time or try again. I burned through some lumber achieving that task. the location of the drivers is not exact but it is close and will be phase- arrayed in the end. So, when the time comes i'll get back to you with more.


Friday, November 30, 2007

begin design

Greetings, earlier I explained the sound one wants to hear from your audio system, now we shall attempt to do it. The first order of business are the equasions used to determine crossover frequency and your db/oct filter. These can all be found at the link provided at an earlier post, you may have to pay for and then chase some information but it can all be found. There are some considerations though, first being the actual driver impedance vs the rated impedance. Even though a speaker is rated at lets say 8 ohms the actual may be different say 6.5 or 7. You need the actual number in the equasion in order for the numbers to come out right. So that is what will be done next, purchase drivers for the speakers,check the resistance, and use those figures to design the crossover network, then purchase the parts and build the network. This may take 2 or 3 weeks but we will be ready for photos soon after that. Oh gees, look at what time it is, I need to run and catch a banquet. Have a good day.

Saturday, November 17, 2007


Hi, as you can see I would prefer to remain anonomous. I will tell you that my name is Mark Chembrovich and I have a bachlor of science degree in engineering from a small mid-west liberal arts college. I am currently turning my attention to audio engineering and the development of advanced passive crossovers and speaker design.

Friday, November 16, 2007

A chain is only as strong as its weakest link. The same holds true when reproducing natural sound. From audios' infancy, through modern day improvement's like the transistor and compact disk, the speaker remains the last moving part and most inefficient link in the "reproduction of natural sound" chain.
So, lets focus on the speaker and more importantly the crossover network that filters sound through those individual speakers or drivers. Natural sound or sound as it is heard in nature can best be described with an orchestra. Imagine a full orchestra spread across a stage, if you sat in a good hearing position, and concentrated, you could hear each instrument coming from a particular point across the stage. This is the sound you want your audio system to deliver, a pair of multi driven speakers that do not just emanate sound from two points of origin, but the entire space between the speakers delivering the spread across stage sound of an orchestra. Achieving this goal is difficult but possible.
In order to understand this concept one must start with the frequency of sound, or hearing range, which is measured in hertz(hz). Humans can hear from 20 hz, the low end or bass, to 20,000 hz, the high end or tweeter, and everywhere in between. furthurmore the range is divided into octaves(Oct). An octave is derived by starting at 20 hz and doubling the number. This gives you 10 octaves in the human hearing range.
Now, the best sound reproductive system would consist of 10 drivers each delivering the sound of its own octave and only its own octave, verses one driver attempting to deliver the 10 octaves of the hearing range all at one time. The former is impractical in portable applications and the latter produces distortion by overworking the range of sound. The solution is to meet in the middle with a speaker that has 4 or 5 drivers to carry the 10 octaves and limit, or filter out sound, leaving only the sound designed to carried by each driver to be carried by that driver.
This bring us to the crossover network, The device used to filter sound. Sound is measured in decibels(db), whispers at a few db and loud sound being 100-150 db. The crossovers job is to filter out the hearing range not designed to carried by an individual driver, ie. only high sound through the tweeter driver, only low sound through the woofer driver, and mid sound through the midrange drivers. The problem with many home and concert speakers today is the efficiency of the crossover used to filter the sound through the drivers. This efficiency is measured in decibels/octave(db/Oct) or the amount os sound that leaks into the next octave above or below. A lot of speakers only filter 3or 6 db/Oct and at louder levels the range of sound is forced through drivers not designed to carry that range, producing distortion. The answer to this problem is to design a more efficient crossover network that filters 24 or 48 db/Oct allowing only the sound meant to be delivered by an individual driver, to actually be sent through that driver.
When designing crossover networks there are two basic choices, active and passive. Active networks separate the sound before the amplifier and the passive separate sound after the amp. The best active network that I have heard of only filters 12-18 db/Oct and at higher volume will still allow sound to leak through octaves, not to mention that these speakers are rather expensive. This is why I believe the passive network is more efficient but also requires more power in the amp to overcome the sound that is filtered out. In my opinion the best passive crossover filter is the butterworth filter design perfected by the linkwitz-Riley filter www.linkwitzlab.com/crossovers.htm which can reach limits to 48db/Oct. This limit however is best used commercially where high volume is a consideration. Home speakers are best when 4 or 5 drivers are phase-arrayed( drivers aligned at the source of inception) and filtered at 24 db/Oct, due to complications of available space required to mount the crossover device.
This is why I have undertaken the development of advanced passive crossovers, to offer the audiophile an opportunity to enjoy the reproduction of natural sound. If there is interest I will update this blog as to my progress on this issue.

Thursday, November 15, 2007

The reproduction of natural sound, crossover networks, and speaker design

Are you interested in this subject? I will be posting more information soon and am just checking the responce.