G.Kleinschmidt said:
The voice coil inductance driving the complex impedance of the driver usually produces a downward tilt in the passband when the VC inductance is large. Some of the monster drivers with very large VC inductance never reach the passband sensitivity because the droop starts so early. This is often part of the BSC. But there is really no way to predict cone breakup, and cavity/tube resonance modes from the parameters; you need measured data. Many subwoofers are very poor in this regard, others not so bad. The AE drivers are well known for their clean response.
I was about to use this Peerless XXLS as an example of a driver with poor passband performance based on what I remember of the XLS version. It now (in the updated XXLS version) looks fairly clean up to about 500 Hz where there is a notch in the response, this new curve is probably smoothed, the old XLS was quite bad. This driver has a 2.5 ohm RDC with the coils in parallel, you might want to consider it:
http://www.madisound.com/catalog/product_info.php?cPath=45_338_381&products_id=1468
This Aura with a stiff aluminum cone has the first peaking breakup mode pushed up above 1 kHz, it also has a low RDC of 2.86 ohms:
http://www.madisound.com/catalog/product_info.php?cPath=45_338_381&products_id=1730
Both of these drivers have much less Xmax than the Shiva-X, the Aura is interesting with the underhung coil. I believe that you would need two of these also, for sensitivity and max SPL reasons.
Keep in mind that these plots are often heavily smoothed so you will want to measure the driver that you choose before you proceed. I've been surprised lately to hear that most suppliers will take returns, I've never done it myself because it doesn't seem right to me once I've opened the box and applied power.
Pete B.
Re: Re: Re: Re: Re: Half space sensitivity
Most would plan on padding the M-T by about 3 dB so that then you need about 91 dB for the woofer. Or work the other way and just calculate the pad that you need based on the woofer's sensitivity. I would not go under 88 or 89 dB for whatever configuration woofer you come up with.
Something interesting to consider is that part of what we percieve as loud is the harsh edgyness of the distortion and when you have lots of power and low distortion speakers one can tend to turn it up yet not find it to sound loud. Really, it does not sound strained. You really need to be cautious about your hearing with a system like this as I think you know.
Pete B.
G.Kleinschmidt said:
Most would plan on padding the M-T by about 3 dB so that then you need about 91 dB for the woofer. Or work the other way and just calculate the pad that you need based on the woofer's sensitivity. I would not go under 88 or 89 dB for whatever configuration woofer you come up with.
Something interesting to consider is that part of what we percieve as loud is the harsh edgyness of the distortion and when you have lots of power and low distortion speakers one can tend to turn it up yet not find it to sound loud. Really, it does not sound strained. You really need to be cautious about your hearing with a system like this as I think you know.
Pete B.
G.Kleinschmidt said:
Yes, it certainly is impressive, and keep in mind that the wall boundaries help reinforce the bass below about 50 Hz where they are close from a wavelength perspective. Being down 6 or 9 dB at 20 Hz is not so bad. On the other hand, you might want to try some EQ or an LT to bring it up at 20 Hz and see what it is like to have the walls and floor shake without having so much average SPL.
Boundary reinforcement is taken to an extreme in the NHT3.3 where the woofer is close to the floor, wall, and large cabinet side pushing the gain up higher in frequency. There are usually dips in the mid bass response when the woofer is further from the boundary due to destructive interference. The floor bounce for example.
I have actually found Allison speakers to be a bit mid bass heavy, perhaps because recordings are mixed and EQed for typical speakers, not sure. However, this was just a casual observation years ago and not in my home where I could do extended listening. In real life we hear the floor or ceiling bounce dip even just hearing a person speak or sing, close mic'ed this is lost and perhaps it is needed in the speaker to some extent. Just a theory for now.
I'm not familiar with that simulation package, does it give you displacement limited max SPL? This tends to be a limitation with sealed box systems. Unibox is a good free simulator that provides this data.
Pete B.
For when you catch up on your reading:
A link to Roy Allison's AES articles covering room effects:
http://www.classicspeakerpages.net/..._by_roy_f/technical_articles_by_roy_f_15.html
A program to simulate the effects of room placement, choose other speaker to input your woofer location:
http://www.classicspeakerpages.net/_Media/bstplac.zip
The AR-9 owner's manual which discusses some of the woofer placement considerations:
http://www.classicspeakerpages.net/...es_1978-1981/ar-9_series_manuals/ar-9_manual/
The AR-9 crossover schematic is another example of a higher order 4 way. Note that there are a few errors or revisions:
http://www.classicspeakerpages.net/.../ar-9_series_schematicsservi/ar-9_schematics/
Pete B.
A link to Roy Allison's AES articles covering room effects:
http://www.classicspeakerpages.net/..._by_roy_f/technical_articles_by_roy_f_15.html
A program to simulate the effects of room placement, choose other speaker to input your woofer location:
http://www.classicspeakerpages.net/_Media/bstplac.zip
The AR-9 owner's manual which discusses some of the woofer placement considerations:
http://www.classicspeakerpages.net/...es_1978-1981/ar-9_series_manuals/ar-9_manual/
The AR-9 crossover schematic is another example of a higher order 4 way. Note that there are a few errors or revisions:
http://www.classicspeakerpages.net/.../ar-9_series_schematicsservi/ar-9_schematics/
Pete B.
Re: Re: Re: Re: Re: Re: Half space sensitivity
Hi Pete.
I could pad down the M-T and use a single high sensitivity woofer for the bass (with all the compromises between high sensitivity and low frequency, closed box performance), but then I would just have an ordinary 3-way for an ordinary amplifier.
The woofers that work best with low fc in a closed box all have low sensitivities.
For instance, that Aura 12" has a 2V (1W/1m into 4 ohm) sensitivity of 85dB. It has both a low EBP of 59 and a low fs.
It should produce some pretty good bass in a sealed box.
With two of these in parallel, I get my target 94dB 2.83V sensitivity.
This speaker looks on paper to be an excellent candidate.
I'll run it through the sim.
PB2 said:
Hi Pete.
I could pad down the M-T and use a single high sensitivity woofer for the bass (with all the compromises between high sensitivity and low frequency, closed box performance), but then I would just have an ordinary 3-way for an ordinary amplifier.
The woofers that work best with low fc in a closed box all have low sensitivities.
For instance, that Aura 12" has a 2V (1W/1m into 4 ohm) sensitivity of 85dB. It has both a low EBP of 59 and a low fs.
It should produce some pretty good bass in a sealed box.
With two of these in parallel, I get my target 94dB 2.83V sensitivity.
This speaker looks on paper to be an excellent candidate.
I'll run it through the sim.
jupiterjune said:
Re: Re: Re: Re: Re: Re: Re: Half space sensitivity
Me again,
I've looked up the Dayton unit and the above can be shown to be correct in the data provided.
In the specitications detailed on the page here.....
http://www.partsexpress.com/pe/pshowdetl.cfm?&DID=7&Partnumber=295-464&ctab=2#Tabs
.... an SPL of 89dB is given at 2.82V/1m
However, the .pdf datasheet here.........
http://partsexpress.com/pdf/RSS315HF-4_specsheet.pdf
.....specifies the SPL in 1W/1m (or 2V/1m) - it is 3dB lower at 86dB.
Anyway, this is an excellent driver! Huge power handling, 14mm Vmax, fs=25Hz and a high Qes giving an low EPB of 48 - perfect for a closed box.
With a 2.83V/1m SPL of 89dB, I will have 95dB with two in parallel.
That gives a baffle step compensation of 5dB with the selected 90dB (2.83/1m) mid-range and tweeter - just right.
Cheers!
Glen
G.Kleinschmidt said:
Me again,
I've looked up the Dayton unit and the above can be shown to be correct in the data provided.
In the specitications detailed on the page here.....
http://www.partsexpress.com/pe/pshowdetl.cfm?&DID=7&Partnumber=295-464&ctab=2#Tabs
.... an SPL of 89dB is given at 2.82V/1m
However, the .pdf datasheet here.........
http://partsexpress.com/pdf/RSS315HF-4_specsheet.pdf
.....specifies the SPL in 1W/1m (or 2V/1m) - it is 3dB lower at 86dB.
Anyway, this is an excellent driver! Huge power handling, 14mm Vmax, fs=25Hz and a high Qes giving an low EPB of 48 - perfect for a closed box.
With a 2.83V/1m SPL of 89dB, I will have 95dB with two in parallel.
That gives a baffle step compensation of 5dB with the selected 90dB (2.83/1m) mid-range and tweeter - just right.
Cheers!
Glen
PB2 said:
OK Pete, I just Uniboxed that Dayton unit.
Even with a series RS as high as 0.5 ohms, the closed box only needs to be 75L for Fc=35Hz at Qtc=0.707.
Maximum power handling is over 200W for a single unit and Vmax limiting well beyond the desired 250Hz crossover.
This is a fantastic unit.
I think I've finally finished with the driver selection.
Cheers,
Glen
Attachments
Re: Re: Re: Re: Re: Re: Re: Re: Half space sensitivity
That Dayton does look quite good but the VD is no where near the Shiva-X as I'm sure you noticed. The Dayton has a heavier cone than most of the drivers we've considered and this suggests that it will work well in a smaller box or provide a lower Fc in the same box. Fc is determined by the moving mass, for equal sized drivers, and the box size primarily as long as Vas is significantly larger than the box volume. The air spring dominates.
I cover this a bit more in this old post:
http://www.classicspeakerpages.net/IP.Board/index.php?s=&showtopic=1090&view=findpost&p=54625
Let's consider motor linearity for a moment here. The force on the secondary of the motor is expressed as f=Bli(t), B is the magnetic field strength (a constant), l is the length of VC wire in the gap, and i(t) is the current in the voice coil. The number of turns or length of wire is constant as long as the driver does not exceed Xmax, and if we assume that B is perfectly focused in the gap, with no fringe field then the motor is perfectly linear within Xmax for this simple model. However, there is a (usually) asymmetric fringe field and other factors that cause Bl to be non-flat or highly non-linear and it is interesting to look at the measured Bl curve for different drivers. Here's the Bl curve for a 10" NHT woofer that has the same motor as the NHT1259 used in the NHT3.3:
http://home.comcast.net/~jhidley/Klippel_reports/11-075-1_LSI_Detailed.htm
This is a fairly simple motor that is commonly used, there is a trick to get more symmetry about the X=0 point which is to leave a few turns off the voice coil at one end to match the asymmetry in the fringe field, simple, elegant and it was used as far back as in the Large Advent around 1969. Audax and others sometimes use it in some of their tweeters. AR and NHT never bothered to use it as far as I know. Your probably aware that asymmetrical non-linearities cause rectification and in this case DC offset in the zero point of the cone motion ... not good.
You can also see the other driver non-linearities, most drivers are pathetic, amazing that they sound as good as they do.
Here is the XBL marketing info:
http://www.adireaudio.com/Files/XBL2DetailsPaper.pdf
and:
http://www.rawacoustics.com/CMS__About_Technology.html
I believe that this is the measured curve for the 6.5" XBL driver with about 8 or 10 mm Xmax, quite nice:
http://www.aranmaracoustics.com.au/New_ADIRE/eX6BL.gif
Correction: I believe that XBL motors are rated for Xmax at the point where BL is 71% of the max value, so this last curve is for a 13mm Xmax driver. Their Xmax is a bit over stated, but they are still dramatically more linear than most motors.
Pete B.
G.Kleinschmidt said:
That Dayton does look quite good but the VD is no where near the Shiva-X as I'm sure you noticed. The Dayton has a heavier cone than most of the drivers we've considered and this suggests that it will work well in a smaller box or provide a lower Fc in the same box. Fc is determined by the moving mass, for equal sized drivers, and the box size primarily as long as Vas is significantly larger than the box volume. The air spring dominates.
I cover this a bit more in this old post:
http://www.classicspeakerpages.net/IP.Board/index.php?s=&showtopic=1090&view=findpost&p=54625
Let's consider motor linearity for a moment here. The force on the secondary of the motor is expressed as f=Bli(t), B is the magnetic field strength (a constant), l is the length of VC wire in the gap, and i(t) is the current in the voice coil. The number of turns or length of wire is constant as long as the driver does not exceed Xmax, and if we assume that B is perfectly focused in the gap, with no fringe field then the motor is perfectly linear within Xmax for this simple model. However, there is a (usually) asymmetric fringe field and other factors that cause Bl to be non-flat or highly non-linear and it is interesting to look at the measured Bl curve for different drivers. Here's the Bl curve for a 10" NHT woofer that has the same motor as the NHT1259 used in the NHT3.3:
http://home.comcast.net/~jhidley/Klippel_reports/11-075-1_LSI_Detailed.htm
This is a fairly simple motor that is commonly used, there is a trick to get more symmetry about the X=0 point which is to leave a few turns off the voice coil at one end to match the asymmetry in the fringe field, simple, elegant and it was used as far back as in the Large Advent around 1969. Audax and others sometimes use it in some of their tweeters. AR and NHT never bothered to use it as far as I know. Your probably aware that asymmetrical non-linearities cause rectification and in this case DC offset in the zero point of the cone motion ... not good.
You can also see the other driver non-linearities, most drivers are pathetic, amazing that they sound as good as they do.
Here is the XBL marketing info:
http://www.adireaudio.com/Files/XBL2DetailsPaper.pdf
and:
http://www.rawacoustics.com/CMS__About_Technology.html
I believe that this is the measured curve for the 6.5" XBL driver with about 8 or 10 mm Xmax, quite nice:
http://www.aranmaracoustics.com.au/New_ADIRE/eX6BL.gif
Correction: I believe that XBL motors are rated for Xmax at the point where BL is 71% of the max value, so this last curve is for a 13mm Xmax driver. Their Xmax is a bit over stated, but they are still dramatically more linear than most motors.
Pete B.
G.Kleinschmidt said:
Looks good Glen, what is the excursion limited SPL at say 20 Hz? You can set the driver power rating to 1000W to get an idea of just the excursion limted SPL since the system will handle much more peak power than continuous? What is it at 16 Hz?
This will help to determine how much VD is required to obtain your required max SPL at your chosen lower freq limit. Most drivers work well beyond Xmax, 20% more is usually easy, some will do even more.
Just a note, you can set the the configuration to 2 drivers, and there is a check box to show the passband droop as a result of Le. Nice work!
Pete B.
Hi Glen, you may also need to add in the series resistance of all the wire in the circuit including the speaker wire from the amplifier and back.
Half an Ohm of resistance in that big an inductor will probably mean laminate core, and will need 1.6mm wire at least.
Ted
Half an Ohm of resistance in that big an inductor will probably mean laminate core, and will need 1.6mm wire at least.
Ted
Moondog55 said:
Hi Ted.
For the 250Hz sub woofer crossover inductor, I'm looking at the permite cored units from SpeakerBug:
"These cored induction coils use a Permite material which provides significant lower hysteresis compared to coils with ferrite cores. Power handling before saturation 400 watts. Cored inductors exhibit lower DCR when compared to air cored coils of the same wire gauge the benefits of which includes the preservation of amplifier damping as well as driver efficiency."
http://speakerbug.com.au/shop/index.php?main_page=index&cPath=4&sort=20a&page=2
I will use a seperate crossover filter for each 4-ohm woofer.
This means the inductor will be really loafing.
For a -12dB/oct crossover I will need a 3.6mH inductor.
Speaker bug sell a 3.9mH P-core with a DCR of only 0.19ohms - 21 times less than the nominal speaker impedance, so power loss isn't much of a problem. I'll use a pair of these with a few turns wound off for the desired inductance.
One thing I've learnt so far is that a little series resistance with the woofer is actually a good thing. It lowers Qtc, which means you have to compensate with a larger closed box volume to restore the damping factor - but that comes with a lower F3.
Two sets of extra long automotive jumper leads from Supercheap Auto will solve the problem WRT to speaker lead resistance (yes, I am being serious).
Cheers,
Glen
An externally hosted image should be here but it was not working when we last tested it.
PB2 said:
OK, I have the HF rolling off and a pair of drivers in parallel, but not sure how to work out the cone excursion. 😕
Cheers,
Glen
While using big speaker cables helps that would be considered overkill by most, some people regard the path back to the amplifier as more important and the speaker cables I use are made from 3core 15amp power flex cable 2.5square mm, I solder the earth and neutral together as the return and use the brown HOT wire as positive, works well for me.
If you can find 4mm 20Amp flex that would work well. I am assuming that the cables aren't going to be longer than 1.5/3 metres.
Those jumper leads have very thick PVC insulation but the conductor is (from memory ) only 2.5mm in dia (5mm sqare )
An alternative is to use multiple runs of 2.5mm*2 plaited around a nonconducting core.
It is sensible to use cable that can handle the peak current without getting too warm.
using seperate inductors on each woofer is a technique I've used myself to get the inline resistance lower without having to pay for custom winding
If you can find 4mm 20Amp flex that would work well. I am assuming that the cables aren't going to be longer than 1.5/3 metres.
Those jumper leads have very thick PVC insulation but the conductor is (from memory ) only 2.5mm in dia (5mm sqare )
An alternative is to use multiple runs of 2.5mm*2 plaited around a nonconducting core.
It is sensible to use cable that can handle the peak current without getting too warm.
using seperate inductors on each woofer is a technique I've used myself to get the inline resistance lower without having to pay for custom winding
G.Kleinschmidt said:
You're aware that the required inductance goes down if you put the drivers in parallel? And I'd rather see an air core if at all possible.
You might want to look for tests of those if you must use them.
G.Kleinschmidt said:
It's on a tab to the right, there are more sheets if that's what you're asking.
PB2 said:
Hi Pete.
Yes, but then a single inductor would have to handle the load current of two woofers.
PB2 said:
Ahh! (still learning to use Excel)
Cheers,
Glen
- Status
- Not open for further replies.