I really can't understand why you (both) cannot see how wrong you are. The purported benefits are only realized in the passband. For a subwoofer application using a low Qts driver you are essentially operating the driver out of (below) its passband where it is much less efficient.
Last edited:
Charlie,
You need to look at efficiency NOT voltage sensitivity. I don't know how better to explain it. The graph that you posted says nothing about the system efficiency without the context of the system impedances. It is an unknown. The statement that the high BL driver from your comparison would need more power is factually incorrect. It will use less power overall. This is the only reason I took the time to post in this thread is because this long held fiction is wrong and still commonly believed.
Nothing in your posts indicate that you understand how the efficiency of a speaker is different from the frequency response shape.
Going back to HIL...Let's say that we have a 100L sealed enclosure. We can chose to use something like an B&C 18DS115 18". Yes it will be down something like 16-18dB by 20Hz. Or we can follow the traditional way of thinking and find a driver to get the response flat and the corner as low as possible. By necessity this FR requirement requires using a driver or drivers with lower overall efficiency to do this. It's in the AES papers that were linked, or a few quick simulations in a modeling program like HR or even winisd will show this. Hint if you are looking at just the voltage response you aren't looking at the right things.
Last edited:
Any loudness that arises from Qts variation has to be illusory, at least as the concepts are ordinarily used. So we are back to unhelpful terms. I believe that any "enclosure" that traffics in resonances and low Qts to gain loudness have to be crummy concepts (and you know who you are).
The allusion to torque curves versus horsepower versus elapsed time is specific to IC motor behaviour and their multi-dimensional characteristics. But torque is horsepower corrected for speed. So you could say a mouthful like "average HP in the relevant speed band.... versus car weight....". Of course, for speakers maybe we also need a mouthful like that.
Footnote: I agree that the Iron Law has no basis, except as a rule of thumb.
B.
Here is what should be a pretty clear example. I used SL's SplMax worksheet that I linked to in a recent post.
The driver that is already populated in that spreadsheet is the Peerless 830500. It's a 12" subwoofer with BL=17.2 N/A and Qts=0.2. I simply changed the BL value to 12 N/A to raise the Qts to around 0.4. I then plotted the SPL values predicted in each case on the same graph. Everything else is exactly the same about the driver, box volume, input voltage, etc. The plot is attached.
I cannot read this plot in any other way except to say that, in the range of 20Hz - 80Hz, the LOWER BL DRIVER HAS HIGHER VOLTAGE SENSITIVITY. That also means that, over that range of frequencies, it is more efficient.
It doesn't matter what is happening outside this frequency range if the driver will be used as a subwoofer - the higher frequency portion of the driver's response is not being used at all in this application.
I can only guess that people are mesmerized by seeing a high efficiency listed in tables of TS parameters for a driver, without thinking about whether that is really relevant to the application at hand. In this case, it's not. The "efficiency" that is listed on a driver datasheet refers to the PASSBAND efficiency, where the passband is defined as the region above the un-EQ'd F3 of the system or, more generally, "mid band".
The driver that is already populated in that spreadsheet is the Peerless 830500. It's a 12" subwoofer with BL=17.2 N/A and Qts=0.2. I simply changed the BL value to 12 N/A to raise the Qts to around 0.4. I then plotted the SPL values predicted in each case on the same graph. Everything else is exactly the same about the driver, box volume, input voltage, etc. The plot is attached.
I cannot read this plot in any other way except to say that, in the range of 20Hz - 80Hz, the LOWER BL DRIVER HAS HIGHER VOLTAGE SENSITIVITY. That also means that, over that range of frequencies, it is more efficient.
It doesn't matter what is happening outside this frequency range if the driver will be used as a subwoofer - the higher frequency portion of the driver's response is not being used at all in this application.
I can only guess that people are mesmerized by seeing a high efficiency listed in tables of TS parameters for a driver, without thinking about whether that is really relevant to the application at hand. In this case, it's not. The "efficiency" that is listed on a driver datasheet refers to the PASSBAND efficiency, where the passband is defined as the region above the un-EQ'd F3 of the system or, more generally, "mid band".
Attachments
Last edited:
The present discussion is not about cars or comparing large enclosures with smaller ones unless I'm mistaken. Larger enclosures are more efficient especially at low bass frequencies and have many advantages. Small enclosures suffer penalties in a number of areas. This is correct and was quantified decades ago.
The engineering going on in many modern drivers is geared towards how to make the transducer more efficient, louder, linear, rugged and extract as much performance as possible from the amount of space available. The goal is more performance from less space. This is why the drivers are the way they are. It is more expensive to make and harder to design a very low Qts woofer, with inductance countermeasures especially in the large 18 and 21" sizes. It is much easier and cheaper to make a traditional ferrite based, moderate to highish Qts woofer with no Faraday rings.
Extremely lightweight is always advantageous on paper. It is not always advantageous in reality. Light weight is also an indicator of fragility and potential weakness. Bass speakers are driven much harder than any other speaker in the chain. I can tell you numerous tales of lightweight cones becoming damaged, ripping, flexing and distorting heavily, or low mass voice coils heating up and losing 3+dB of sensitivity or cooking themselves to death.
The damping and restorative forces of the driver are expressed in the electrical and mechanical parameters. Heavy Mms causing "slowness" or "overhang" or other subjective problems with the sound is yet another misunderstood myth that is widely believed. it does lower efficiency, but like any other characteristic of a complex system it cannot be looked at in a vacuum. It must be considered against suspension compliance and motor force, etc. The driver specifications literally take all of this into account, but the myth continues.
The first statement is correct the second statement however, is wrong.
I don't think anyone in this discussion is talking about datasheet efficiency or sensitivity specs. I would assume everyone here knows those should be taken with a grain of salt or two and that it varies with frequency and is thinking in terms of a response measurement over a broad bandwidth.
Charlie, that doesn't show max SPL. That's at some nominal voltage (2.83 Vrms?). I am not reading anyone arguing that point. Voltage sensitivity isn't efficiency. Never has been, never will be.
If I can channel Josh (and hopefully he'll correct me where I go astray), the idea is that we have a preponderance of inexpensive high power amplifiers that can swing outrageous voltages, so don't let that be some sort of arbitrary restriction. So if the effective impedance of a high-BL/low Qes driver in a smaller box is higher, yet requires *less* power, it is more efficient. Less voltage sensitive, sure, but higher efficiency. We're ultimately going to be limited by Xmax.
If I can channel Josh (and hopefully he'll correct me where I go astray), the idea is that we have a preponderance of inexpensive high power amplifiers that can swing outrageous voltages, so don't let that be some sort of arbitrary restriction. So if the effective impedance of a high-BL/low Qes driver in a smaller box is higher, yet requires *less* power, it is more efficient. Less voltage sensitive, sure, but higher efficiency. We're ultimately going to be limited by Xmax.
Max SPL? That's related to Xmax and Sd. Anyway, I made my point clear - the driver will produce more SPL in the below 100Hz frequency range per watt of input power. You can call that "voltage sensitivity" or whatever you want, but I call that "more efficient". My assumption is that the load impedance is roughly the same, so voltage and power are roughly the same thing. Maybe this is not all that accurate... but I think the sentiment is correct.
I'm not following what you mean by this word salad. Please see my former comments on "efficiency".
Last edited:
Charlie, you're unfortunately being extremely obtuse.
Throw your two examples into Hornresp, especially in a pretty small box, and look at the "efficiency" graph. It's an option under "tools" when you run calculate and look at "4 Acoustical Power" under "Window". Swapping Qes and nothing else will show the very effect being described about efficiency.
Throw your two examples into Hornresp, especially in a pretty small box, and look at the "efficiency" graph. It's an option under "tools" when you run calculate and look at "4 Acoustical Power" under "Window". Swapping Qes and nothing else will show the very effect being described about efficiency.
Old but still relevant: Hornresp for Dum... hmm... Everyone 
- Home Theater Forum and Systems - HomeTheaterShack.com
David McBean has done a great job making it quite available to all kinds of users, and a sealed box will get you out the door in 10 minutes (and give you that intuitive feel that I cannot provide).
Efficiency is just that: P_acoustic/P_electrical
- Home Theater Forum and Systems - HomeTheaterShack.comDavid McBean has done a great job making it quite available to all kinds of users, and a sealed box will get you out the door in 10 minutes (and give you that intuitive feel that I cannot provide).
Efficiency is just that: P_acoustic/P_electrical
Let's get back a little bit to the source of the discussion: do you still plan a relatively 'small' box or a large one ? How deep do you want to go.. do you really need that 20Hz ?
OK, but I'd like to add these questions:
Do you want a low distortion or high distortion subwoofer?
Do you prefer to use multi kilo Watt pro audio amplifiers with fan cooling or consumer amps up to a few hundred Watts and noiseless heat sinks?
Do you consider thermal compression to be of concern?
Do you want a low distortion or high distortion subwoofer?
Do you prefer to use multi kilo Watt pro audio amplifiers with fan cooling or consumer amps up to a few hundred Watts and noiseless heat sinks?
Do you consider thermal compression to be of concern?
I think he started just like me a year ago, but then based on many-many feedbacks and discussions I slowly re-imagined my needs and fine tuned the to-be-final project (still in design phase).
He'll need quite a lot of help but sometimes discussion sub-threads just go into very deep discussions while loosing focus of his original needs. This is good btw, tons of different opinions during this, but at the end he must have some kind of conclusion what his needs became 'cause they need to be changed I think. It's always a fun part of dreaming
He'll need quite a lot of help but sometimes discussion sub-threads just go into very deep discussions while loosing focus of his original needs. This is good btw, tons of different opinions during this, but at the end he must have some kind of conclusion what his needs became 'cause they need to be changed I think. It's always a fun part of dreaming
I was able to use the SPLmax spreadsheet to calculate efficiency, and I WAS WRONG ABOUT THE EFFICIENCY! Read on for more details...
Attached are three plots. In the plot with the lighter background, the orange and blue curves are the power (Vrms*Irms) needed to reach Xmax for each case. The blue curve is the "high BL/low Qts" version of the driver and the orange curve is the "lower BL/higher Qts" version of the driver. The only thing that was changed was the BL strength of the motor and all other parameters remain the same. It's obvious that the blue curve is always below the orange curve, which means that the high BL motor will always require less power to reach Xmax. Wow, OK, that is new to me and very interesting. The gray curve shows the percent less power that is needed by the high BL version of the driver. This drops to a minimum of about 10% around resonance, but more significantly rises to 25% by 20Hz and keeps climbing to 50% by 10Hz.
Ok, the driver with higher BL is more efficient. I never bothered to look at it this way, and I can't argue with the numbers here.
But let's look at this a little more. I have attached two additional plots of voltage and current required to reach Xmax as a function of frequency for each case. The lower BL motor requires much more current at the lowest frequencies, as much as 45% more current. On the other hand, the high BL motor requires higher voltage especially around resonance where it needs almost 60% higher rail voltage to reach Xmax.
If you really want to push the driver to Xmax at the lowest frequencies (20Hz and below) in a closed box then the high BL system is clearly better in terms of efficiency. I had not looked at this angle before, so it is a surprise to me. The lower current demand is a strong benefit because it drives VC heating. But at higher frequencies you will really need a very high power amplifier (into 4R/8R) because of the need to apply higher voltage to reach Xmax. In this case that is around 35Hz, but this will change depending on the specifics of each driver and box. I might plug the specs from the iPAL into the SPLmax spreadsheet and check it out in a little more detail to learn more.
I can now see why the high BL subwoofer might have some pluses, especially for the lowest bass, but will likely require pro amps to reach its potential. I am sure that is fine in many cases but not my cup of tea (but maybe I'm learning something new here). Now that there are some high Xmax, high BL drivers available there are some interesting opportunities.
Attached are three plots. In the plot with the lighter background, the orange and blue curves are the power (Vrms*Irms) needed to reach Xmax for each case. The blue curve is the "high BL/low Qts" version of the driver and the orange curve is the "lower BL/higher Qts" version of the driver. The only thing that was changed was the BL strength of the motor and all other parameters remain the same. It's obvious that the blue curve is always below the orange curve, which means that the high BL motor will always require less power to reach Xmax. Wow, OK, that is new to me and very interesting. The gray curve shows the percent less power that is needed by the high BL version of the driver. This drops to a minimum of about 10% around resonance, but more significantly rises to 25% by 20Hz and keeps climbing to 50% by 10Hz.
Ok, the driver with higher BL is more efficient. I never bothered to look at it this way, and I can't argue with the numbers here.
But let's look at this a little more. I have attached two additional plots of voltage and current required to reach Xmax as a function of frequency for each case. The lower BL motor requires much more current at the lowest frequencies, as much as 45% more current. On the other hand, the high BL motor requires higher voltage especially around resonance where it needs almost 60% higher rail voltage to reach Xmax.
If you really want to push the driver to Xmax at the lowest frequencies (20Hz and below) in a closed box then the high BL system is clearly better in terms of efficiency. I had not looked at this angle before, so it is a surprise to me. The lower current demand is a strong benefit because it drives VC heating. But at higher frequencies you will really need a very high power amplifier (into 4R/8R) because of the need to apply higher voltage to reach Xmax. In this case that is around 35Hz, but this will change depending on the specifics of each driver and box. I might plug the specs from the iPAL into the SPLmax spreadsheet and check it out in a little more detail to learn more.
I can now see why the high BL subwoofer might have some pluses, especially for the lowest bass, but will likely require pro amps to reach its potential. I am sure that is fine in many cases but not my cup of tea (but maybe I'm learning something new here). Now that there are some high Xmax, high BL drivers available there are some interesting opportunities.
Attachments
Hey glad you played around with the sims, Charlie, and can understand at least what such a woofer design can do for you. It may not be what you want to design for, but the lightbulb went on for me when I started looking at these kinds of woofers for home applications. Class D amps such as the TPA 3255 I already own seem to be able to get to xmax-limited output (and above my wildest desires) with a 15"/18" smaller boxes than I could achieve with more standard 12" driver. That's like a 20x20x16 outside dimension sealed box for an 18", which will need some creative bracing/placement to keep something like that from walking. But that kind of packaging has an appeal.
You get a lot more for your money with a pro amp:
Review and Measurements of Crown XLS 1502 Amp | Audio Science Review (ASR) Forum
Behringer inuke NU6000 vs KAM KXD7200 bench tested - Speakerplans.com Forums - Page 1
2kW modules are available from hypex, you could put these in a box and only have a fan come on if they got hot:
Hypex Electronics B.V.
I could design a multi kw class D that didn't need fans but I don't have any issues with the cheap pro amps.
Review and Measurements of Crown XLS 1502 Amp | Audio Science Review (ASR) Forum
Behringer inuke NU6000 vs KAM KXD7200 bench tested - Speakerplans.com Forums - Page 1
2kW modules are available from hypex, you could put these in a box and only have a fan come on if they got hot:
Hypex Electronics B.V.
I could design a multi kw class D that didn't need fans but I don't have any issues with the cheap pro amps.
Excellent!
Here is another thought experiment along these lines...
We know that the higher BL driver produces more output regardless of frequency with equal input power. This stays true after EQ is applied. Even if you use a greatly undersized amplifier the high BL driver should provide more overall headroom from the amplifier. You may run into voltage limitations rather than power or current where the impedance is high and the superior voltage sensitivity of the low BL driver might allow it to have greater headroom over a small range of frequencies near the impedance peaks where the amp is voltage limited.
However there are a few other things to consider.
Most content we listen to is complex, harmonically rich, wide bandwidth and varied. It isn't single tones. Your amplifier only has so much current, power and voltage capability. Once one of those limitations is met that's it. During a movie soundtrack it's not uncommon to have something like the following scenario: an explosion and rumble down near 20-25Hz, really loud gunshots occurring over it with a 40Hz fundamental and other noises up near 90Hz.
Systems usually respond in the most linear, compression free manner, near impedance peaks where they are most efficient. Distortion is also usually very low in this same area. Large amounts of current through a voice coil increases inductive and BL related distortions. It also causes heat in the voice coil and motor which increases resistance and lowers the sensitivity. Output compression occurs. The high BL driver will produce the same amount of output with less power and less current through the coil, resulting in less heating of the coil and less thermal related compression effects. It should also result in less distortion.
I just hate the sound of ports, I like tight, controlled and non resonant bass. I'm not using them in a club, a concert house or anything, just on my living room, and it is not big. I just think I can get a much greater bang for my buck with 18 inch LaVoce or B&C subwoofers than with those 12 inch drivers. And I won't use much power, just enought to reach their excursion limit at 20 hertz, 500 watts just will do it in a 108 liters box, I will not need much cooling.
Thanks for the recommendation, I can get a 500 watts Hypex plate amp for 330€. Does this Crown have DSP? Can I set my response curve flat at 20 hertz with it?