This update is about a fully passive version of the 3-Way Retro. Why would you like a passive version? I personally like to swap between my DIY amplifiers and my restored vintage gear. Active filters requires multiple power amps. Not much of a problem if you keep the active XO (and/or DSP) in a separate box but it is still a lot of gear compared to the simplicity of a single integrated amp.
An active high order filter usually performs better than any passive filter, at least on paper. But like so much else in electronics and Hi-Fi, everything is a tradeoff. A fully passive build is a tradeoff, convenience and interoperability over response and performance. I will personally take convenience and interoperability any day as long as the tradeoff isn't too big.
I have used VituixCAD long enough to actually trust its simulations. I therefor went with the following second order filter from my previous simulations. Second order because it simulated well enough and because it is half the amount of expensive components for a LF filter @ 600 Hz.
I had to adjust the simulated values a bit to values in stock at my favourite local dealer
HiFiKit. I went with Jantzens Cross Caps as usual, they are hard to beat from a cost/performance perspective. I would have liked a thicker winding on the Janzen inductors but had a hard time justifying it to my self. I rather go air core with a bit thinner wire than iron core.
Simulated values to selected component values:
- 33.2 uF -> 33 uF
- 16.6 uF -> 18 uF
- 2.12 mH -> 2 mH
- 4.24 mH -> 4 mH
- 1.17 R -> 1.2 R
- 9.7 R -> 10 R
This is what the new board looks like. I made separate input and output connections in order to enable bi-wired input connection to the speakers Neutrik speakON connectors. The LF output connects to the woofer and the HF output connects to the existing board (the XO between mid and tweeter). This makes it easy to bypass any board in the future if I want to run passive, active, uni-wire or bi-wire etc.
The new board sits to the right of the existing board. It looks a bit like a rats nest but bi-wired input requires a bit extra and I have also made interconnections to mid and tweeter so that I can invert the input without having to dismount the elements. Any phase corrections can therefore be done at the XO end. Picture taken before the "speaker fluff" was put back in (note polarization of HF, it has to be swapped, more about it later).
I did this conversion a while ago but I did not want to publish it before I got the opportunity to do some outdoor measurements. I did not make a lot of measurements on a turn table since I already built the XOs and I know what to expect when it comes to directivity (see
previous measurements as part of my 3WC build). I did not calibrate for SPL either but the sensitivity of the woofer is 88.5 dB.
A note on the importance of measurements and polarity. This is my first un-smoothen far field measurement:
It is easy to spot the cancellation around the lower XO point. Reverse the polarity to the HF branch and you get this un-smoothen far field measurement:
Not bad from a second order text book LR filter @ 600 Hz. Next up the near field measurement of the woofer (un-smoothen and measured between 5 Hz and 1 kHz):
I used the VituixCAD Merger Tool to splice the far- and near field measurements together (baffle step taken into account) and this is what it looks like un-smoothen:
And like this with the Psychoacoustic Smoothing applied in REW:
I am super happy with the result and I was pleasantly surprised how good it sounded with just a regular integrated amp without a DSP. It isn't ruler flat but the small imperfections are super easy to correct with a DSP. The F3 is about 50 Hz and it works well with music. The speakers are supposed to be positioned close to the wall and it will add some bass making up a
more preferable target line with a bit of slope from bass to treble.
Another thing I wanted to try with a fully passive build was to extend the bass with EQ. This is a sealed enclosure and I personally tend to think about extension with a Linkwitz transform as something that requires active cross overs. But it doesn't have to be if you got the opportunity to extend the bass before amplification. My new
DIY integrated got a DSP as the preamp making it easy to try this concept out on this fully passive version of my 3WR. I just used the
Single Band Automatic Speaker EQ block in SigmaStudio instead of the Two or Three Way Speaker PEQ block. Extending the response down to 30 Hz will require around 10 dB amplification but the rest of the filters will be minor. I used 8 filters on each channel and applied a -6 dB slope from 20 Hz to 20 kHz. The sound is great without EQ but it is lacking the grunt of a subwoofer. With bass extended down to 30 Hz I personally doesn't find the need for a sub, not even for TV and movies on a big screen. I also think the SB23NRXS45-8 deserve the extra extension since it sound so good in a sealed enclosure with a bit of EQ down to 30 Hz.
This might be the end of the road for my
3-Way Retro build (except for new cabinets whenever I get the time). It has been a long and winding road but the end result was well worth it. It all started with
Troels Gravesens SB Accoustics 3-Way Classic. I personally thought the
SB29RDC-C000-4 or
SB29RDAC-C000-4 a bit of overkill and went with the 3/4-inch
SB21RDC-C000-4 instead. Troels XO points @ 600 Hz and 3.7 kHz was spot on for these drivers but I personally didn't like the sound from the ported enclosure. I therefor made another attempt with the same drivers but in a sealed enclosure. Initially as a semi-active build with a Linkwitz transform to extend bass. A straight forward build that works great even fully passive. I love it this way since I can use it with all my vintage stuff as well as with my DSP enabled
Ultimate Integrate.
A final note: Is it worth building this speaker? The cost for the elements is around €400, add parts for fully passive XOs and its another €200. Enclosures, veneer and connectors, another €150. Say around €700 in total for a finished pair. So how do they stack up against some commercial offerings? Keep in mind that I might be a bit biased and that sound preferences are quite subjective, but here goes...
I got a friend who is selling speakers for a living. He was nice enough to lend me his listening room for two hours. I picked a pair of best in class speakers representing the following price points; €500, €1000, €2000, €4000 and €8000 (the €8000 one was not best in class but still quite good). No competition at the €500 and €1000 price points. I would put the fully passive version without EQ somewhere between the €1000 and €2000 price point (closer to €2000 than €1000). The €4000 speaker was a clear winner but with a couple of caveats. It was a modern looking slim floor stander and it was butt ugly. The €8000 speakers was even better but more than twice the size of the €4000 speakers (and I doubt they would sound that good in my living room). I would put the EQ and semi-active version well above the €2000 price point but not all the way up to €4000. But who knows, I think there is still room for improvements, especially with a better, less resonant cabinet...
