Subwoofer driver optimized for sealed cabinet vs. for reflex cabinet

[ppataki]

This is more of a theoretical topic from my end (but it might very soon turn into a practical one, depending on how this discussion will evolve)

I have done many DIY projects in the last five years, many of those were subwoofer projects (the last several ones were DOS types - I love those)
For all my projects I had been using sealed cabinets but for the vast majority of those projects I had put drivers in those sealed cabinets that were actually designed/optimized for reflex cabinets (i.e. EBP value higher than 100 and very low Qts value)

I did not really care about that to be honest, simulations always looked fine; in reality the response did need some correction with EQ in the very low end (see below graph as one example) but distortion has always been low and I have always been happy with the sound subjectively too

Lately I have been pondering if it would make any sense to try a different approach and use a driver in a sealed cabinet that was actually designed to be used in a sealed cabinet

I went ahead and simulated the response of my current 2x15" Lavoce SSF153.00 based DIY subwoofer vs. putting 1x15" Dayton Audio RSS390HF-4 in the same cca. 74 liter sealed cabinet

The normalized response looks like this (orange - Dayton, red - Lavoce):

The Dayton has almost 14dB higher SPL at 20Hz

When I check the custom amplitude response with a load of 250W, the difference decreases to 2.5dB

Of course cone excursion is much higher in case of the Dayton (both relatively and in absolute terms)

And GD looks worse too

Phase however, looks better (I assume less rotation means better but please keep me honest here)

Of course the low-shelf correction (boost) that is needed for the Lavoce speakers are not taken into consideration here....
A 6dB low shelf filter would double the cone displacement and increase GD by approx. 2.4ms - so the above graphs would get much closer to each other

The theoretical question is: would an experiment of putting the above Dayton driver in the same cabinet make any sense?
Or would the room 'kill' all the difference anyway?

I would appreciate any comments, especially if anybody here has any real-life experience with a similar topic
Thank you

 

[bmc0]

GD looks worse [...] Phase however, looks better

If one is equalized using minimum phase filters to have the same anechoic frequency response as the other, the group delay and phase will also be the same (for small signals). In other words, the anechoic phase and group delay are linked to the frequency response so there's little point in comparing them if you're going to use EQ.

AM distortion caused by nonlinear Cms(x) and Bl(x) would be a bit worse when extending the LF response with EQ, all else equal. Since all else isn't equal, you'd need good large signal data of both drivers to predict which option might perform better overall. For a given SPL at a given frequency, the excursion of the single Dayton driver will be about double (since the total cone area is approx. halved), but it's also rated for twice the Xmax.

I personally have two RSS390HF-4s in sealed boxes and they perform well, but I've not compared them to those LaVoce drivers (or any other pro audio subwoofer, for that matter).

 

[mcdn]

The published specs for xMax are 6.7mm for the Lavoce and 14mm for the Dayton, so in principle one Dayton can give the exact same output as two Lavoce drivers in a sealed cabinet, as you have shown above.

But... That requires some justification. I asked a lot of stupid questions about sub drivers in a thread a while back and learned a _lot_. I'll try to summarise and see if I get it right this time.

1. xMax as a T/S parameter is intended to be a measure of "Linear xMax", the maximum excursion while keeping distortion below a certain level. This is not spelled out well for early learners like me, though it is mentioned in the Wikipedia article on T/S parameters.
2. Unfortunately there isn't one single definition of that "certain level" for subs. Car people often use 20% THD, 10% THD is more commonly accepted for other drivers, I'm sure there are 5 other standards. Sometimes it's not measured, just eyeballed based on voicecoil length.
3. Therefore published xMax is important but not determinative when choosing sub drivers unless you either make some assumptions or can find comparable independent measurements
4. Power handling is the other primary factor for sub driver performance - if the driver can handle the required power to get to xMax in a given cabinet, that's about all we need to know. (OK, there is potentially the nonlinearity of air in very small cabinets, but it will mostly be dominated by driver distortion)

So let's make some assumptions:

1. The Dayton and Lavoce linear xMax published specs are comparable enough for your purposes
2. The Dayton and Lavoce RMS power specs are also comparable enough (400W Lavoce, 500W Dayton)
3. We are using EQ of some kind

Then the question is "what difference does the cabinet make?". To simplify a bit, and talking about sealed cabinets for now:

1. The short answer is, you can make the box smaller but you'll need more power for the same response.
2. A cabinet that is "too small" will roll off the driver's output sooner with a higher resonant peak
3. A cabinet that is "too large" won't give any support to the driver through resonance, but otherwise has no ill effects
4. With EQ we can tailor the response, so we can use smaller boxes at the expense of requiring more power. See Linkwitz Transform at sound-au.com for a great explanation

Then what about ported VS sealed?

1. A port reduces driver excursion over its passband, and also driver power requirements. Something has to move the air in the port but it's resonating so is quite easy to drive. This is why designers love them
2. Port sizes become enormous unless the box volume is big enough
3. As a resonator, ports inevitably have more GD
4. Below port resonance the drivers are just exchanging air between the box interior and the room without making a pressure wave (the steep rolloff you see in measurements). This is why sealed designs may be better for true sub-bass e.g. 10Hz, because in principle they can be smaller than a 10Hz ported design

 

[bmc0]

for example in an infinitely large cabinet, or infinite baffle, the driver will have a big response peak response at Fs

You've got that backwards. Smaller volume means higher Q (>0.7 results in a peak before rolloff) and larger means lower Q.

Assuming the same driver, a smaller box will require less power within its bandwidth, but its bandwidth will be narrower compared to a larger box.

 

[LSPhil]

The theoretical question is: would an experiment of putting the above Dayton driver in the same cabinet make any sense?
Or would the room 'kill' all the difference anyway?

Yes, you will be happy. I have RSS315HF4, which play up to 500Hz.

 

[mcdn]

You've got that backwards. Smaller volume means higher Q (>0.7 results in a peak before rolloff) and larger means lower Q.

Assuming the same driver, a smaller box will require less power within its bandwidth, but its bandwidth will be narrower compared to a larger box.

That's a bad error, edited!

 

[Holmz]

That's a bad error, edited!

Can you just rewrite it correctly please?

 

[mcdn]

Can you just rewrite it correctly please?

That’s what I thought I’d done with the edit? Or do you mean the reference to power? In which case I should move that after the EQ point and make clear I mean that the power requirement after EQ might get beyond the driver’s capability, I.e the EQ isn’t a magic wand.

 

[mcdn]

Can you just rewrite it correctly please?

I've had another go at that section.

 

[Line Array]

get yourself HornResp and use the built in "efficiency" tool and you will see the efficiency of the Dayton is LOWER than that of LaVoce AT ALL FREQUENCIES including 20 hz, 16hz, 12hz etc ...

yes it has higher SENSITIVITY but lower EFFICIEINCY

sensitivity is ratio of output to VOLTS

efficiency is ratio of output to WATTS

then research PowerSoft iPal technology and PowerSoft M-Force technology

the bottom line is more motor force is always better - you just need to have the right electronics and DSP to manage it

the average DIYer is not sophisticated enough to manage drivers with too high Bl^2/Re or too low MMS etc.

what the average DIYer uses is RULES OF THUMB which they mistake for physics.

PowerSoft uses physics which is why nobody uses iPal or M-Force because nobody understands physics.

the "build it and they will come" approach doesn't work.

look at Tesla - people buy and sell Teslas based on anything but the merits ( or lack thereof ) of the actual car. they bought Teslas on hype and are now selling them on negative publicity. the car itself never mattered.

this is also how people buy speakers. they quite frankly don't have what it takes to objectively analyze what they're buying.

as for Dayton there are only two reasons to use Dayton:

1 - price
2 - availability

there is not a single Dayton driver for which there doesn't exist a better alternative from another brand, but typically it will cost double.

is it worth paying double to make sure you're not using Dayton ?

for me it is. but this is subjective of course.

is a Porsche worth 2X Tesla ?

opinions are split ...

personally, i will use Dayton when a company like L-Acoustics decides to use their drivers in the sound system used to cover a festival like Coachella ... in the mean time i will stick to transducer brands that actually do get used in such applications such as B&C etc.

and yes, Dayton now has "prosound" drivers, so they no longer have any excuses. if i pull apart an L-Acoustics box and find a Dayton driver in there i will convert.

 

[LSPhil]

there is not a single Dayton driver for which there doesn't exist a better alternative from another brand, but typically it will cost double.

A very interesting post. Could you please suggest a better alternative to the Dayton Audio RSS315HF4?

 

[Heinrich]

I went ahead and simulated the response of my current 2x15" Lavoce SSF153.00 based DIY subwoofer vs. putting 1x15" Dayton Audio RSS390HF-4 in the same cca. 74 liter sealed cabinet

The normalized response looks like this (orange - Dayton, red - Lavoce):

View attachment 436746

A word of caution: you compare the "transfer function / magnitude", which is normalized to "0", as you see in the graph. Once you consider that the sound per watt equation depends nearly solely on the internal volume (reflex +6dB), you would expect the "made-for-reflex"-driver to have the same absolute (!) level below ~100Hz, but way more above that.

There's no good reason to exclude even strong motorized speakers from use in a sealed enclosure. See example below, the 18sound has an exeptional strong motor, the Ciare is quite weak even by HiFi standards, same enclosure of 22 liters.

Transfer function: bad bass for the strong model

actual SPL at one watt: even in subbass the 18Sound is stronger by around 3dB, in the mids its more than 15dB (!!) stronger

 

[ppataki]

A word of caution: you compare the "transfer function / magnitude", which is normalized to "0", as you see in the graph. Once you consider that the sound per watt equation depends nearly solely on the internal volume (reflex +6dB), you would expect the "made-for-reflex"-driver to have the same absolute (!) level below ~100Hz, but way more above that.

There's no good reason to exclude even strong motorized speakers from use in a sealed enclosure. See example below, the 18sound has an exeptional strong motor, the Ciare is quite weak even by HiFi standards, same enclosure of 22 liters.

Transfer function: bad bass for the strong model

actual SPL at one watt: even in subbass the 18Sound is stronger by around 3dB, in the mids its more than 15dB (!!) stronger

View attachment 441311

Yes, you are right, and I know this, this is exactly what is shown in the first post in the second graph

After reading all the comments (thank you to all of you!) I am not sure if trying the Dayton would make any sense

Just one last thing that popped to my mind: would it make any difference that the Dayton has a way lower Fs vs. the Lavoce (19.5Hz vs 47Hz)?
Meaning that I am currently 'forcing' my Lavoce drivers to operate way below their Fs

 

[Heinrich]

Yes, you are right, and I know this, this is exactly what is shown in the first post in the second graph

After reading all the comments (thank you to all of you!) I am not sure if trying the Dayton would make any sense

Just one last thing that popped to my mind: would it make any difference that the Dayton has a way lower Fs vs. the Lavoce (19.5Hz vs 47Hz)?
Meaning that I am currently 'forcing' my Lavoce drivers to operate way below their Fs

Huh, again a tl;dr 8-(

Can't for sure answer the second question. But as long as the term 'force into' is used, I don't see a technical reason to decide for one or the other.

 

[DVDdoug]

Just one last thing that popped to my mind: would it make any difference that the Dayton has a way lower Fs vs. the Lavoce (19.5Hz vs 47Hz)?

I'm NOT an expert. I just plug the numbers into the software.... You could do the same thing.... Leave the other parameters unchanged and plug-in a different Fs.

But Fs is a free-air characteristic and I don't THINK it's as important in a "small" sealed cabinet where the resonance/cut-off is higher anyway. You can "force" the woofer to move at low frequencies with EQ and wattage.

 

[J.M. Noble]

No offense, but starting with a so-called "subwoofer" that's at least -6dB at 40Hz (or 80Hz for the sealed?!?!) and going nuts with EQ to manhandle the bottom octave out of it seems like a losing fight with physics.

Yes, I am aware that this is how a lot of subs, including good ones with solid low 20s response, are made. But they are almost universally doing that with much smaller drivers and the designs are compromised in favor of the least obtrusive box possible. Anyone playing with dual 15s has obviously abandoned stealth and has no neighbors, so just build a big box around the right driver(s) (low Fs with almost certainly a very large Vas) and reap the performance rewards.

 

[ZolaIII]

https://www.diyaudio.com/community/attachments/linkwitz-transform-closed-box-and-vented-box-tutorial-toc-pdf.1253761/

 

[Heinrich]

No offense, but starting with a so-called "subwoofer" that's at least -6dB at 40Hz (or 80Hz for the sealed?!?!) and going nuts with EQ ...

Anecdotally I confirm it works. Power, and withstanding the power is cheap today, living room is not. 24bits, hence lots of room for arbitrary equalization is readily at hand. At a certain point, dictated by physics, the distortion of the compressed air will take over, but only so much as the compression is adiabatic. If stuffed the enclose operates more linear in the isothermal regime.

@ppataki, that might answer your question above. The tension of the suspension combines happily with the pressure difference from the box to result in the one and only resonance frequency. Some p/a drivers sport an intentionally non-linear suspension as a protection measure. Such won't be of any concern here.

btw, I gave up on nit picking with Thiele/Small parameters and derivatives, simulations. The room, position, heat, program, too many variables. I just throw in what I have and what a box of a given size can bear, equalize later.

 

[J.M. Noble]

Anecdotally I confirm it works. Power, and withstanding the power is cheap today, living room is not. 24bits, hence lots of room for arbitrary equalization is readily at hand. At a certain point, dictated by physics, the distortion of the compressed air will take over, but only so much as the compression is adiabatic. If stuffed the enclose operates more linear in the isothermal regime.

I refer to this above in the context of smaller drivers. OP has 15" drivers, so even an acoustically "small" box is going to be physically large.

As someone who designed subwoofers professionally with LEAP/LMS from the early 90s, I dislike the "jam it in a shoebox and shove watts into the voicecoil" approach: that huge Xmax is unlikely to be very linear (play some test tones, you'll hear it--no need for a distortion analyzer), and its inefficiency offends my engineering sensibilities.

 

[Heinrich]

... and its inefficiency offends my engineering sensibilities.

I know how you must feel ;-) As Boris Vian, the highly esteemed pataphysicist, said: “I hate wear and tear!”

But what would you do?

Take a box of 8 liter.

The usual approach would be to install a 10cm chassis and somehow get a port to 40Hz. Wind noise, resonances are the result, and not much low bass, which you would hardly hear because of Fletcher/Munson's curve.

Because it would be reflex, the nominal efficiency would be about 6dB higher than sealed, but only at the tuning frequency. Everwhere else the efficiency drops to virtually nil.

The alternative is therefore closed with a maximum diaphragm area. Although the resonance is then very high, this is not a disadvantage compared to the same sealed speakerbox equiped with a smaller driver. This is because the efficiency depends practically only on the volume, not on the alignment. Above the low bass, the larger cone surface always has an advantage.