Parallel driver interaction in a 4-way system.

Status
Not open for further replies.
Jump to Latest
#201
Guys, thanks again for all the help so far and I apologise now for entirely changing the direction of this thread, but I’ve just sound something that changes everything.

All my problems seem to be solved. What do you guys make of this (say, in a nice WMT 3-way with the coils in parallel):

http://www.soundlabsgroup.com.au/p/M-103260-SPH380TC/15"+SPH-380TC+2+x+4+Ohm

A good reason to forget about 8" woofer line arrays? 🙂

Cheers,
Glen
 
#202
One Shiva-X with 2.6L VD has more than twice the VD of that driver. I suggest 2 of these per side so that your system will easily do 16 Hz, for better thermal capacity having two large voice coils, and to raise the efficiency of the single driver.
Look at the Bl curves when selecting a driver, the XBL motors
are usually far more linear.

Pete B.
 
#204
PB2 said:
One Shiva-X with 2.6L VD has more than twice the VD of that driver. I suggest 2 of these per side so that your system will easily do 16 Hz, for better thermal capacity having two large voice coils, and to raise the efficiency of the single driver.
Look at the Bl curves when selecting a driver, the XBL motors
are usually far more linear.

Pete B.
Click to expand...


OK, thanks Pete, I'll check it out.

Cheers,
Glen
 
#206
G.Kleinschmidt said:
Hmmmmm....

Now that monster will require a cabinet 0.5m wide. Baffle step will therefore be at ~226Hz.

That would be a nice place to crossover, so the BSC can all be done in the bass.

It has a sensitivity of 89dB (1W/1m). Minus baffle step gain, how about that 84dB Visaton 8" "sub woofer" for the mid-range?

http://www.soundlabsgroup.com.au/p/V-1340-TIW200XS/TIW+200+XS+-+8+Ohm


Cheers,
Glen
Click to expand...

You need to consider voltage sensitivity rather than the per watt per meter figure since your driving it with a voltage source, keeping in mind the min impedance which of course will not be a problem with your amp. My feeling is that that mid is too inefficient.
 
#207
G.Kleinschmidt said:

That would be a nice place to crossover, so the BSC can all be done in the bass.

Cheers,
Glen
Click to expand...

Am I reading you correctly that your thinking of steping the sensitivity levels at the XO point to implement BSC? This would be a compromise it really should be gradual from about 100 Hz to 1 kHz, roughly.
Actual, will depend on placement of the drivers and room boundary destructive interference.

The room boundary issue was covered years ago in designs and papers by Allison, the AR-9, and more recently in the NHT 3.3. Worth reading about even if you don't use their methods.

Pete B.
 
#208
PB2 said:
I mentioned the NHT 3.3, here is a revised version of the 12" woofer used in that design. It was very good in its day, however the motor is not as good as the latest XBL types:
http://www.madisound.com/catalog/product_info.php?manufacturers_id=139&products_id=1402

Pete B.
Click to expand...


Ok, I've just downloaded the Shiva-X *.pdf
The dual voice coils are unfortunately only 8 ohms each, so my amp will never be able to drive them to capacity.

I could use two of them per side to get the impdeance down, but it seems a waste. On the plus side, I guess that linearity would be fantastic as each unit would never be pushed to its limit.

My feeling ATM is that that unit with the dual 4 ohm coils would be a better compromise all round (would certainly get my living room shaking).


PB2 said:
You need to consider voltage sensitivity rather than the per watt per meter figure since your driving it with a voltage source, keeping in mind the min impedance which of course will not be a problem with your amp. My feeling is that that mid is too inefficient.
Click to expand...


D'oh - good point.


PB2 said:


Am I reading you correctly that your thinking of steping the sensitivity levels at the XO point to implement BSC? This would be a compromise it really should be gradual from about 100 Hz to 1 kHz, roughly.
Actual, will depend on placement of the drivers and room boundary destructive interference.

The room boundary issue was covered years ago in designs and papers by Allison, the AR-9, and more recently in the NHT 3.3. Worth reading about even if you don't use their methods.

Pete B.
Click to expand...


Pete, I don't have enough information on BSC to know what the best alternative is.

I'm still working through the huge number of links that have been provided in this thread.

Cheers,
Glen
 
#209
A quick BSC overview:
Let's say for example that your woofer XO is 500 Hz, about half way through the full BSC range. BSC is typically implemented simply by making the woofer XO inductor significantly larger than the calculated value. You might think that this would cause a too rapid 6 dB/oct slope, however once you take into consideration the rising impedance of the woofer VC inductance it is usually much more gradual as is required. This combined with an RC across the woofer that is not an exact Zobel can be used to control the slope, and the actual crossover point.

Sometimes, it is necessary to split the woofer XO inductor into two, one with a parallel resistor for gradual BSC, and the other to roughly provide the actual XO.

So you'd do roughly half of your target BSC step up to 500 Hz in the woofer XO, and then use the same method in the lower mid XO continuing to 1 kHz.

This is very commonly done in commercial designs.

You can SPICE it using an RC for the driver but it is much better to use an actual speaker simulation package.

Pete B.
 
#212
PB2 said:
A quick BSC overview:
Let's say for example that your woofer XO is 500 Hz, about half way through the full BSC range. BSC is typically implemented simply by making the woofer XO inductor significantly larger than the calculated value. You might think that this would cause a too rapid 6 dB/oct slope, however once you take into consideration the rising impedance of the woofer VC inductance it is usually much more gradual as is required. This combined with an RC across the woofer that is not an exact Zobel can be used to control the slope, and the actual crossover point.

Sometimes, it is necessary to split the woofer XO inductor into two, one with a parallel resistor for gradual BSC, and the other to roughly provide the actual XO.

So you'd do roughly half of your target BSC step up to 500 Hz in the woofer XO, and then use the same method in the lower mid XO continuing to 1 kHz.

This is very commonly done in commercial designs.

You can SPICE it using an RC for the driver but it is much better to use an actual speaker simulation package.

Pete B.
Click to expand...


Thanks Pete.

This makes sense. A couple of Q's though:
What order is the baffle step gain - 3dB/octave, 6dB/octave?
I've been told that BS is at ~1/3 the baffle width wavelength and would be all over by 1kHz for a 0.3m wide enclosure. A 15" sub would require a 0.5m wide enclosure.

That Monacor dual 4-ohm coil unit I found has a recommended crossover frequency of 400Hz.
If I built a WMT 3-way with one of these crossed over at 400Hz in a 0.5m wide enclosure, would it be feasible to do the baffle step correction entirely in the bass X-over?

Also, is the full +6dB of baffle step gain typically realised?
The reason I ask is because I've been looking through speaker projects published in various electronics magazines and the effect seem to be universally ignored to the point of not even being mentioned.

If the effect in real life is significantly lower than +6dB (say 2 or 3 dB) and heavily dependent on room acoustics and speaker position, could this be just something the average designer leaves aside for active equalisation (if desired)?

Cheers,
Glen
 
#213
Moondog55 said:
Also look at the "Maelstrom" 15 inch subwoofer, but again if you start talking 8inch mid bass we are back into 4-ways, or the bottom woofer could be your .5.
I've also looked at the Vifa M22-WR -09-08
Avaliable from Speaker bits or WES if you have a trade account

http://www.speakerbits.com/Default.aspx?menupage=150&pi=1

Rabbitz likes this bass driver, and I have moddelled it in a small sealed box,
Click to expand...


Hi Ted. Do you mean this?

http://www.audiojunkies.com/blog/1380/maelstrom-x-has-landed

That's a serious speaker, but 18" would take a wider enclosure than I'd like and its dual coils are only 8 ohms each.

BTW, what software package are you using?

Cheers,
Glen
 
#214
Half space sensitivity

Does the half space sensitivity actually take into account the baffle step gain?

For example, would an 8-ohm 89dB (1W/1m) woofer need to be paired with an 8-ohm 89dB (1W/1m) tweeter, or a ~83dB tweeter?

Also, how on earth do I translate these "1W/1m" sensitivity figures between speakers of different impedances that I intend to connect in parallel?


😕

Cheers,
Glen
 
#215
Ebp+ Xo Dcr

OK, more babble


Besides downloading WinISD, I haven’t started modelling any sealed woofer enclosures yet, but I have a question WRT to selecting woofers suitable for sealed box operation.

The Cookbook states “In terms of driver Qts, closed-box loudspeakers generally require a fairly high Qts of greater than 0.3.” This means drivers with moderately sized magnet structures.

R. Small’s “Efficiency Bandwidth Product” is then cited as the criterion for determining if a woofer is more suitable for a closed or vented box.

EBP = Resonate Frequency / Driver Electrical Q
= fs/Qes.

An EPB in the vicinity of 50 or less means a sealed box is more appropriate. An EBP of about 100 really indicates a driver best suited to a vented enclosure.

Very few drivers that I’ve looked at , even those recommended for closed box operation, have an EBP less than 50. Most are around 70 or so.

However, one driver I’ve found has and extremely low Qes of 0.2:

http://www.soundlabsgroup.com.au/p/M-101830-SPH300CTC/12"+SPH-300CTC+2+x+8+Ohm

With a resonant frequency of 18Hz, its EBP = 90. So, this woofer, according to R. Small, needs a vented enclosure.

What I find interesting is the “only with series resistor” note under the closed box specification. In other words, a series resistance is required to reduce Qes/Qts, thus bringing the EBP down.

So, the question – since most drivers (even those marketed for closed box operation) do not have an EBP as low as the stringent standard of 50 or less, wouldn’t the Qts reducing DCR of the average crossover inductor actually be an advantage?

Cheers,
Glen
 
#216
I don't know all the answers to those questions; but if you read back a bit, all series resistance including driver DCR is parasitic and reduces SPL.
As I understand it low bass plus loud bass is a trade-off between efficiency and box size; if you could tolerate a huge box it becomes less of a problem, as a high Qts driver needs a huge box to give a reasonably flat response. Infinite baffles ( see the huge threads on this subject) with sealed boxes ie: air suspension the plus is a small box.
It is usually recommended to use the lowest DCR inductor you can afford to avoid losing SPL, put some figures into a box program and see how HARD it is to achieve concert levels of 120dB even momentarily.I am using Jeff Bagby's free downloads

avaliable here


http://audio.claub.net/software/jbagby.html

The Vifa woofer I pointed to has a Qts of 0.33 works well in a small sealed box but its almost twice as loud in a ported box.

And deeper but needs BIGGER box,
 
#217
Cheap 12inch driver for sealed box

http://www.soundlabsgroup.com.au/p/V-9069-W300/W+300+-+8+Ohm
or the equivalent Peerless here #830669

http://www.speakerbits.com/Default.aspx?menupage=150&pi=3

both need big boxes to acheive a flat in box response, otherwise there is a big peak in the response, think of cheap boombox type speakers, really heavy on kick-drum but cannot reproduce organ music or tympani tones or real double bass

http://www.audiomarketplace.com.au/...product_id,2/option,com_virtuemart/Itemid,49/

The Tempest is here also , my mistake the Maelstrom is BIG actually 18inches across, the Tempest is the 15incher
 
#218
PB2 said:


You need to consider voltage sensitivity rather than the per watt per meter figure since your driving it with a voltage source, keeping in mind the min impedance which of course will not be a problem with your amp. My feeling is that that mid is too inefficient.
Click to expand...

That is not actually what is represented, it is showing the sensitivity. It's unfortunate that even some manufacturers use the 1W/1m as the description. The graph and sensitivity rating is for voltage sensitivity. The 1W/1m refers to the voltage applied that would result in 1W/1m if the load were a constant 8 ohms, that is, 2.83v. The result is the standard voltage sensitivity rating/curve for 2.83v @1m.

The driver is not too inefficient, it doesn't have a high enough sensitivity. References to the term efficiency are incorrect. Efficiency is a specific reference to the power out/power in, that is, acoustic power out/electrical power in and will be shown as a percentage. The graph/sensitivity number provided can be used as shown directly as being 2-pi.

Dave
 
#219
Moondog55 said:
Cheap 12inch driver for sealed box

http://www.soundlabsgroup.com.au/p/V-9069-W300/W+300+-+8+Ohm
or the equivalent Peerless here #830669

http://www.speakerbits.com/Default.aspx?menupage=150&pi=3

both need big boxes to acheive a flat in box response, otherwise there is a big peak in the response, think of cheap boombox type speakers, really heavy on kick-drum but cannot reproduce organ music or tympani tones or real double bass

http://www.audiomarketplace.com.au/...product_id,2/option,com_virtuemart/Itemid,49/

The Tempest is here also , my mistake the Maelstrom is BIG actually 18inches across, the Tempest is the 15incher
Click to expand...


Hi Ted

The problem with some of these big subs (especially the Shiva X) is that they don't have a wide enough bandwidth.

I currently quite like the look of the Monacor dual 4-ohm coil unit:

http://www.soundlabsgroup.com.au/p/M-103260-SPH380TC/15"+SPH-380TC+2+x+4+Ohm

This single 15" unit will give me my 2-ohms nominal impedance, it can take all the power my amp can give it and it has seemingly excellent bandwidth (fu-800Hz) with a recommended crossover of 400Hz.

An externally hosted image should be here but it was not working when we last tested it.


It has a sensitivity of 89dB (1W/1m). The EBP is 69, but the manufacturer claims operation in small cabinets and recommends only 55L for a sealed enclosure.

I've downloaded those programs from Jeff Bagby's site (thanks!).
I'll see how this unit performs in the simulator.

Now I'm currently at a loss to work out the 1m/1W SPL figure I need from an 8 ohm midrange driver to go with it.

Cheers,
Glen
 
#220
Glen are you beginning to see why so many of us are using Electronic crossovers??
I would never attempt to build some of my stuff with passive networks, it is simply too easy to use active systems,
I admire your drive.
I have a liking for 12inch woofers, but I do have a sub I built a while ago using a car driver, Kicker ES or EX can't remember, Fs 18Hz, Qts of 0.42 and it still took three boxes to get it to sound right, last box is 140 litres fully stuffed and a passive first order high pass at 20Hz that took $114 dollars worth of capacitors
1440uF/100V
As for band width, its only appliccable if you want to cross high, in a small room ( say 5 X 8 metres ) I'd never go over 250 with a 15 and probably prefer to stay under 100, crossing to 10's or 8's and crossing to the midrange around 250/300 tops.
Bass is the engine room, the fundamentals on which the music is built and to be good has to be stong, solid and clear ( so called fast bass ) but clarity in the midrange is what I try and build for
 
Status
Not open for further replies.
Top Bottom