Martin,
Thanks for pointing all that out! There is a more complex relationship between the amplification and loudspeaker than I assumed.
One more question, are you using 2 alpha 15A on the same baffle dimension as ur whitepaper?
If so, would using 2 Alpha 15A per side merely means a change in low pass xo to suit a 4ohm load due to wiring the woofers in parallel?
Thanks for pointing all that out! There is a more complex relationship between the amplification and loudspeaker than I assumed.
One more question, are you using 2 alpha 15A on the same baffle dimension as ur whitepaper?
If so, would using 2 Alpha 15A per side merely means a change in low pass xo to suit a 4ohm load due to wiring the woofers in parallel?
fishball79 said:
Greets!
No, it just contradicts some folk's way of doing it and there's more than one way to do it, though with their way you have more tuning flexibility with the trade-off being typically more component cost and often mediocre dynamics, at least by my standards.
WRT your Ciares, at a glance the baffle appears too big assuming the 4 ft is its width, allowing the 'tail to wag the dog' so to speak which would cause the mids/HF to be modulated somewhat.
GM
fishball79 said:
No, the two Alpha 15A system is a completely different baffle. It is a brute force design and I would probably do it a bit differently based on what I have learned over the past year. The double Alpha 15A system is also actively biamped so the crossover is handled before the amp. If you wanted to use a passive crossover, you would have to run the numbers for the lower impedance of two driver in parallel.
fishball79 said:
fishball79 said:
Hi,
In principle two woofers per side gives +6dB (4 ohm) to suit
a 95/96dB main driver, so you would be in the right ballpark.
Arguably though you'd want a 10" to cover more of the range.
(Presumably you would need to add 3" to the baffle hieght)
No doubt the details will differ somewhat - e.g. floor dip related,
the upper driver being more on axis, and interference effects
between the bass drivers higher up, but the principles do apply.
I'm not saying its ideal, and you do not get any extra bass
extension, but it is a better starting point than an ad-hoc
arrangement, ideally you would do the simulations.
🙂/sreten.
"Greets!
You're welcome!
FWIW, of the two, the 1.43 Qts driver is the better choice IMO since by the time you get the more efficient one's Qts that high through mass loading its efficiency will be much lower, though it will have an impressively low Fs.
True.........
GM"
Hmmmm this is interesting!
That (Qts 1.43) driver is much cheaper too... a bit ugly though...!
If I understand correctly though if i did go the lower Qts route, I could put some wings on it and get down to about 50Hz without too much worry, then with a little bit of mass loading and maybe just lowering the crossover a dash.... all things could balance out without losing too much sensitivity.
Whaddayareckon?!
P.S. or I could just choose the fe83e instead! My room isn't that big!
Stroop
You're welcome!
FWIW, of the two, the 1.43 Qts driver is the better choice IMO since by the time you get the more efficient one's Qts that high through mass loading its efficiency will be much lower, though it will have an impressively low Fs.
True.........
GM"
Hmmmm this is interesting!
That (Qts 1.43) driver is much cheaper too... a bit ugly though...!
If I understand correctly though if i did go the lower Qts route, I could put some wings on it and get down to about 50Hz without too much worry, then with a little bit of mass loading and maybe just lowering the crossover a dash.... all things could balance out without losing too much sensitivity.
Whaddayareckon?!
P.S. or I could just choose the fe83e instead! My room isn't that big!
Stroop
I really don't like to push my own software, it seems kind of unethical on a DIY forum. But, the questions you are asking can be easily answered by running MathCad simulations. You could optimize the baffle size and shape, the driver selection, the driver locations on the baffle, and the crossover frequencies and slopes. You might save yourself money by going with the right drivers and time wasted if your educated guesses don't work out on the first try.
Hi MJK
I personally have no problem with you putting forward your software that will of course help me, especially when you do so much for this community.
I'm looking into!
Does it work on mac? I'm truly and spectacularly hopeless on computers....
Stroop
I personally have no problem with you putting forward your software that will of course help me, especially when you do so much for this community.
I'm looking into!
Does it work on mac? I'm truly and spectacularly hopeless on computers....
Stroop
MathCad is PC/Windows based software. I think it will run on a MAC using some form of emulation software. I have not used a MAC so I can't say much more.
stroop said:
As a general plan of action it's fine, but as MJK implied, 'God is in the details' and while to my way of thinking doing 99% of it with software defeats much of the point of DIY as a hobby, it's a huge time-saver and generally allows for much more space efficient designs to be realized without much/any experimentation, saving some $$$ on materials too.
MJK said:
If you have a new Intel Mac you have one of the best Windows boxes available. Windows (XP or Vista) can be run natively as the boot OS from Boot Camp (free except for the cost of XP) or as a virtual machine (still running natively & real Windows machines & do require an OS) in a window under OS X using Parallels, VMWare Fusion, or OpenOS X (these will run almost any x86 OS, ie various Windows, DOS 6,2, Linux, Sun OS, BeOS, OS/2 -- and with enuff RAM & disk, simultaneously. Well behaved applications can be run under the OSX version of WINE (Crossover Mac) without needing to run Windows.
dave
MJK
Great article. Nice to see good engineering and good writing.
GM
I respect your POV, but I couldn't agree less. 🙂
DIY is real engineering in the purest sense.
We decide our budget, which parts to buy and what to build,
which design tools to use, and what tools are appropriate for us to create. Also time to realization and effort expended.
To me, I can create an end result where the compromises are appropriate to my application.
Just my two cents.
Doug
Great article. Nice to see good engineering and good writing.
GM
I respect your POV, but I couldn't agree less. 🙂
DIY is real engineering in the purest sense.
We decide our budget, which parts to buy and what to build,
which design tools to use, and what tools are appropriate for us to create. Also time to realization and effort expended.
To me, I can create an end result where the compromises are appropriate to my application.
Just my two cents.
Doug
I have a Macbook and using parallels with a XP virtual machine have used MathCad.
BTW Dave Red Hat runs well too.
BTW Dave Red Hat runs well too.
Unless you wrote the software yourself.
Amen!
At work if i didnt use dynamic simulation(not mine) it would take me a great deal more time to accomplish my tasks. I can load a structure and bring the temperature up to a set point and see the actual dynamic bending and even measure the deflection. In the old days this would take me weeks , now its several hours.
World Trade Center,point of fact, at 1K degrees F the strength of steel is around 90%less. This can help engineers to design in greater strength thru alloy combination to maybe prevent the loss.
The point i am trying to get at (and doing a poor job of it is) simulations are not the total end answer, but does eliminate a great deal of prototyping and testing.
Its very easy to incorporate math into a program, it saves time and energy and mistakes if the programming is correctly structured.
To me, I can create an end result where the compromises are appropriate to my application.
Very well stated!
ron
Amen!
At work if i didnt use dynamic simulation(not mine) it would take me a great deal more time to accomplish my tasks. I can load a structure and bring the temperature up to a set point and see the actual dynamic bending and even measure the deflection. In the old days this would take me weeks , now its several hours.
World Trade Center,point of fact, at 1K degrees F the strength of steel is around 90%less. This can help engineers to design in greater strength thru alloy combination to maybe prevent the loss.
The point i am trying to get at (and doing a poor job of it is) simulations are not the total end answer, but does eliminate a great deal of prototyping and testing.
Its very easy to incorporate math into a program, it saves time and energy and mistakes if the programming is correctly structured.
To me, I can create an end result where the compromises are appropriate to my application.
Very well stated!
ron
Thanks for the well worded, easy to read and understand article MJK. Few papers are as well written to be easy to be read.
Few questions though,
Question 1:
What happens if the baffle was tilted further back to say 45 or 30 degrees?
My intention is mount the speaker on the ceiling. I suppose there would be greater rear wave bounce?
Question 2:
Can anyone clarify why when the woofer extends lower in frequency, we should expect higher transient response? (page 18 last paragraph of the article)
Few questions though,
Question 1:
What happens if the baffle was tilted further back to say 45 or 30 degrees?
My intention is mount the speaker on the ceiling. I suppose there would be greater rear wave bounce?
Question 2:
Can anyone clarify why when the woofer extends lower in frequency, we should expect higher transient response? (page 18 last paragraph of the article)
zenesh said:
Presumably this is upside down and the mid/ tweeter pointed
towards ear height ? Note that the article sidesteps the issue
of distance to the rear wall. Near a rear wall dipoles do not
produce bass, it sounds like you are considering this.
zenesh said:
On page 8 it states the lower the frequency extension the longer
the transient response. Consider the opposite (a tweeter) the less
bass extension you have the shorter will be the transient response.
In other words the "settling time" to an impulse depends on the
high pass function applicable, the lower the frequency of this
function the longer it takes to settle for a given high pass.
The shape of the response depends on the high pass alignment.
This shape extends in the time domain as you lower the frequency
of a given low pass alignment.
🙂/sreten.
zenesh,
I don't know the answer to what happens if you tip the baffle back 30 to 45 degrees. Obviously it will have some impact but I have no way of calculating the response at the moment. I will have to think about this some more.
sreten,
You are correct I did not include the rear wall reflection in any of the plots. However, when I do simulate the speaker placed 1 m off of the back wall, the response is not so bad at either a 1 m or 3 m listening position. You get some comb filtering due to the reflections arriving in and out of phase with the main signal but these are very sharp nulls, not sure how audible this would be. The same phenominon occurs with a traditional boxed speaker to some extent.
When I place the speaker 0.001 m off of the rear wall the bass does cancel out completely and the response only returns when the baffle dimensions are sufficiently large with respect to the wavelength of sound and the Fostex driver starts to become directional.
I am sure that placing the OB even further from the wall would be a benefit but I don't think that 1 m is a total disaster. My Lowther OB system is about 1 m from the rear wall, my room is really too small for such a large speaker, and it seems to work very well. Nobody who has heard it has complained about lack of bass.
I don't know the answer to what happens if you tip the baffle back 30 to 45 degrees. Obviously it will have some impact but I have no way of calculating the response at the moment. I will have to think about this some more.
sreten,
You are correct I did not include the rear wall reflection in any of the plots. However, when I do simulate the speaker placed 1 m off of the back wall, the response is not so bad at either a 1 m or 3 m listening position. You get some comb filtering due to the reflections arriving in and out of phase with the main signal but these are very sharp nulls, not sure how audible this would be. The same phenominon occurs with a traditional boxed speaker to some extent.
When I place the speaker 0.001 m off of the rear wall the bass does cancel out completely and the response only returns when the baffle dimensions are sufficiently large with respect to the wavelength of sound and the Fostex driver starts to become directional.
I am sure that placing the OB even further from the wall would be a benefit but I don't think that 1 m is a total disaster. My Lowther OB system is about 1 m from the rear wall, my room is really too small for such a large speaker, and it seems to work very well. Nobody who has heard it has complained about lack of bass.
- Status
- Not open for further replies.