I see a lot of info on matching drivers according to their response curve but I don't see much in terms of spl. Basically, I'm looking to figure out how to select drivers that are "equal" in terms of spl so I don't have a mid that runs out of steam way before the tweeter and things such as that. Is it just as easy as taking the sensitivity and calculating the max dB based on wattage for each driver or is it more complicated than that?
I'm just thinking that this AMT tweeter is 88dB 1w/1m and has a rating of 15w rms giving it a max rms output of ~100dB
Dayton Audio AMT Mini-8 Air Motion Transformer Tweeter 8 Ohm
This 15" sub also has a sensitivity of 88dB 1w/1m and is rated for 500w rms giving it an rms output of 115dB.
Dayton Audio RSS390HF-4 15" Reference HF Subwoofer 4 Ohm
Crossover design, room gain, and enclosures aside; and just using quick maths - in order to get 115dB out of those tweeters, I would need 6 of them just to keep up with the single 15". Surely that can't be right....?
I'm just thinking that this AMT tweeter is 88dB 1w/1m and has a rating of 15w rms giving it a max rms output of ~100dB
Dayton Audio AMT Mini-8 Air Motion Transformer Tweeter 8 Ohm
This 15" sub also has a sensitivity of 88dB 1w/1m and is rated for 500w rms giving it an rms output of 115dB.
Dayton Audio RSS390HF-4 15" Reference HF Subwoofer 4 Ohm
Crossover design, room gain, and enclosures aside; and just using quick maths - in order to get 115dB out of those tweeters, I would need 6 of them just to keep up with the single 15". Surely that can't be right....?
Hey KCD!
Good question.
The answer is a little complicated, so let's take a 2-way, and we can even use my online example, if you want to play with a simulated crossover here:
A Speaker Maker's Journey: The LM-1 Bookshelf Version
In general, in a 2-way, you want to end up with a tweeter that is equal or higher in sensitivity to the woofer. Why? Power. If you pad the woofer down with resistors, you waste a lot of power in heat. Tweeters take very little power, in terms of music energy, so they are less wasteful to pad down.
Looking at a woofer's sensitivity in free-air, or even in your cabinet does not tell you where you'll end up. Generally, you trade bass for efficiency, so you end up loosing a lot of that to get more bass. An 89 dB woofer may loose 3-6 dB in the crossover to get deep enough. So you end up with 83 around the crossover point.
THAT is where you want to match your tweeter. 🙂
Does this help?
Take a look at the crossover files for the LM-1 and you can see all of this happening.
Best,
E
Good question.
The answer is a little complicated, so let's take a 2-way, and we can even use my online example, if you want to play with a simulated crossover here:
A Speaker Maker's Journey: The LM-1 Bookshelf Version
In general, in a 2-way, you want to end up with a tweeter that is equal or higher in sensitivity to the woofer. Why? Power. If you pad the woofer down with resistors, you waste a lot of power in heat. Tweeters take very little power, in terms of music energy, so they are less wasteful to pad down.
Looking at a woofer's sensitivity in free-air, or even in your cabinet does not tell you where you'll end up. Generally, you trade bass for efficiency, so you end up loosing a lot of that to get more bass. An 89 dB woofer may loose 3-6 dB in the crossover to get deep enough. So you end up with 83 around the crossover point.
THAT is where you want to match your tweeter. 🙂
Does this help?
Take a look at the crossover files for the LM-1 and you can see all of this happening.
Best,
E
I think you are nearly right with the math.
Otherwise, 100dB is already ear deafening loud in average home environment, where this tweeter planned to be used. Maybe for short periods (e.g. musical peaks) it gives higher SPL without self destruction, but i bet, distortion is showing up quickly anyway.
If you want very high (100+dB) SPL for treble range, use a good compression driver.
Last edited:
You have to consider some things.
First in a speaker you have to use only one tweeter, in order to eliminate destructive interference between drivers at high frequencies. As a consequence you need to have a tweeter with a higher SPL than the woofer, so you can either add more woofers and/or pad down the tweeter in order to match the SPL of the drivers at crossover point. As Erik said padding down a woofer is a big no-no. Second point is that in order to match SPL you have to compare 2.83V measures and not 1W ones.
With all that in mind, the woofer has higher sensitivity than the tweeter as it is 91dB@2.83V vs 88 of the tweeter, so based on that the drivers aren't matched well. But looking at the SPL of the woofer alone doesn't tell the whole story as you in general want to partially or fully compensate with the crossover for the baffle step effect. This effects only the woofer part so you generally end up with a woofer with an effective SPL of 3-6dB down from the nominal value, and in your case it will be between 85 and 88dB@2.83V. The amount of compensation of the baffle step effect depends on the placement of the speaker, but is any case the chosen drivers could be made to work together.
Only at this point I'd look at max SPL capabilities. Unfortunately RMS power handling numbers are pretty useless because they measure voice coil meltdown or something like that and you really want to look deeper at x-max values (i.e. physical point of high distortion) and box/tuning for the woofer and crossover point and slope for the tweeter.
What I usually do is that after I found well matched drivers based on SPL @2.83V (and crossover point...), I run a simulation on Unibox for chosing the box volume and tuning, and this will set the maximum output capability of the woofer. In all the speakers I built or simulated the maximum output of the woofer wasn't really a constraint for the tweeter, but if you end up with a number bigger than 100dB you should choose a large diameter dome with high SPL and sufficient x-max, or better a CD.
The 3-6 dB lost in the crossover are for addressing the baffle step effect and not the bass capabilities.
Ralf
First in a speaker you have to use only one tweeter, in order to eliminate destructive interference between drivers at high frequencies. As a consequence you need to have a tweeter with a higher SPL than the woofer, so you can either add more woofers and/or pad down the tweeter in order to match the SPL of the drivers at crossover point. As Erik said padding down a woofer is a big no-no. Second point is that in order to match SPL you have to compare 2.83V measures and not 1W ones.
With all that in mind, the woofer has higher sensitivity than the tweeter as it is 91dB@2.83V vs 88 of the tweeter, so based on that the drivers aren't matched well. But looking at the SPL of the woofer alone doesn't tell the whole story as you in general want to partially or fully compensate with the crossover for the baffle step effect. This effects only the woofer part so you generally end up with a woofer with an effective SPL of 3-6dB down from the nominal value, and in your case it will be between 85 and 88dB@2.83V. The amount of compensation of the baffle step effect depends on the placement of the speaker, but is any case the chosen drivers could be made to work together.
Only at this point I'd look at max SPL capabilities. Unfortunately RMS power handling numbers are pretty useless because they measure voice coil meltdown or something like that and you really want to look deeper at x-max values (i.e. physical point of high distortion) and box/tuning for the woofer and crossover point and slope for the tweeter.
What I usually do is that after I found well matched drivers based on SPL @2.83V (and crossover point...), I run a simulation on Unibox for chosing the box volume and tuning, and this will set the maximum output capability of the woofer. In all the speakers I built or simulated the maximum output of the woofer wasn't really a constraint for the tweeter, but if you end up with a number bigger than 100dB you should choose a large diameter dome with high SPL and sufficient x-max, or better a CD.
That's not really true. Trading bass for efficiency is done at the driver design stage point by the manufacturer. That is because efficiency, deep bass, and (low) volume displacement are interconnected and you can have only 2 in a driver. So if you want to have deep bass in a driver you need to have either a low efficiency or a huge volume box. In PA woofers high efficiency and relatively small volume box are a must so they have not deep bass capability (relatively to the diameter), in HiFi woofers deep bass is somewhat desired and is traded with low efficiency and/or huge boxes.
The 3-6 dB lost in the crossover are for addressing the baffle step effect and not the bass capabilities.
Ralf
Music isn't one sine tone for each driver, it is closer to a spectrum, similar to pink noise. Each driver gets a chunk of it. Give it a larger chunk and it is strained harder and can play less loud.
Most of the power of music is at low frequencies. Added to that, many people like exaggerated bass, increasing the demand on low frequency drivers.
I guess in practice most designers pick their drivers based on previous experience with similar drivers and then test whether it works. If not, there is some room to shift power from one driver to another by shifting the crossover frequency. Increase the crossover between mid and tweeter and the tweeter gets less power, for example.
Most of the power of music is at low frequencies. Added to that, many people like exaggerated bass, increasing the demand on low frequency drivers.
I guess in practice most designers pick their drivers based on previous experience with similar drivers and then test whether it works. If not, there is some room to shift power from one driver to another by shifting the crossover frequency. Increase the crossover between mid and tweeter and the tweeter gets less power, for example.
Last edited:
There are 2 different things to look at here:
1 - driver sensitivity requirements: Always start with the woofer and then subtract the expected baffle step loss. This will now be the minimum sensitivity the other drivers need to match and it doesn't hurt if you have a little extra dB to play with just in case some adjustments are necessary.
2 - max SPL: First decide on your requirements. What is the max SPL that you need in your room at your LP? And at what frequencies? If you don't know, grab a SPL meter and do some measurements. Most people think they need a whole lot more than they actually do, but a little headroom never hurts either. In general, expect to get less SPL out of a 2-way than from a bigger speaker.
Once you've got your targets, model your drivers to see what the max SPL is at your lowest frequency target before xmax is exceeded. Now make sure that the driver is spec'd to handle the power required to do so and at least another 3dB of headroom. And just because a driver is rated at 500W doesn't mean you have to use all of it.
This will work for the mid and woofers. For a tweeter, just go directly to the sensitivity rating and calculate max SPL with its max power rating. Again, best to have some headroom left over.
This is a slight simplification, as you may or may not be dealing with baffle step loss and there is also a difference in SPL at 1m vs say, 2 or 3m at the listening position. And it will help if your modeling can add in any HP filters which will strongly affect a driver's excursion demands and thus max SPL.
1 - driver sensitivity requirements: Always start with the woofer and then subtract the expected baffle step loss. This will now be the minimum sensitivity the other drivers need to match and it doesn't hurt if you have a little extra dB to play with just in case some adjustments are necessary.
2 - max SPL: First decide on your requirements. What is the max SPL that you need in your room at your LP? And at what frequencies? If you don't know, grab a SPL meter and do some measurements. Most people think they need a whole lot more than they actually do, but a little headroom never hurts either. In general, expect to get less SPL out of a 2-way than from a bigger speaker.
Once you've got your targets, model your drivers to see what the max SPL is at your lowest frequency target before xmax is exceeded. Now make sure that the driver is spec'd to handle the power required to do so and at least another 3dB of headroom. And just because a driver is rated at 500W doesn't mean you have to use all of it.
This will work for the mid and woofers. For a tweeter, just go directly to the sensitivity rating and calculate max SPL with its max power rating. Again, best to have some headroom left over.
This is a slight simplification, as you may or may not be dealing with baffle step loss and there is also a difference in SPL at 1m vs say, 2 or 3m at the listening position. And it will help if your modeling can add in any HP filters which will strongly affect a driver's excursion demands and thus max SPL.
Ralf:
I disagree but I'm to tired to argue with you about it. There are just too many simple examples which prove my point.
Be well, enjoy your time here.
Erik
I disagree but I'm to tired to argue with you about it. There are just too many simple examples which prove my point.
Be well, enjoy your time here.
Erik
Erik,
sorry, but disagreeing without showing proof of the contrary is only a waste of time.
And while your ending statement is partially true (losing 3-6dB from spec), your argument is not. You are mixing (or at least your sentence shows that) bass capabilities derived from T/S parameters, and baffle step effect. Since baffle step occurs, depending on baffle width, at a point in the hundreds of Hz range, if you want to address the effect you are simply setting the total SPL of the woofer to a SPL that the woofer mounted in that particular baffle exhibit somewath in the above range. You do it with the crossover, and in this case you are trading bandwidth for efficiency: you can have a full bandwidth from the woofer (say 50 to 3000 Hz), at a reduced SPL level (3-6dB less than nominal), or a highly reduced bandwidth (say 500 to 3000 Hz) at nominal SPL level. Cutting the 50 to 500 Hz range, more than 3 octaves, is hardly losing some bass, is losing all bass and part of the midrange. I have never seen a woofer, used as a woofer, high passed at the baffle step point, because this will become a midrange.
All of this could not be interesting for the OP, but your assertion was imprecise and I felt a clarification was due.
Ralf
sorry, but disagreeing without showing proof of the contrary is only a waste of time.
And while your ending statement is partially true (losing 3-6dB from spec), your argument is not. You are mixing (or at least your sentence shows that) bass capabilities derived from T/S parameters, and baffle step effect. Since baffle step occurs, depending on baffle width, at a point in the hundreds of Hz range, if you want to address the effect you are simply setting the total SPL of the woofer to a SPL that the woofer mounted in that particular baffle exhibit somewath in the above range. You do it with the crossover, and in this case you are trading bandwidth for efficiency: you can have a full bandwidth from the woofer (say 50 to 3000 Hz), at a reduced SPL level (3-6dB less than nominal), or a highly reduced bandwidth (say 500 to 3000 Hz) at nominal SPL level. Cutting the 50 to 500 Hz range, more than 3 octaves, is hardly losing some bass, is losing all bass and part of the midrange. I have never seen a woofer, used as a woofer, high passed at the baffle step point, because this will become a midrange.
All of this could not be interesting for the OP, but your assertion was imprecise and I felt a clarification was due.
Ralf
I've never seen a need to peak match drivers like you are doing and personally think its a waste of time and effort, the sound content is what will dictate if thats ever needed and some how I doubt if you will ever find anything that plays that way. And if you are worried, just find a midrange and tweeter that will over perform db wise the woofer and you are set.
- Status
- Not open for further replies.