wiring parallel or series sensitivity rules?

There does appear to be a slight confusion of terms here, even my misunderstanding of it.
From what you have said above, a driver producing a certain acoustic output (acoustic energy) is then paralleled with an identical driver producing the same acoustic output the overall acoustic energy produced by the two drivers is double, that is +3db. But a microphone doesn't measure total acoustic power output from a driver (or drivers) but the sound pressure level SPL at a particular point of measurement, that is in half space for ... db@2.83V/1metre and that is taken as the sensitivity of the driver. Adding a second complementary driver in close proximity simply raises the overall sound pressure level by 3db midrange. Not 6 db as I mentioned but will be for lower bass. Sorry for the confusion.

WinISD is correct.

C.M
 
Last edited:
In my Near Field Line Array White Paper (NFLAWP) I deal with how to calculate the efficiency gain and sensitivity gain (or loss) for multiple drivers. See the discussion on pages 17 and 18 for how to relate these factors to the overall speaker SPL.

As noted in the above discussion the actual (measured) SPL is subject to how closely the drivers are spaced (coupled) to each other.

http://www.audioroundtable.com/misc/nflawp.pdf

As noted in the white paper how drivers radiate in their horizontal and vertical planes may differ (cone drivers versus say planar/ribbons).
So ultimately the measured results will be the truth test for the resultant SPL.
 
Last edited:
A few things to note:

As the driver becomes "non-directive" (and below) acoustic center for that driver is typically somewhere near the driver's surround (spanning most of the driver's surface). Some driver's though effect a wider range of dispersion with cone structure (to a point) usually with a particular dust-cap profile.

Practically speaking then, a driver is (for the purposes of acoustic center near the surround) non-directive about an octave below the wavelength for that driver's diameter. Ex. operational diameter for the driver is 100mm. Wavelength for 100mm's is about 3400 Hz. Half of that freq. is an octave lower = 1700 Hz. 1700 Hz then is about that point where acoustic center nears the surround, or the operational diameter of the driver.

Other things to keep in mind with respect to gain:

Diffraction with multiple sources/drivers usually adds a bit more gain to the midrange below the diffraction "peak" (..of course IN-ADDITION to the diffraction "peak").

Gain's from halving impedance rely on the partnering amplifier's output impedance AND the power supply of that amplifier and its ability to provide sufficient current into the lower output impedance with the ability to "double-down" power output into lower impedance loads. Some amplifiers with a high output impedance will actually provide losses into lower impedance loads and GAINS into higher output impedance loads.
 
Last edited:
Let's take a different approach: 2 drivers = double the cone size = double efficiency (which is acoustic/electric power). This normalized to power gives +3dB but normalized to voltage gives +6dB because power is proportional to voltage^2 . Connecting a second speaker to the original in parallel with leaving everything else as it was before, power doubles because I halved the impedance. At the end, double power + double efficiency = 4x acoustic power which is then +6dB lift.

True or false ? :)
True, when the two drivers can be considered as coupled.
 
Again thank you to all of you for this still valuable conversation.

To the last one @ ScottG: it might then also happen that at some lower frequencies the midranges built next to eachother are interfering with eachother producing another ~ +3dB lift while after a certain frequency (higher enough to narrow the dispersion) they get decoupled and a normal +3dB gain remains as stated by Tweet in #41. Do I see it right ?

From that point of view - no big news for you probably :) - I think it's extremely difficult to get a somewhat-flat response even in an echoless room, in a sterile environment, not to mention a normal living room where we're definitely screwed up here and there. If we move out 2-tweeters apart from eachother we cure somewhat the interference on lower midrange freqs as well, on the other hand we're getting a less-pointlike source of sound. Doing this horizontally is less advised due to horizontal dispersion "issues", doing this vertically might make sense (and we then probably beginning to discover the good old D'Appolito MTM design).

Considering thisarticle showing the freq response of a "Ferrari", I'm beginning to scratch my head. And how about the IRS V-s used here with the bass segment pushed quite much to the back (definitely needs a huge amount of time correction and still blows people away)..

There seems to be an insanely huge gap between theories what we tend to discuss hours long and then an Infinity reference model from 1985 comes and seems to f**k physics in doggy style.. :scratch1: :bigeyes: :spin: :hot: :D Oh God.. I'm better off with 1 stronger midrange than 2 ones.

Thank you all, again. (Shall this topic continue or rise again, I stay to learn further about both parties' opinions - as I see there's STILL a split). :judge:

Edit: AndrewT :worship:
 
Here is another factor, Treble always sounds louder than Mid and Bass. That is a tweeter at 90db sounds louder than a Midrange at 90db. It is the nature of audible treble frequencies.

So, trying to take rate Sensitivity of drivers, and combine them to equal a tweeter would typically not work. Though rather than adding more and more drivers, you could simply spend about $5 on Resistors and Pad the drivers down until they are balanced.

I'm not saying don't choose your speaker to create an overall level match, I'm just saying at some point you have to move from the theoretical to the practical. Generally, by their nature, a woofer do not have the Sensitivity of a Tweeter. There are some tweeter that are well over 100db output. It is going to take a LOT of woofers and midranges to match that output.

Matching the Impedance and Sensitivity of driver is one of the tricky parts of finding drivers for a particular project. You also can't trust the rated frequency response of most bass drivers, while they may be functionally loud to quite high frequencies, they are not flat much above 1000hz. Of course there are exceptions. Larger bass drivers are going to have far less upper bass and lower mid. Smaller bass driver will have less bass but better upper bass and lower mid. So, it matters whether we are talking 10" and 12" bass drivers or whether we are talking 5.25" and 6.5" bass drivers.

Also, driver material matter relative to workable frequency range and the necessity of resistive elements. For example, metal cone driver (usually aluminum) are very good within their working range, but above their working range they have a peak output caused by resonance that needs to be dealt with. You must either limit the Pass Band to a very low crossover frequency, or you must add additional crossover element to deal with the break up area.

That's why engineers usually design speakers instead of pizza boys.

Then there are phase issues that depend on the type of crossover you use. If you look at 3-way speakers that use -12db/octave crossovers, you will typically see that the Midrange is wired backwards. On rare occasion, you will also see tweeters wired backward. That is partly due to phase shift introduced by the -12db/octave crossovers.

It take a pretty high threshold of basic knowledge to create a decent DIY speaker system.

Steve/bluewizard
 
Last edited:
diyAudio Moderator
Joined 2008
Paid Member
when the two drivers can be considered as coupled.
This won't apply to a pressure measurement.. If the two drivers are within 1/4 wavelength such that comb filtering is avoided, the room power won't be reduced by cancellations. On the other hand you could have a 1" tweeter and a 15" woofer on opposite sides of the room and if you measure from somewhere that the pressure contributions are equal and in phase, you get +6dB.
 
Two coherent (in-phase) sources added together is +6dB, or technically +6.02dB because:
20*log(1) = 0
20*log(2) = 6.02

The equation 20*log(x) is a convertion from linear to logarithmic (since decibels is a logarithmic scale) where x is a linear quantity. In the above we increased the linear amplitude from 1 to 2 since we went from 1 speaker to 2 speakers.

So this means if you have a speaker connected to an amplifier producing 90dB SPL and you then decide to connect another idential speaker in parallel you will achieve 96dB SPL. However since adding the second speaker the amplifier is now producing twice the power output since the amplifier is a voltage source and will simply provide as much current as needed to maintain the target voltage. Therefore the current produced by the amplifier doubled and so did the power since power = voltage*current.

If we decide to now turn down the volume control so the amplifier only outputs as much power as it did into 1 speaker, the voltage (and therefore current) will be reduced to 0.707. Power = voltage*current = 0.707*0.707 = 0.5 (half the power)
Now we find what happened to the sound level at the speaker drivers. We reduced the current to 0.707 of the original value so therefore the linear amplitude will reduce by the same amount. Lets find out what that is in decibels:
20*log(0.707) = -3.01

So when we double the number of coherent noise sources we get +6dB but when we halve the power per noise source to get back to our original total power we get -3dB.
This is why when you tell WinISD there are twice as many drivers but maintain the same power output you only get +3dB and not +6dB.

For the sake of completeness, when you connect two identical drivers in series you get the same +6dB from having two coherent noise sources but now the voltage and current to each driver is halved (-6dB), net effect is 0dB compared to a single driver and the power to each driver is 1/4 of the original power. If you increase the power back to the original power by turning up the volume control you get +3dB.
 
Last edited:
..
Steve/bluewizard
I think you might know what my original plan was: 1x 18" woofer (97dB), 4x midrange (91dB -> lifts to 97) and 1x AMT tweeter (98dB). Tri-amping active setup, LR4, XO: ~ 125-150Hz and ~2500Hz. Now I'm thinking on 1x stronger midrange around 97dB (and flat) instead of 4x91dB smaller ones if the debate still goes on how much I get from 91dB: 94 or 97. Too much of worries all you could say but I still want to plan good, in theory at least. The standalone midrange seems to be a good choice too I think. In an active setup I can correct sensitivity (efficiency?) :) differences via gain adjustments but for a while I might also try this setup as a passive one with lower volume levels and this is why I'm juggling with numbers in my head, what to do and how to select midrange. 1 piece (with a good power handling capability, good sensitivity and more importantly, "flat" response) or 4 slightly better ones, introducing this trouble of more speakers while wanting to achieve an even better freq curve. I think I add more complexity and mess with 4 midranges than by choosing 1x standalone midrange. Anyway, this is my next build on paper, just made the sketch for the woodworker for the bass box. But the theory is just as important for me, behind this 'cause you know.. knowing is good and we all learn a lot in these forums. I really enjoy the journey with you all.. *respect*
 
This won't apply to a pressure measurement.. If the two drivers are within 1/4 wavelength such that comb filtering is avoided, the room power won't be reduced by cancellations. On the other hand you could have a 1" tweeter and a 15" woofer on opposite sides of the room and if you measure from somewhere that the pressure contributions are equal and in phase, you get +6dB.
That's the constructive interference looming again.
 
I think you might know what my original plan was: 1x 18" woofer (97dB), 4x midrange (91dB -> lifts to 97) and 1x AMT tweeter (98dB). Tri-amping active setup, LR4, XO: ~ 125-150Hz and ~2500Hz. Now I'm thinking on 1x stronger midrange around 97dB (and flat) instead of 4x91dB smaller ones if the debate still goes on how much I get from 91dB: 94 or 97. Too much of worries all you could say but I still want to plan good, in theory at least. ...

Here we encounter the complexities of Speaker Design.

Good Woofer, exceptionally wide frequency response -

Dayton PA460-8 - 18" -

https://www.parts-express.com/pedocs/specs/295-036-dayton-audio-pa460-8-specifications-46885.pdf?

https://www.parts-express.com/dayton-audio-pa460-8-18-pro-woofer--295-036

Don't know much about the midrange, but it looks pretty good -

Faital Pro 6FE100 - 6" Midrange Driver -

http://www.faitalpro.com/en/products/LF_Loudspeakers/product_details/datasheet.php?id=401020100

Aurum Cantus Aero Striction Tweeter - either Ribbon or Air Motion (not sure) -

https://www.parts-express.com/pedocs/specs/276-446--ast25120-spec-sheet.pdf

https://www.parts-express.com/aurum-cantus-ast25120-aero-striction-tweeter-8-ohm--276-446

The Arum Cantus are good, but they are not cheap. Amazingly wide frequency response on this tweeter.

But notice you are spending $300 on the Tweeter and $100 on the Woofer. Again, not well balanced, but your choice.

It would take slight exception to the crossovers you have chosen, but that's a matter of personal opinion. Give the very wide response of the Woofer, I would be inclined to think 250hz and 2000hz to 2500hz. 150hz takes the Midrange down to the bottom of its workable range. The more inside the workable range, the better the Mid is going to perform as is true of just about everything in life. Pushed to the limited, it is going to be strained, well inside the limits, substantially less strained.

Also, especially on the low end, the closer you are to the low end limit, the less power handling you have. Though in your case with 4 Midrange, that's probably not a concern. An array of 4 would give you double the power handling, half the excursion, and a +3db boost in the output.

Best off the top of my head guess, a 2x2 array of 6" Midrange would need about a 21" wide cabinet. A 21" wide cabinet would probably work well with an 18" woofer. However, I'm not sure how 21" wide would work with the Tweeter which itself is 84mm (3.35") wide.

A 21" baffle width corresponds to the frequency of 643hz (unless I calculated wrong). So a 250hz Bass/Mid Crossover is below the Baffle Step.

If you look at the Part Express page for this woofer, they have determined using software called "Bass Box Pro" that the optimum Sealed Cabinet is 2.64 cubic feet, and the optimum Ported Cabinet is 5.68 cubic feet. You might want to work out your cabinet dimensions for boxes that size.

Cabinet Size also effect frequency response. The 2.64 Ft³ cabinet is rated at 73hz, and the 5.68 Ft³ cabinet is rated at 47hz at -3db. Perhaps not what you would expect from a Driver that is rated down to 26hz.

Generally PA Woofers do not go as low as Hi-Fi Subwoofers. The disadvantage to a Subwoofer is that the impedance is relatively low, and that implies you have to drive it with a separate amp. Though you were planning to do that anyway.

For example,

https://www.parts-express.com/cat/s...eca:matches(.,"P_Searchable","1")]&PortalID=1

Note the highest impedance is 4 ohms -

https://www.parts-express.com/dayton-audio-dcs450-4-18-classic-subwoofer-4-ohm--295-475

https://www.parts-express.com/pedocs/specs/295-475-dayton-audio-dcs450-4-specifications-46179.pdf


https://www.parts-express.com/dayton-audio-rss460ho-4-18-reference-ho-subwoofer-4-ohm--295-472

https://www.parts-express.com/pedocs/specs/295-472-dayton-audio-rss460ho-4-specifications.pdf

The $125 Dayton Classic 18" Subwoofer has these specs -

Optimum Cabinet Size (determined using BassBox 6 Pro High Fidelity suggestion)

Sealed Volume = 3.85 ft.³
Sealed F3 = 46 Hz

Vented Volume= 13.01 ft.³
Vented F3 = 25 Hz


That's better, but larger cabinets. Work out those dimensions assuming a 21" face on the cabinet.

1.75 ft (21")W x 2.5 ft H x 3ft D = 13.125 Ft³

Keeping in mind you have to subtract the volume of the drivers from the internal volume of the cabinet. So, let's throw in a bit more for the volume of the driver.

1.75 ft (21")W x 2.75 ft H x 3ft D = 14.44 Ft³

Is that a workable cabinet for you?

The Midrange would have the same face, roughly 21" x 21" but probably would not need to be as deep.

Also note that using the best guess estimate, the speakers will be about 21" wide and 60" high and best guess a bit over 3ft deep, and given the size, they are going to need to be set out from the wall a considerable distance. I would speculate about 2ft minimum. Can you handle that?

So, consider the size of room it is going to take to reasonably house these speakers. I'm guessing 20ft x 30ft room would work. Though the exact dimensions are up for debate.

To get good dispersion and sound stage, you probably don't want the tweeter to be on a full 21" wide baffle, though you are free to do that. But if you look at many high end speakers, the cabinets are trimmed away in the area of the Tweeter.

https://www.canton.de/en/reference-k/reference-1-k

In the case of Bowers-Wilkins, they use an Outboard Tweeter to very impressive effect -

http://www.bowers-wilkins.com/Speakers/Home_Audio/700-Series/702-S2.html

http://www.bowers-wilkins.com/Speakers/Home_Audio/800_Series_Diamond/800-D3-products.html

But that is a design decision that is up to you.

Lastly, if a full Active Speaker system, what Crossovers were you planning to use?

PA Crossover are cheap enough, but they tend to be lock at either 18dB/Octave or 24db/Octave which are very steep slopes. Fine for PA, but a bit much for Hi-Fi.

If you can spend more money, then get a Speaker Management System, which is also a Crossover made for Studios and for PAs, but it has adjustable Slopes on the Crossovers. However, these convert everything to the Digital Domain, apply the processing, then convert the various frequency ranges back to analog on separate outputs.


Also, some PA Amps have built in Crossovers, and can be bridge into a mono amp with very considerable power. You could then go for Passive Crossovers on the Mid/High. But again that is a design decision you would have to make.

Here is one in a series of Crown Amps that has a built in crossover (low pass, band pass, high pass), and substantial power. The person in this video is planning to use the amp on a Dayton Subwoofer.

https://www.youtube.com/watch?v=BZD2Pc_tHnE

There are more powerful and less powerful Amp in the Crown XLS series.

https://www.sweetwater.com/store/search.php?s=Crown+XLS+amp&Go=Search

https://www.zzounds.com/prodsearch?q=Crown+XLS+amp&pa=29&key=q&form=search

This is the Amp in the Video -

https://www.sweetwater.com/store/detail/XLS1502--crown-xls-1502-power-amplifier

https://www.zzounds.com/item--CWNXLS1502

- Power Rating:
- Stereo 2-ohms: 775W
- Stereo 4-ohms: 525W
- Stereo 8-ohms: 300W/channel
- Bridged 4-ohms: 1550W
- Bridged 8-ohms: 1050W

The process is complex with many details that need to be considered.

Steve/bluewizard
 
Last edited:
That's better, but larger cabinets. Work out those dimensions assuming a 21" face on the cabinet.

1.75 ft (21")W x 2.5 ft H x 3ft D = 13.125 Ft³ (21"W x 30"H x 36"D)

Keeping in mind you have to subtract the volume of the drivers from the internal volume of the cabinet. So, let's throw in a bit more for the volume of the driver.

1.75 ft (21")W x 2.75 ft H x 3ft D = 14.44 Ft³ (21"H x 33"W x 36"D)

Is that a workable cabinet for you?

Add the dimensions in inches.

Steve/bluewizard
 
Vortex, It might be easier to build an open baffle design, you might like to read this article first. The articles are very informative and very well written.

http://www.hifizine.com/2010/12/prototyping-4-way-open-baffle-speaker-with-the-minidsp-2x4/
http://www.hifizine.com/2011/03/refining-a-4-way-open-baffle-speaker-minidsp-2x4/

Prototyping a Dipole Bass System

Minidsp makes a good workable crossover design much easier than trying to configure a 'more or less hit and miss' passive design and to get it right. Just a suggestion.

C.M
 
Last edited:
Thank you all for the valuable input, what you say sounds interesting. Room is average, rather mid-size, but I'll move soon to a new one with plenty space for everything. Speaker size and placement is not an issue. I have a separate thread for my project. I calculated a huge vented box for the original Daytons, port tuned at fs, with -3dB at 24Hz, having taken everyting else into consideration (acoustic suspension/stiffness, delays, little SPL decrease, etc). An experiment, I have a good axe in case I hate the sound, don't worry. Mdf burns good. ;) For the latest sketch (idea), check #11 here, I mean the drawing now. I'll introduce a little trapezoid effect with the big side panels so they won't be parallel, less standing waves.

Back to topic: 2 speakers in series halves excursion (we can say halves power falling onto 1 speaker if they're identical). This website's calculator tells thr same for parallel connected speakers too which in my opinion is not the case. In parallel connection I think both individual speakers retain their power and excursion. Can be that these calculators and figures are wrong here ?
 
Last edited: