Further tinkering with the PMS active build is on hold for now - Mrs Mad Scientist is growing tired of my measurement stuff in the lounge, so I decided to have a go at the Man Cave setup.
It’s a 2-way speaker, so only four channels needed for active conversion. My first mistake - I snagged a really nice condition Motu M4 from Reverb only to find Roon doesn’t do 4 channels, only 2, 6 (5.1) or 8 (7.1) 
So, I managed to snag another Motu Ultralite Mk5 from eBay. Also picked up a second Topping PA5II. Anyone who’s seen my other speaker build thread will know I’m a fan of Neutrik Speakon connectors:
The PA5IIs have very small binding posts, so some surgery was in order:
Both PA5IIs now sport a pair of Neutrik 2-pole sockets.
For the speakers, I wanted to go beyond just converting them to active. I’ve been intrigued by wavegudes lately, especially in relation to matching midrange/midwoofer directivity with the tweeter. My desktop speakers used to live in the lounge - a pair of Zaph Audio ZRTs in a 20 litre sealed box. Really nice Scan-Speak drivers, the sliced paper 18W8531G00 Revelator midwoofer and the D3004/660000 “AirCirc” Illuminator tweeter.
With the desktop setup, they were always a little “bass-light”, so I used to run them with a +3 dB low shelf PEQ in Roon. I didn’t want to go with a ported enclosure, so opted for a Scan-Speak Flow Resistor, which effectively increases the sealed box size by about 20% and gives a 3rd order roll-off from the box tuning frequency:
I have a bunch of tweeters and waveguides kicking around, but for this decided to grab a pair of Monacor DT-300 tweeters and WG-300 waveguides (discontinued, though showing as in stock at one of my regular suppliers). Turns out they weren’t in stock, but the supplier had access to an alternative waveguide and 3D printed and assembled an adaptor to fit.
The waveguide is more or less the same size as the midwoofer, and from measurements of the DT-300/WG300 elsewhere should be a perfect directivity match to the 18W8531G00.
So I took the drivers out of my cabinets, ripped off the front baffle on the table saw and set about making a test baffle for the new layout.
First set of Acourate measurements of the tweeter showed a bit of a problem:
There’s a big @ss suckout centred around 3990 Hz. I tried lots of different measurement distances from the width of the waveguide out to 3 m from the front baffle. No dice, it was still there, even off-axis. Need to try it on a big baffle to see if it’s still there. If it is, it’s a waveguide loading issue - the wavelength @3990 Hz is 8.6 cm, approximately the radius of the wavegude. I ran some sweeps in REW too - nothing funky on the CSD or distortion plots, just a big drop in amplitude.
Anyhoo, with Acourate’s driver linearisation, it wasn’t much of an issue.
Linearised crossover at 1700 Hz:
The DT-300 with waveguide is about 98 dB sensitivity vs the 18W’s 87 dB, so the dip wasn’t going to lose much overall in the scheme of things.
Because of the tweeter being recessed in the waveguide, time alignment adjustment with the midwoofer was minor. The tweeter (at sample 6000) is delayed for test purposes by 1000 samples which puts the relative midwoofer peaks around ~5000:
Left midwoofer was 4996 samples and right was 4997 samples, so 4 and 3 samples early respectively which means they needed a 4 and 3 sample delay. ~ 0.08 and 0.06 ms. Acourate’s time alignment is mind-blowingly overkill in it’s level of precision.
Initial sweeps L & R and target curve (Bruegemann):
Inversions relative to Target:
Measurement of applied final filters:
Pretty much +/- 2 dB 30 Hz to 20 kHz
Step Response:
ICCC (IACC):
RT60 is good enough for a recording studio:
The convolution filters are uploaded to Roon’s Muse and my RPI4B is connected via USB to the Motu Ultrlite Mk5.
So, how do they sound?
Just utterly jaw-droppingly incredible. For a near-filed setup, I’m utterly blown away by the size of the soundstage, and the pin-point precision of every element within it. I’m hearing details that I can’t pick out with my PMS wide-baffle 3-way. And the bass is phenomenal for a 6 1/2" midwoofer. I’ve spent hours listening to track after track on these since finishing them yesterday.
All that remains is to make a proper double thickness baffle, reduce the driver c-c distance (some remeasuring and tweaked filters will be required) and make it look nice. I already have the red polyester/carbon fibre skinning fabric:
One peculiarity I came across is that although Acourate can export Roon compatible multichannel convolution filters, the 4-channel versions don’t work. Roon doesn’t recognise 4-channel audio. I had no output from the right tweeter (channel 3, the fourth starting from 0)
44100 2 4 0
0 0
0 0 0 0
Cor1S44.wav
0
0.0
0.0
Cor1S44.wav
1
1.0
1.0
Cor2S44.wav
0
0.0
2.0
Cor2S44.wav
1
1.0
3.0
I had to manually tweak the config files to trick Roon into thinking it was a 6-channel setup, even though there are only 4 channels mapped with 2 filters:
44100 2 6 0
0 0
0 0 0 0 0 0
Cor1S44.wav
0
0.0
0.0
Cor1S44.wav
1
1.0
1.0
Cor2S44.wav
0
0.0
2.0
Cor2S44.wav
1
1.0
3.0
There’s a feature request for 4-channel support in Roon, but very few votes, so it’s not going to happen anytime soon, if ever. I guess not many folks use Roon for 2-way active stereo setups.
