Nuance W10-NS1020-12 10" Woofer for OB project?

[charlielaub]

I have been mulling around an OB project for a long while. I have a good mid that I would like to use, but I have been held back for a long time trying to find some woofers that would work well but will not break the bank.

Today I happened to do some casual modeling of three of the Nuance W10-NS1020-12 10" Woofers (p/n 299-083) connected in parallel on a huge (48"x48") planar open baffle (don't ask!). According to the specs posted, the drivers have a high Q (1.4 ish), high Xmax (9mm) and Re is 9.5 ohms so you could safely do about three in parallel. According to my modeling, these seem to fit the need nicely and are inexpensive.

I have been following with close interest the approach that Martin King has recently espoused, which uses high Q drivers to partly compensate for the open baffle loss. The idea, as far as I can make out, is to find a driver that has a resonance frequency (Fs) just below where you would like the system cutoff to be located (e.g. 40Hz system cutoff, shoot for Fs=30-35Hz) and Qts>1.3 or so. This all seems to make sense to me.

Anyway, I was wondering if anyone else has thought about using this driver for this purpose (an OB system) or has purchased one or more of these drivers and actually measured the T/S parameters. I would like to know how accurate the parameters listed on the product page are...

So what are your thoughts about using these drivers for OB woofers?

I could post some graphs from the modeling if anyone is really interested...

-Charlie

[Mayhem13]

I think there was a discussion about these for OB use once before and yes.....the specs look good for MJK's approach on a budget. Given the Xmax, three of these should displace as much as the Orion's pair of Peerless XLS 10's given the cone area comparison. With three of these, maybe a fullrange driver ala MJK with two W10's under and a single OVER? Those new Mark Audio fullrangers are really getting some great praise from the fullrange/OB crowd.

[charlielaub]

I think there was a discussion about these for OB use once before and yes.....the specs look good for MJK's approach on a budget. Given the Xmax, three of these should displace as much as the Orion's pair of Peerless XLS 10's given the cone area comparison. With three of these, maybe a fullrange driver ala MJK with two W10's under and a single OVER? Those new Mark Audio fullrangers are really getting some great praise from the fullrange/OB crowd.

Thanks for your thoughts... the Nuance drivers do seem to work for a budget OB application...

I already have a pair of B&C 8PE21 that I will use as the "mids". I'm aiming to put together a 3-way active OB system with all of these, although I am not sure about which tweeter to throw in to the mix yet. I have some around that might work well, but I will probably wait until I can actually measure the response of the mid in baffle before deciding what to use.

-Charlie

[CUtiger]

Well I just got four of them in yesterday and am still waiting on my wt3. With all the talk going on about dipoles recently I decided to go ahead and get four of these to experiment with (I needed like 35 more dollars for free shipping anywas so in reality the quad only cost me a net $55 or so). Nothing to report on in terms of parameters or performance but the build quality actually seems to be quite good considering the price. No terribly rough edges, misaligned bits or sloppy glue lines. I had some old 6.5" axon 6s1 buyouts from a long time ago just laying around I'm gonna try and make work as a midrange, just need to get some tweeters that look capable and cheap as this is just a budget experiment.

[Mayhem13]

Given the paper cones i suspect these will play quite cleanly to 300hz....where they;d work quite nicely with a full range driver sans the complex crossover work and multiple drive units. Reading Martins paper (again!) linked by Charlie indicates a parallel pair of these would be a good match to a driver in the 84-85db efficiency range...or in other words just about every 3-4" tang band offering available. The Fountek FR89 would be an EXCELLENT candidate as well.

[blair]

Hello Charlie,

I am very interested in you measurements.

Thanks!

[charlielaub]

Hello Charlie,

I am very interested in you measurements.

Thanks!

It's only a model...

-Charlie

[charlielaub]

OK, here are the details of my modeling approach, and the results:

In order to get the OB response, following Martin King's concept, you need to take a few separate steps:
1. Model the baffle response - I did this using "The Edge". Although I think that this program does not model floor or room boundary response effects, it can handle multiple drivers (three in this case). I used the following baffle model for a 48"x48" plain baffle and three of the Nuance drivers:



2. I then wanted to compare this to the response of the same system (same size front baffle) if it were an infinite baffle system and then normalize the open-baffle response to the passband level for the IB system. This is because I will be using a driver response modeling program (WB&CD) to mimic the driver's response in an open baffle by using a closed box model and setting the box volume to, essentially, a very large number. Thus, the correction that I would like to apply to the closed box model should take in to account the difference between the closed box model and an equivalent open baffle. Since I believe that WB&CD models acoustic radiation in to 2-pi space, I need to model the OB response as deviation from this. Thus the response should be modeled as a "loss" with 0dB reference level as the passband response level. This graph explains this concept, except that rather than a loss it shows the passband level as a gain over the SPL at DC. While the shape is right, I need the passband level to be at 0dB:



Using The Edge, I found the following responses for a closed box, infinite baffle system with a 48" x 48" front baffle:



...and the equivalent open baffle system:



As you can see, the response between 200 Hz and 500 Hz is about 1dB higher for the OB system. The CB/IB response shows the usual 6dB loss between DC and the passband level, while the OB system output continuously falls at 6dB/octave below about 150Hz. Both of these models are for a listening position that is about 2m off axis, leading to the roll off at frequencies above 500 Hz. The interference of the three sources also contributes to the high frequency roll off.

To obtain the response correction that I was looking for, I subtracted about 5dB from The Edge's OB response model to get the following response curve:



Again, 0dB in the above graph is the infinite baffle passband response level. This is finally what I needed to apply to my CB/IB model of the three Nuance drivers to convert it to the OB response.

Next, this was imported in to WB&CD using the "Room or Baffle Response" import functionality.

3. After importing the normalized response/correction to WB&CD, I found that adding a second order HP filter (Q=1.4 @ 37Hz, about 4dB boost), the response was flat down to about 40Hz. This is definitely a low frequency cut off that I can be happy with, for an OB system. A better option, which would result in additional low frequency extension, is a first-order shelving LP filter, but unfortunately I can't accurately model that in this software.

Here is the resulting model of the final OB system at 60W input power:



This is the cone excursion for the drivers, also at 60W input power:



-Charlie

[diy speaker guy]

OK, here are the details of my modeling approach, and the results:

In order to get the OB response, following Martin King's concept, you need to take a few separate steps...

-Charlie

I have a couple of comments and questions.

1. Why did you decide to use such a massive baffle? A quick glance at the t/s parameters suggests to me that you could get the same LF extension with a much narrower baffle at the cost of a bit of spl.

2. There are much easier ways to model this. Essentially, this is what MJK's software is for, so why not use that? With that software you could model what you have here in about 5 minutes, starting with a model that you could actually afford to make with passive components. Even using Bagby's software it doesn't have to be this complicated. I'm not sure why you are using IB/closed box modeling, there are much easier ways to do this.

3. Technically, (if I understand what you are doing here) this isn't how MJK does it in his famous OB paper. He uses the first dipole peak as the xo point (around 400 hz or so with his 20 inch wide baffle) and it works very well. Even with high q drivers, it's the only way to raise the xo point high enough to work with reasonably priced passive components.

[charlielaub]

I have a couple of comments and questions.

1. Why did you decide to use such a massive baffle? A quick glance at the t/s parameters suggests to me that you could get the same LF extension with a much narrower baffle at the cost of a bit of spl.

2. There are much easier ways to model this. Essentially, this is what MJK's software is for, so why not use that? With that software you could model what you have here in about 5 minutes. Even using Bagby's software it doesn't have to be this complicated. I'm not sure why you are using IB/closed box modeling, there are much easier ways to do this.

1. The baffle dimensions are just something I have been playing around with. They also happen to work well for my midrange, which has a very low Q and small Xmax and therefore needs the help of a wide baffle.
As you say, you could use a narrower baffle, but you will lose more and more output on the low end as the baffle narrows. The system will require more input power and will become excursion limited at a lower SPL. IMHO 105dB at 40Hz is already about as limited as I would like it. Since I don't know what the distortion performance of these drivers is like, I would rather err on the conservative side WRT excursion requirements.

2. I don't have MJK's software. I am just using what I have...
Jeff Bagby's DBS modeler cannot model multiple drivers and since I do not care about the floor or room boundary reflections, the Edge fits the bill nicely.

Also, your on point #3 (that you edited in while I was responding to your original post) - I'm not making a passive speaker, so I don't strictly need to push the crossover point up to 400Hz, although if you look at the final response graph the response does extend up that far... So, the focus of everything that I posted was done to model the low end response. I will worry about the crossover (to the mid) later, since I have much more flexibility when using active electronics for that purpose.

Anyway, if you want to offer up something and add it to the thread, be my guest.

-Charlie

[diy speaker guy]

1. If your mid's fs is about 2 octaves below the xo point, q and xmax are not really a factor.

2a. IIRC there's at least 2 workarounds to the number of drivers used with Bagby's software. Pete S. told me how but I don't remember exactly, ask him.
2b. I'll model the woofers for you with MJK's software, using his method with a narrow baffle just to let you know what it looks like. Hopefully quite soon, although I might get hijacked into babysitting for the next few hours. If you don't like how that looks I'll model it on a larger baffle, but the wider the baffle gets, the lower the dipole peak gets, the more expensive the xo components get (if passively crossed).

[charlielaub]

1. If your mid's fs is about 2 octaves below the xo point, q and xmax are not really a factor.

2a. IIRC there's at least 2 workarounds to the number of drivers used with Bagby's software. Pete S. told me how but I don't remember exactly, ask him.
2b. I'll model the woofers for you with MJK's software, using his method with a narrow baffle just to let you know what it looks like. Hopefully quite soon, although I might get hijacked into babysitting for the next few hours. If you don't like how that looks I'll model it on a larger baffle, but the wider the baffle gets, the lower the dipole peak gets, the more expensive the xo components get (if passively crossed).

2a. Well, that is a generalization that doesn't exactly apply to my particular driver - I believe that this is a rule of thumb that is followed when using passive crossovers. Anyway, you are welcome to check it out for yourself - it's a B&C 8PE21. Locate it right at the top of the 48x48 baffle and slightly off center...

2b. OK, cool. I am definitely interested in seeing the results of your efforts using MJK's software. Can you do one try at the 48x48 baffle to see how the results compare? Does his software show the driver excursion as a function of frequency or only frequency response? One of the reasons that I go through WB&CD is so that I can see the excursion, impulse response, and add in filter sections as needed.

-Charlie

[diy speaker guy]

Ok, so I have some good news and some bad news. Let's start with the bad.

MJK's software is not as sophisticated as Bagby's in some ways but what it does it does very well and allows for very quick changes. It does not use frd's, only t/s specs. It does show excursion vs frequency, impulse response, allows for filters and much more. But the OB worksheet is only set up to allow for a maximum of 3 drivers, and it is assumed that you will be modeling a single fullrange driver (or midrange) on top of a pair of identical woofers. So, for example, there's no way I could model all 3 woofers with a crossover and your midrange. (There is a workaround but I wouldn't trust it to be accurate enough for final design xo work.) This is why I only volunteered to model the woofer section. I assume this is good enough, since you were only modeling the woofer section anyway.

The above noted limitation also means that I cannot show you a single graph with all 3 woofers optimally filtered, since one of these 3 drivers can only be high passed - or run without any filter at all. (I totally forgot about this limitation and I apologize.)

So the bad news is that if you stray too far from MJK's fullrange and woofer(s) method, the software is limited. Even more limited than I thought. But I can still show a bunch of very useful information.

And now the good news. If the published specs are reasonably accurate, this looks like an ideal woofer for this type of OB project. You can build it active if you want but these are a great candidate for an MJK style passive xo with a 12 db/oct filter around 200 or 250 hz.

On to the first graph. This shows the unfiltered output of all 3 woofers at xmax (9mm) on a baffle 20 inches wide by 48 inches high. The woofers are all centered horizontally and are located at 6, 17 and 28 inches (centered) from the bottom of the baffle. This leaves enough room at the top for a mid and tweeter. This model includes floor bounce (it actually helps a bit) but no other boundaries. The dotted red line is the top woofer, the blue dotted line is the combination of the 2 bottom woofers. The solid red line is the combination of all 3 woofers. The virtual mic location is arbitrarily set at 40 inches high.



Remember that this is completely unfiltered output. It looks pretty good. From these results, I would guess that a simple 12 db/oct lowpass around 200 - 250 hz or so would net a result of 105 db from 40 - 200 hz (+/- 1 db). Without the floor bounce effect you might not get quite up to 200 hz as flat.

Based on these results, I personally don't see any reason to go wider than 20 inches, and in fact I'd probably make it even narrower than that. I'll still model it at 48 inches wide if you like but I don't see a need for that kind of width.

Next, if you like, I can model each of the 3 drivers individually, unfiltered and filtered (at 200 or 250 hz, whichever looks better) for max flat bandwidth. But due to the incredible amount of variability in the settings I'd ask you to define the parameters you would like to see. Just tell me the baffle width, height, driver locations, mic height, boundary conditions (floor, side and back wall, ceiling) on or off.

And just to make sure, before going any further, please confirm the specs I'm using.
Fs: 34 Hz
Le: 1.77 mH
Re: 9.5 ohms
vas 72.5L
Qes: 1.75
Qms: 6.79
sd 345 (just a guess)
bl 7.7

[charlielaub]

It looks like the system, using a 48" baffle, is about 3dB more efficient than the same thing using a 20" wide baffle. The first maximum shifts in frequency a little as well, but this is not big deal. IMHO the main issue is that to get up to realistic SPLs, you need as much from three of these drivers as they can possibly deliver. If you want to go with the narrower baffle, you might want to use four drivers, although connecting them all in parallel requires an amp that is capable of driving that (low of a) load, e.g. can put out a lot of current in to 2.75 ohms.

I think what you are showing with the MJK modeling package is very similar to what I am getting, despite the differences in the programs and approaches. The bottom line is that both the Fs and Qts of these drivers is very well suited to an open baffle system that provides extension down to 40Hz.

-Charlie

[diy speaker guy]

I agree. I was just trying to point out there are easier ways to model the woofer section and easier ways to implement the xo, especially if the design is to be compatible with passive components, which is a benefit and adds value (IMO) even if you don't plan to use passive components in the short term.

OTOH, if you are happy with your multi step modeling routine and having to use active xo then that's fine too. I don't think the extra 3 db is worth the extra 28 inches but that's just personal preference.

FWIW, this high q woofer method is by no means the only way to get wide bandwidth flat response if you have an active xo in the mix. You might also want to look into low q woofers (around .3) with a very low (around fs) and very shallow low pass to get the same type of results.

With active xo and a bit of dsp any woofer will work.

[brianpowers27]

I think you have a winner with the 20" configuration. The narrower configuration will provide some directivity benefit over the larger configuration. 104db per side at xmax will likely yield the ability to peak around 115db when both are run.

[charlielaub]

I think you have a winner with the 20" configuration. The narrower configuration will provide some directivity benefit over the larger configuration. 104db per side at xmax will likely yield the ability to peak around 115db when both are run.

Brian...

I don't quite follow your math. If one speaker (having three of the Nuance drivers) can produce 104dB at Xmax, how does two speakers product 115dB? Adding the second speaker should only increase output by 3dB.

-Charlie

[brianpowers27]

Brian...

I don't quite follow your math. If one speaker (having three of the Nuance drivers) can produce 104dB at Xmax, how does two speakers product 115dB? Adding the second speaker should only increase output by 3dB.

-Charlie

Maybe you are right that you will not gain that much. Consider that the simulations are built using sine waves and we seldom listen to those. Also, x max figures tend to be somewhat conservative. Bottom line, I bet it would work fine for your application.

[Mayhem13]

This is true for most drivers where xmax and xmech are around 70% respectively.

I just recieved and unexpected Bonus from work and i'm pretty much sold on a budget built dipole now Charlie using the Fountek FR89EX 3" fullranger ala MJK's models. I figuring a passive cross at 300hz for the fullranger should work nicely.

https://www.madisound.com/store/prod...0808617147e451

[diy speaker guy]

I didn't mention this before but if you look at the graph I posted you can get near 110 db out of those 3 woofers on the 20 inch wide baffle if you can live with a less than perfectly flat response down to 40 hz. It drops about 5 db between 200 and 40, and it would match room gain better that way too. Just cross over higher.

As far as adding a second speaker (another 3 woofers), I was under the impression that doubling the cone area nets a 3 db increase and doubling the power gains another 3 db, for a grand total of 6 db.

Either way, I'd still vote for a baffle even smaller than 20 inches wide. 12 inches wide would only lose a couple of db from the graph I posted, it would look much better, fit in your car if you wanted to show them off at at diy show and would still be plenty loud enough if you set them up as described above, letting them roll off naturally below 200 hz instead of shooting for perfectly flat response down to 40 hz.