Here is my thought; the Impulse from the speaker, is the carrier signal. The electrical signal is now the modulating signal. The worse the carrier signal the worse the modulating signal. The best transient in this graph is the Raw driver (Red line). At what point which Flaw becomes audible, Overhang Time versus Attack time, is the next discussion.
So sound travels 2.5m in 7.29 ms? Which relates to approximately 137hz... So distortion will increase after 137hz... We can change overhang by High passing LF... you could design the LF crossover point and the midrange rolloff around this idea.
So sound travels 2.5m in 7.29 ms? Which relates to approximately 137hz... So distortion will increase after 137hz... We can change overhang by High passing LF... you could design the LF crossover point and the midrange rolloff around this idea.
Last edited:
Not really following you........
Transient response = speed, faster = wider BW, ergo a 'faster' transient = higher peak SPL for a given frequency and the lower its inductance, the wider its BW till the driver's vibration inertance/'stiction' rolls it off.
The added mass actually damps the driver's resonances above its upper mass corner (Fhm) where Le dominates, ergo no Qes, BL, mass ratio theory and above Fhm arguably better performing.
Increase inductance though and the HF response begins rolling off above Fhm to a lower HF corner frequency.
All this is clearly seen in the driver's measured response(s) in the 3rd plot you didn't post.
Transient response = speed, faster = wider BW, ergo a 'faster' transient = higher peak SPL for a given frequency and the lower its inductance, the wider its BW till the driver's vibration inertance/'stiction' rolls it off.
The added mass actually damps the driver's resonances above its upper mass corner (Fhm) where Le dominates, ergo no Qes, BL, mass ratio theory and above Fhm arguably better performing.
Increase inductance though and the HF response begins rolling off above Fhm to a lower HF corner frequency.
All this is clearly seen in the driver's measured response(s) in the 3rd plot you didn't post.
I think I can word this a little better.
Flaws to criticize; Attack time (time from take off to peak), Time and intensity of the Overhang and Overshoot, time until completely settled.
I think this also leads us back to excursion and velocity and the idea that impulse is worsened with increasing velocity. We coincidentally listen within thiele/small parameters at short listening distance distances. I always felt that an anechoic measurement at peak levels should be common, but non the less, this IR represents the consequence of the transients. The larger and longer overshoots and overhangs. Bass, generally will be the more constant source of excursion... higher velocities bring higher overshoot and overhang, which are worse if BL is low compared to mass but without mass be get deformity of the cone....
For a sub - You'd want a cone that was only heavy enough to hold shape at anticipated highest velocities, with enough BL to counteract inertia, and to low pass at the limitations of the resulting flaws.
For a mid - You'd basically want all the same things but you might high pass in order to create enough resolution.
Removing excursion, Obviously is the key. One side of the Spectrum would be a light coned system where all large excursion has been remedied.
Why don't we say that Le and Mass both dampen the HF sensitivity. Le and Mass both dominate, Le did not defeat mass, just like a super light cone can't defeat high induction
"Transient response = speed, faster" _ exactly except you havent stated that the Raw driver is performing faster than the Weighted
View attachment 1258034
Last edited:
The quote below (which is now correct) was originally and incorrectly showing as being by camplo. Not sure how that cam e to be, however...
Once again, these measurements are incorrect. The added mass should reduce a peak value of the IR graph compared to the raw driver. While both the frequency response in the Hi and Mid range and initial part of the IR graph of the driver with the added mass is almost is the same as the raw driver. This contradicts to fundamental laws of physics. The higher the mass is , the lower driver efficiency is (lower efficiency leads to lower sensitivity and lower peak value of IR graph ).
Last edited by a moderator:
Not only stiffness is important, but also damping. Ideal pistonic motion of the cone can be achieved if we have a very high damping value. I once upon a time considered the breakup modes in a typical paper cone, and compared the effect of increasing the Young's modulus and damping on the breakup of the cone. In both cases, in a membrane with a low Young's modulus but high damping value and with a high Young's modulus but low damping, we have almost the same frequency response (the picture below). Another thing is that a high value of damping is associated with the viscoelastic behavior of materials, which is by no means a simple issue, and there is a separate branch of physics called rheology, which in one or another way studies the viscoelastic properties of matter.
But typical DIY user hardly needs to dive deeply into various branches of physics in order to make the right choice of speaker. It seems to me that focusing on the ability to interpret measurement results and understanding some fundamental principles of electroacoustics will be much more useful for the DIY user.
Note that you provided the power response graph which is not the same as the on-axis SPL response
Have you browsed the books I mentioned before?
Because the Mms and Le effects are described in there very well, incl pictures and measurements.
Practically speaking, just look at the freq resp graphs as well as distortion.
That's all we need to know.
Again, for home use I very much doubt you will notice any differenced.
Because the Mms and Le effects are described in there very well, incl pictures and measurements.
Practically speaking, just look at the freq resp graphs as well as distortion.
That's all we need to know.
Again, for home use I very much doubt you will notice any differenced.
Why don't we say that Le and Mass both lower the HF sensitivity. Le and Mass both dominate, Le did not defeat mass, just like a super light cone can't defeat high induction in terms of HF sensitivity.
"Transient response = speed, faster" _ exactly except you haven't stated that the Raw driver is performing faster than the Weighted
The raw driver is ahead of the weighted driver, I highlighted it in the areas that it is. The impulse isn't the same as the steady state, but it should give some clues about it. It wants to do what IR does on every Cycle, its just that the voltage is defeating it. Every new transient which is every peak and trough, it is influenced by the IRs character, to some degree. In the above example they added pennies on the dollar, there are 300g mms drivers with the same BL as another with a 190g mss driver, those areas in yellow only get farther behind as mass increases but 28g is not enough to super crazy is all. Look at the FR below.... they barely made a dent in the HF sensitivity with le or Mass.... All 3 examples have the ability to play up to the same frequency... not the best example to show effect. Increasing mass will lower Q specs if only mass alone is increased. Is that enough said? These penalties will only be exaggerated in the outskirts of excursion, especially with additional mass void of additional damping and motor to control that mass.
I am proposing that, the decay of a system can also detiriate transient response. I am not claiming to no all the details but if there is residual energy in the system from previous impulses, it would seem that they would interfere with the real time signal.
I understand Bl Qes and Le better now. I see that increasing Mass will lower Damping of the system. No one is in denial of the effects of Q on the transient but in this case, Q is in check. Q is the check actually. Regardless of what Mms/Mmd says, no matter what BL says or Le apparently, if Qtc is 0.22 Things are obviously dampened. Q specs represent the system that those other factors help to cause. That's one point I kept missing. So you learned me there... But Mass and BL being completely irrelevant to transient response? I could see this being true; Mass does not have an effect on transient response within the masses commonly used. Inductance and Mass are related. It seems they both want to resist change in direction of energy. Still you say, mass doesn't matter...
Lets start here.... The BL to mass ratio is ridiculous on a 6.5" driver in comparison to 18"s and 20" which I've been studying for the last 2 days. 0.53BL per gram?!?!?! Unheard of for a large woofer. The driver in this test has 0.39Bl per gram. Unheard of in a large Woofer. Mass loaded test, additional 28.5g bring us to 0.17BL per gram... Thats a high performance 18" lol.
If I recall, Mass and Le are linear effects.... So basically if doubling mass doesn't do much of nothing to the FR in the example, similar results are had with 15's and 18's. So the main issue if there was one, would not be in the acceleration, but in the decay and worst at high excursion. As drivers get bigger, Sd of the cone increases faster than the surface area of the suspension. Bl per gram also lower. Acceleration decreases due to lowered acoustical impedance but Cone mass increases.
Its almost like this is the only thing worth looking at. The one thing I see, is energy gets larger with lower resonances. I cannot predict the decay of 30g vs 300g with matching Qes specs. HR might give some insight I guess but seems to not be completely accurate on these matters. Here Is 720g qts0.17 Purple versus 72g qts0.17 Green. It is missing the attenuation, the impulses aren't power matched, Increasing Mass should lower efficiency.
Here Is 720g qts0.17 Purple versus 72g qts0.10 Green
720g Qtc=0.17 Purple vs 720g Qtc=0.34 Green
I think the missing element is FR matching below. Until you match SPL you aren't seeing the final transient response. Thats if what we see below is the representation of each response IR we saw earlier. The already bad Decay of the Weighted driver will be even worse than what's seen here. I believe the difference between a lot of drivers will be at higher velocity
SO guess the goal is 16of these
Trying to follow the last few pages, but I have to admit that none of it makes much sense to me. What is it that you are trying to understand Camplo? There are too many errors in your comments to correct them all - I don't have all day.
PS. The GW-1858 data seems like nonsense since the impedance peak and the response peak are way off of each other and they should be at the same frequency.
PS. The GW-1858 data seems like nonsense since the impedance peak and the response peak are way off of each other and they should be at the same frequency.
Yeah, that's certainly been my experience, but not familiar with any of these fancy measurement programs and considering the source and the many experts that agree with this 'proof' have just assumed it was either some special type measurement or tweaked for easy graphic comparison.
Me too, 'took a shot', now time to move on. Hopefully not too many errors in my responses.
Yeah, they removed it from their site some time ago, but a couple of actual DIY measured ones I've seen show each graph as accurate enough.
@gedlee @GM Im trying to understand the chain of events. I'm not sure what errors I spoke but I concluded that Le is responsible for transient response. Now I am able to agree with you guys, without just blindly following. You guys are on another level, you see errors in thiele specs, like a woofer whisperer.
The other thing I spoke on, that I'm not sure how you guys think, is the issue of the performance of decay versus transient response. I think that the worse decay is, the worse transient response is. I cannot suggest to what degree. I was theorizing that the decay character of the IR is present to some degree at every oscillation. Voltage of the source can
Dampen this, and how it is expressed in the signal will be complex. This was relevant to experiment recently discussed, as the driver with the added mass had worse decay.
The other thing I spoke on, that I'm not sure how you guys think, is the issue of the performance of decay versus transient response. I think that the worse decay is, the worse transient response is. I cannot suggest to what degree. I was theorizing that the decay character of the IR is present to some degree at every oscillation. Voltage of the source can
Dampen this, and how it is expressed in the signal will be complex. This was relevant to experiment recently discussed, as the driver with the added mass had worse decay.
I am not sure I follow completely. The transient response is given by the shape of the transfer function / frequency response. Impulse response is just another way of displaying the same thing. That is theory. Are you trying to find out how much does the reality differ from theory or what is your aim? I mean real drivers vs TS parameters based model?
If nonlinearity of the system is weak nonlinear convolution/deconvolution technique can be applyed to separate linear and nonlinear parts of the impulse response. This is actually what REW do. The sweep sine technique is resistant to ninlinearities to a certain extent (while MLS is not).