[jmimac351]

Sharing this for the Wilson Audio Duette (Series 1) speaker sent in for measurement / upgrade.




Initial speaker Measurements / Comments from Danny:

The (original) on axis response:   It looks like they did not compensate for the baffle step loss at all. So there is quite a swing from 300Hz to 600Hz.



The (original) horizontal off axis:



The (original) vertical off axis:



Changing the supplied tweeter resistors:

Note: The factory Novel Crossover comes with 2 sets of resistors for "Free Space" and "Near Wall" placement.  The resistors are 5.3Ohm (Free Space) and 4.2Ohm (Near Wall).  There is a pair of binding posts at the top of the crossover for installing these resistors to alter tweeter output. (See pic of crossover above)



Raw drivers versus the on axis measured response:



The impedance curve showed a little internal resonance near the port tuning:



The spectral decay was pretty clean:

[jmimac351]

That’s ridiculous

I think we now see why the speaker sounds so 'hot'.

That's so bad it makes me wonder if some part of the crossover got knocked loose during shipping.

I’ve seen how Wilson crossovers are made. They encase them in resin so I doubt that.

Check the measurements that Jim posted on page two that were taken by the NRC. They match mine really well.

Oh, right you are!  It was just so scaled out it was harder to see:

[jmimac351]

Corrected Measurements by Danny Richie / GR Research

THANK YOU DANNY! You just fixed a ~$16,000 bookshelf speaker system for FREE!  YOU ROCK!!!!

New On Axis Response



New On Axis vs 10 Degree Off Axis Response



New Horizontal Off Axis Response



New Vertical Off Axis Response



New Crossover



New Spectral Decay



New Impedence

[Tyson]

Danny is a master.

[goggle1824]

What year(s) were these $16,000? (MSRP I assume).

[Hobbsmeerkat]

What year(s) were these $16,000? (MSRP I assume).

The V1 like these seem to have been available 2006-2012 timeframe.

[jmimac351]

What year(s) were these $16,000? (MSRP I assume).

Stereophile review from 2012 shows:

https://www.stereophile.com/content/wilson-audio-specialties-duette-loudspeaker

Specifications: Two-way, reflex-loaded bookshelf loudspeaker with separate crossover enclosure.
Drive-units: 8" cone woofer, 1" silk-dome tweeter.
Sensitivity: 90dB/W/m/kHz.
Nominal impedance: 4 ohms. Minimum impedance: 3.96 ohms at 3.1kHz.
Recommended amplifier power: >20W.
Dimensions: 18.4" H by 9.4" W by 13.75" D. Weight: 39 lbs.
Finish: Automotive paints in non-metallic black, Diamond Black, Dark Titanium, Desert Silver, and Argento Silver Twelve.
Price: $13,900/pair plus $1795/pair for matching stands. Approximate number of dealers: 50.
Manufacturer: Wilson Audio Specialties, 2233 Mountain Vista Lane, Provo, UT 84606. Tel: (801) 377-2233. Fax: (801) 377-2282.
Web: www.wilsonaudio.com.

"Overall, however, the Wilson Audio Duettes produced a sound that allowed all the music I listened to during my visit to communicate very effectively.—John Atkinson"

[mkrawcz]

Those speakers are going to sound awesome now.

[jmimac351]

Let it be known... I really like these speakers and just wanted them to not rip my head off when I played them On Axis.  100% going to do this upgrade.  I have a milling machine... I'm thinking of machining a custom enclosure for the crossover... keeping it external, and connecting via umbilical - just like the original.  The external crossover sits at the bottom of the stand in a machined spot, behind the vertical column.  I may have a clear lexan top to see the parts, recessed stainless fasteners, etc.  No permanent mod to the original design. Plug and play.

[Hobbsmeerkat]

Let it be known... I really like these speakers and just wanted them to not rip my head off when I played them On Axis.  100% going to do this upgrade.  I have a milling machine... I'm thinking of machining a custom enclosure for the crossover... keeping it external, and connecting via umbilical - just like the original.  The external crossover sits at the bottom of the stand in a machined spot, behind the vertical column.  I may have a clear lexan top to see the parts, recessed stainless fasteners, etc.  No permanent mod to the original design. Plug and play.

We're all excited to follow along with your upgrade process and hear about your thoughts about the sound before and after. (Especially since the stock crossover is already external)   

[jmimac351]

We're all excited to follow along with your upgrade process and hear about your thoughts about the sound before and after. (Especially since the stock crossover is already external)   

Will do, Hobbs.  THANK YOU to you as well!!!  I know you're right in there with Danny helping with all of this work.  It's so very much appreciated!!!  Danny kept raving about The Bully!  Said his wife was complaining about "The Bass". 

For those wondering... that Duette crossover is encased in epoxy.  There's no getting at those parts.  After chatting with Danny this afternoon, he mentioned the measurements themselves show no baffle step compensation applied... it is what it is.  I asked him... "So what did they DO?

  

If someone reading this knows the answer to that question... I'm sure many would be interested.  Whatever the case, the issue has now been rendered moot.  This speaker is right and proper now... and she looks good too.  And let me tell you... that 8" ScanSpeak woofer will CRANK. They play BIG.

What I was relieved about is the drivers themselves are excellent... there were "good guts" to work with... they stop and start / don't ring... as there would be no fixing that.  I already own a pair of speakers with that problem... 

[Hobbsmeerkat]

The Bully is a very exciting speaker, I personally like it more than the Brute, tbh.
We've just about got everything ready to go for it.

What's sad in a way is that the tweeter's on-axis response is hampered by the foam layer surrounding the tweeter causing the wavy response of the tweeter. (The grill makes it worse along with the hump at 600Hz)
Had they used a little more layered felt and not cut it perfectly round, the on-axis response could have likely been much smoother especially on axis. As it is, there's a regular pattern of peaks/dips, with the peaks in the on-axis becoming dips off axis & vice versa. The cabinet also needed a little more stuffing to control the resonance near the tuning frequency of the port.
I'd imagine that wilson put a lot of time/effort into picking good, quality drivers, cabinetry, paint etc, but it's interesting to see how little things here and there were seemingly overlooked leading to the measurements we get..

As far as what they did do with the crossover on the woofer? If I were to guess, they designed the crossovers using electrical parameters only, rather than the actual acoustic output of the drivers in that box. It certainly wouldn't be the first time baffle-step got missed in the design phase of a speaker..

[jmimac351]

The Bully is a very exciting speaker, I personally like it more than the Brute, tbh.
We've just about got everything ready to go for it.

What's sad in a way is that the tweeter's on-axis response is hampered by the foam layer surrounding the tweeter causing the wavy response of the tweeter. (The grill makes it worse along with the hump at 600Hz)
Had they used a little more layered felt and not cut it perfectly round, the on-axis response could have likely been much smoother especially on axis. As it is, there's a regular pattern of peaks/dips, with the peaks in the on-axis becoming dips off axis & vice versa. The cabinet also needed a little more stuffing to control the resonance near the tuning frequency of the port.
I'd imagine that wilson put a lot of time/effort into picking good, quality drivers, cabinetry, paint etc, but it's interesting to see how little things here and there were seemingly overlooked leading to the measurements we get..

As far as what they did do with the crossover on the woofer? If I were to guess, they designed the crossovers using electrical parameters only, rather than the actual acoustic output of the drivers in that box. It certainly wouldn't be the first time baffle-step got missed in the design phase of a speaker..

Very interesting.  So, you think if the felt were in strips / squared off around the tweeter / woofer, that would make a positive difference in on axis response?  Wilson sells replacement felt, so I could get some and trim to suit. 

I've got a UMIK microphone... I haven't figured out exactly how to do a solid measurement, yet.  I plan to do some other builds and Clio would be cool... (I wish the pricing were lower)

[Hobbsmeerkat]

Very interesting.  So, you think if the felt were in strips / squared off around the tweeter / woofer, that would make a positive difference in on axis response?  Wilson sells replacement felt, so I could get some and trim to suit. 

I've got a UMIK microphone... I haven't figured out exactly how to do a solid measurement, yet.  I plan to do some other builds and Clio would be cool... (I wish the pricing were lower)

The felt isn't really the issue, it's more the rubber/foam material that sits out in front of and surrounds the tweeter/felt.
It sits higher than the felt layer, and is also a perfect circular cut out, which means the diffraction issues that creates will all arrive at the same time, hence the regular patterns of peaks and dips in the tweeter measurements.

[jmimac351]

The felt isn't really the issue, it's more the rubber/foam material that sits out in front of and surrounds the tweeter/felt.
It sits higher than the felt layer, and is also a perfect circular cut out, which means the diffraction issues that creates will all arrive at the same time, hence the regular patterns of peaks and dips in the tweeter measurements.

Got it.  That stuff can come off. I see what you mean about the dips / peaks being inverse between on / off axis.  This is the On vs 10 degree off axis, and it is clear that it's just on the tweeter.  Thanks for pointing that out!!!

Now here's the big question... if I pull that stuff off, is there any chance it would impact the revised crossover design, or is it all just "benefit"?  It looks like if the On Axis peaks are smoothed down... it's all gravy. 

[jmimac351]

Original Response vs Corrected Response

I am going to work on improving response further based on Hobbs' feedback about the circular foam.

[jmimac351]

It appears that foam is in a recessed portion of the baffle and removing it may create other edges.  I wonder if buying replacement, identical foam from Wilson, then cutting it uniformly and placing around the tweeter... if that will have the desired benefit to break up the uniform circle. Hmm... another thing I could do is maybe buy a few more pieces of the current, jagged felt around the tweeter and then stacking them up and / or buy a bigger piece of pad and cut the "felt shape" and just place it over the felt. That's a "tape in place and measure" thing.

Fine tuning... I have no idea whether any of these options would work... HOBBS!!

"Oh, are you the guy modifying the Duette...?" 

[Hobbsmeerkat]

I wouldn't remove the foam.
The trick will be to build layers of soft felt to smooth the transition to the foam layer to help give it a gradual transition to the foam layer.
I played with felt strips a little bit but most the the bits of felt we have are on the larger side, and generally made matters worse in other areas.

It may affect the notch filter at 6Khz a bit, as it won't have to work as aggressive, but everything else should work the same.

[jmimac351]

I wouldn't remove the foam.
The trick will be to build layers of soft felt to smooth the transition to the foam layer to help give it a gradual transition to the foam layer.
I played with felt strips a little bit but most the the bits of felt we have are on the larger side, and only made matters worse in  other areas

It may affect the notch filter at 6Khz a bit, as it won't have to work as aggressive, but everything else should work the same.

Thank you!  I was just thinking this thru more and thinking of building up too.  Thanks again for pointing that out. 

I've got a UMIK microphone.  Will that with REW work as well as your Clio setup?  I've started collecting "How To's" but haven't actually taken a measurement.  I've seen things like keep window under 4ms to eliminate first reflection, etc.  I wish the Clio setup wasn't so pricey.

[Tyson]

If I were you I wouldn't stress too much about it.  It might be a fun project to learn about some stuff but really the overall response now is excellent, far better than most commercial speakers achieve. 

Also, IME, it's peaks in the 800hz to around 2khz that are the most annoying.  Above 5khz is mostly ambience information, so a spike would just be a bit more 'air' and a small dip won't be noticeable at all.