nickmckinney said:
*shrug* I have no idea, I did not write the spreadsheet and I do not understand the math. I only recognized the graphs
You'll understand this much better than I did:
http://www.diyaudio.com/forums/showthread.php?s=&postid=728426&highlight=#post728426
He does say that the room effects are a fudge / first approximation.
I remember Thorsten, good guy.
However on the "fudge" is it Chocolate or Peanut Butter?
I want to see some measurements matching these simulations before I buy that any woofer can do 40Hz at 87dB at one watt at one meter or whatever without needing a cabinet or infinite baffle. That means 107dB at 40Hz for a single driver with 100watts applied? That is some serious output for a boxless woofer.
nickmckinney said:
Keep in mind that oven if all this was correct, with 12.5W, the Alpha15 has already surpassed the rated 3.8mm Xmax at 40hz in this application. The driver is excursion limited to no more than 97dB in this application. Again the simulations show some amount of gain for boundaries. This is going to be very placement sensitive. As your boundaries move, so does your gain from the boundaries. I don't see how it can be generalized
John
Time to throw in this one
http://www.quarter-wave.com/OBs/OB_Theory.html
MJK is called MJK/Martin in here
Picture of Alpha15 ... I suppose it dont like it too loud
http://www.quarter-wave.com/OBs/OB_Theory.html
MJK is called MJK/Martin in here
Picture of Alpha15 ... I suppose it dont like it too loud
Attachments
You can download the spreadsheet and play with it. It requires you to input distance from floor, side and back walls. As those numbers change, the response curves change. I'm not sure if the actual equations are visible or locked; like I said, I don't understand the math and wasn't looking to dig that deep into it. IIRC, there are other tabs in the spreadsheet that take power handling and excursion into account and show SPL limits.
Martin has MathCAD worksheets that also allow you to enter floor and back wall distances. Again, I've just played with the tool, I don't pretend to understand the math.
I hope it's chocolate. I hate peanut butter
I think Thorsten said not to read too much into his spreadsheet, and that it's just a rough guide, or something like that. If nothing else, it gives some idea of where the response might start to fall off, how much EQ may be required, etc. At some point, you have to build and measure.
John_E_Janowitz said:
If you are listening to a pure sine wave with a frequency of 40 Hz and and an input power of 12.5 watts then I agree that you will exceed the 3.8 mm of Xmax (I'll take it on faith that you have done the math correctly) of the Alpha 15A. But I don't listen to pure sine waves, music is the summation of many sine waves of all different frequencies so I think your calculation is extremely pessimistic. I don't worry too much about Xmax when designing a speaker for music. If I was designing for HT maybe it is more important. I can tell you that the Alpha 15A mounted in an OB that gets it down to 40 Hz can play very loud and you have to touch it to prove to yourself that it is even moving, the motion is not easily visible.
Good morning guys (well, it's still morning here).
John, I could go over your post #39 point by point - but there's no need. I agree with all the points! Going with 4 IB15s per side should work great. But that's a monster rig. The bigger baffle, deeper side wings and all that x-max should really shake things up!
Also agree with you on the Alpha power issue. Looks to me like it hits x-max about about 12W or less at 40Hz. Maybe that's why Martin uses 2 per side. But mostly on music the big pro woofers hardly move at all.
For me, a Qts of about 6 or 7 works well.
As for that sim, yes, it's X-Baffle. I thought I had noted that it was and also the fact that it's not perfect - maybe I left that part out. Sorry. But I have found it to be very close to what I measure and hear in room.
The simulation is under the following conditions;
Baffle: 965x510mm (aprox MJK size)
Woofer Height: 400mm
Dist. Rear Wall: 1500mm
Dist. Side Wall: 2000mm
Don 't know if that typical, maybe a bit on the cramped side for room placement. It's similar what I have here in my little cottage by the sea.
The reason you see the drivers going above their rated sensitivities in the midrange is room gain. In this case about 4dB worth.
As to having a driver that will do high levels at 40Hz with little power, yeah - within its x-max. As John importantly notes above. But subjectively it won't sound like it. It will sound very bass shy. Why? Look at all that mid-range energy. That's not going to sound very balanced.
So you have to tame that mid-range (and get rid of the highs) for it to sound balanced.
My original point was that a driver with an sensitivity of 86dB/W is going to need a lot of power on an OB. I still think that's true. But of course, things are rarely as simple as that.
John, I could go over your post #39 point by point - but there's no need. I agree with all the points! Going with 4 IB15s per side should work great. But that's a monster rig. The bigger baffle, deeper side wings and all that x-max should really shake things up!
Also agree with you on the Alpha power issue. Looks to me like it hits x-max about about 12W or less at 40Hz. Maybe that's why Martin uses 2 per side. But mostly on music the big pro woofers hardly move at all.
For me, a Qts of about 6 or 7 works well.
As for that sim, yes, it's X-Baffle. I thought I had noted that it was and also the fact that it's not perfect - maybe I left that part out. Sorry. But I have found it to be very close to what I measure and hear in room.
The simulation is under the following conditions;
Baffle: 965x510mm (aprox MJK size)
Woofer Height: 400mm
Dist. Rear Wall: 1500mm
Dist. Side Wall: 2000mm
Don 't know if that typical, maybe a bit on the cramped side for room placement. It's similar what I have here in my little cottage by the sea.
The reason you see the drivers going above their rated sensitivities in the midrange is room gain. In this case about 4dB worth.
As to having a driver that will do high levels at 40Hz with little power, yeah - within its x-max. As John importantly notes above. But subjectively it won't sound like it. It will sound very bass shy. Why? Look at all that mid-range energy. That's not going to sound very balanced.
So you have to tame that mid-range (and get rid of the highs) for it to sound balanced.My original point was that a driver with an sensitivity of 86dB/W is going to need a lot of power on an OB. I still think that's true. But of course, things are rarely as simple as that.
My goal was NOT make an OB that would play loud, or to use small amps
I am happy if it would do 95db at max level, and still doing better below 40hz
this was to take a different approach as I also knew about MJKs OB
I think Martin has also realized that his small version with one Alpha15 and a smaller fullrange with less sensitivity actually proved to be better than the big one with double Aphas ... at least I think I read he admitted that in another thread
you know, we dont all use fullrange and small tube and chipamps, or play insanely loud, or bottomless low ... thats not OB
I am happy if it would do 95db at max level, and still doing better below 40hz
this was to take a different approach as I also knew about MJKs OB
I think Martin has also realized that his small version with one Alpha15 and a smaller fullrange with less sensitivity actually proved to be better than the big one with double Aphas ... at least I think I read he admitted that in another thread
you know, we dont all use fullrange and small tube and chipamps, or play insanely loud, or bottomless low ... thats not OB
tinitus said:My goal was NOT make an OB that would play loud, or to use small amps
I am happy if it would do 95db, and still doing better below 40hz
this was to take a different approach as I also knew about MJKs OB
I think Martin has also realized that his small version with one Alpha15 and a smaller fullrange with less sensitivity actually proved to be better than the big one with double Aphas ... at least I think I read he admitted that in another thread
you know, we dont all use fullrange and small tube and chipamps
I liked the size of the newer OB design and will be reconstruct the older version with a smaller baffle. I will still use two Alpha's per side. I am trying to decide between OB and H frame. The Lowther driver with two Alpha's is a big step up from the smaller OB I designed as a sample problem.
I think that the principles I laid out for the design studies I have presented also apply to multi-driver designs. I would not change the design philosophy of woofer Qts and efficiency, baffle size, and crossover points. I also don't use a tube amp or chip amp, three 200 watt SS amps and an active crossover do the trick.
MJK said:
Regardless of what you are playing, be it a test tone or music, you simply have to displace enough air to reach that frequency at that point. If you are playing music that has tones over a range from 40hz up to 20KHz the requirements are still the same. To reproduce a kick drum in the 40-50Hz range at a given SPL requires the same amount of excursion as to reproduce a test tone at 40-50hz. If there is electric guitar at say 300hz at the same time, that requires much less excursion to reach the same SPL. However, the electric guitar at 300hz is totally seperate from the kick drum. You need to account for enough displacement at the lowest frequencies you plan to reproduce.
John
John_E_Janowitz said:
I cannot agree with your analogy. Here is something I wrote a few years ago about Xmax.
"Suppose I have a driver with an Xmax = 1 mm. I want to know what kind of input it can take at 20 Hz without exceeding this value. Lets assume that the standard Xmax plot, as a function of frequency from a T/S based computer simulation, shows that 1 watt of input will yield a displacement of 1 mm at 20 Hz. So 1 watt at 20 Hz corresponds to Xmax for a pure sine wave input signal.
Sinusoidal Input :
--------------------
If I excite this driver with a single frequency sine wave at 20 Hz so that the maximum displacement is 1 mm, then the maximum input power is 1 watt as shown in the computer simulation plot. The driver is limited to 1 watt of input at 20 Hz before Xmax is exceeded.
Square Wave Input :
-------------------------
If I excite the driver with a 20 Hz square wave that produces a displacement magnitude of 1 mm, then the input power at 20 Hz can be greater than 1 watt before Xmax is exceeded. Hopefully, the following over simplified derivation will adequately demonstate my reasoning.
Definition of a square wave at 20 Hz :
Time = 0.000 to 0.025 sec the magnitude is +1 mm
Time = 0.025+ to 0.050 sec the magnitude is -1 mm
Using the Fourier Series to represent this wave shape yields the following sine waves that are added to reconstruct the required square wave. Plase recognize that this is an infinite series and I am only showing the first 3 terms.
1st harmonic at 20 Hz = magnitude is 1.273 mm
2nd harmonic at 40 Hz = magnitude is 0.424 mm
3rd harmonic at 60 Hz = magnitude is 0.255 mm
.
.
.
.
Look at the magnitude of the first harmonic at 20 Hz, it exceeds Xmax. But the sum of all harmonics equals Xmax.
Conclusion :
--------------
If I have done my math correctly then a 1 mm square wave would have a component that exceeds Xmax at 20 Hz. Recorded music is much more complicated then a simple square wave, it is made up of many different sine waves all adding together to generate displacement of a driver as a function of time. I do look at the driver displacement plot, calculated in my T/S based MathCad models, and track the input power that will cause Xmax to be exceeded, but I do not consider it to be a rigid limitation in a design."
My point is that you just don't add magnitudes independently. For music it will take more power to reach Xmax then the simple calculation that you originally provided.
MJK said:
What the hell is going on here as I am completely lost. Is this an explanation why we don't need xmax in a woofer? It sounds like everything is based on music being more complicated than a square wave, but in reality a square wave is the hardest thing for both a speaker and amplifier to reproduce.
John_E_Janowitz said:
I agree with you John, who's to say that music is always a complex waveform, a synthesizer can be set up to put out nearly a pure sine wave and guitar notes (and organ) are much closer to being sine like than square. They are even worse than one might think since AM on a guitar waveform produces sidebands above and below the fundamental. This is simple signal analysis as is MJK's Fourier decomposition of a square wave.
What I think many people miss is that most drivers are usable well beyond their Xmax spec because the compression is often gradual. I often look at what I call Xmax50 where the motor strength (BL) is half of its rest value. It is easy to see a woofer with only 3mm Xmax doing what looks like 6 or more mm peak when driven hard and while the bass thickens up it is nothing like hard clipping. I don't suggest normal use of a woofer overdriven like this, but it is typical of party levels with many systems. Most tolerate LF distortion fairly well, but it is easy to hear when compared to a clean reference.
Your high Xmax drivers look very good John, and your point makes perfect sense.
Pete B.
nickmckinney said:
Yes I agree Nick good point to question about those curves.
What is the baffle size used for the Sims, is it IB, or finite OB?
http://www.diyaudio.com/forums/showthread.php?postid=1516125#post1516125
Pete B.
PB2 said:
This is 100% true especially the more limited the range the driver has to cover. You can design compression into the driver through the suspension and all good drivers will have it to some degree. Rockford Fosgate back in the 80's/90's also had Faraday compressors wound into the voice coil for their high output woofers which was the most extreme I have seen yet.
When you have a true uncompressed dynamic recording it is very hard for the system to stay within xmax even at reasonable volume. I have test tracks of various recordings showing this. Many of them are not musical instruments causing the xmax problem, rather most are stomping or smacking of objects while making the music. A Flamenco dance recording for example.
What music offers is masking, heavy content in the mid and upper bass tends to mask the low bass distortion and while this allows a system to be played louder (beyond Xmax) it is not accurate.
Simple solo source material tends to be much more revealing because there is nothing for the distortion to hide behind.
Superposition of sine waves is standard signal analysis, we use it all the time and it is accurate. John's analysis is correct.
Pete B.
Simple solo source material tends to be much more revealing because there is nothing for the distortion to hide behind.
Superposition of sine waves is standard signal analysis, we use it all the time and it is accurate. John's analysis is correct.
Pete B.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.