B&W Matrix 802 Series 3 crossover rebuild - March 2006

I was persuaded to change the crossovers in my Matrix 802 speakers, even though B&W themselves explain in their FAQ that this isn't necessarily a good idea. Over the prior year I had made big improvements to my amplifier, building the Ella kit and slowly upgrading it, but felt that there was still something missing, particularly in the lower frequencies. The Ella is only 50W after all. I'd looked at the original crossovers and seen the electrolytic capacitors, which I've read deteriorate with age, and the inductors that were getting a bit rusty and decided to try changing them.

I had upgraded from B&W 603s2 speakers a couple of years earlier and really liked my 802s so my main worry was that I might damage them or change them in a way that I couldn't reverse. So rather than change individual components one at a time I thought the safest approach was to build complete replacement crossovers.

There were some additional requirements that I decided on at the beginning. First I wanted the crossovers to stay inside the speakers as I didn't have the floor space for additional boxes and didn't want to be bothered with having additional wires and connectors. Also I wanted to make the B&Ws as easy to drive as possible for use with my Ella. Both of these requirements meant that I had to stay with smaller, lower resistance iron or steel cored inductors rather than the larger, better sounding air-cored inductors used by North Creek in their external crossovers and B&W in their newer speakers.

I made a prototype and put it into one speaker first so I could compare the old against the new. This wasn't easy as there aren't many good quality mono recordings, but I did notice the sound was clearer and the speaker seemed more dynamic. So then I built a finished crossover for the other speaker. Once both speakers were updated it was clear that this was a big improvement. There was some harshness in the mid range and I hadn't increased the bass much, but both of these were easily fixed by adding a couple of additional resistors. The speakers seem really well balanced and better in all types of music. Photos of the changes are here.

The downside to the process was the time it took (I suppose about 40 hours in all, although I'm not a very fast worker) and the speakers are now about 2kg heavier. I'll have to be careful next time I move apartment to keep the speakers upright as I'm not sure how well the screws I used will hold the additional weight. The weight might also be an issue if I didn't have the Sound Anchor stands. These stands support the speaker right across the base, but if I was only using the original spikes then the full weight of the crossover would need to be held by the screws that attach the base to the main cabinet. The cost of the change ended up at about US$200 which I feel was definitely money well spent.

Having spent so much time improving the crossovers, it's interesting that I have since upgraded to speakers that have no crossover at all, the MarkAudio Pensil 12s, but I still enjoy listening to the B&Ws on my second system.


The circuit diagram for the Matrix 802 Series 3 is:
Circuit diagram
The original parts used by B&W were:
Item Description Values
C1 Bennic PMT polypropylene 4.7 uF 160 V
C2 Bennic PMT polypropylene 10 uF 160 V
C3 Bennic bipolar electrolytic 26 uF 70 V
C4 Bennic bipolar electrolytic 10 uF 70 V
C5 Bennic bipolar electrolytic 150 uF 70 V
       
R1 Bennic wire-wound in cement 3.3 R 7 W
R1 Bennic wire-wound in cement 3.3 R 7 W
R2 Bennic wire-wound in cement 0.56 R 7 W
       
L1 Air core inductor 0.2 mH  
L2 Ferrite core inductor 1.5 mH  
L3 Ferrite core inductor 1.0 mH  
L4 Wicon Ferrite core inductor 2.0 mH  
L5 Wicon Ferrite core inductor 2.0 mH  
       
  MR/HF PCB board    
  LF PCB board    
       
  Van den Hul CS12 wire (all drivers and posts)    

These were replaced with:
Item Description Values From
C1 Obbligato aluminium film in oil 4.7 uF 630 V DIYhifiSupply, HK
C2 Obbligato aluminium film in oil 10 uF 630 V DIYhifiSupply, HK
C3a SCR MKP polypropylene 18 uF 400 V  
C3b SCR MKP polypropylene 8 uF 400 V  
C4 Obbligato aluminium film in oil 10 uF 630 V DIYhifiSupply, HK
C5 Solen polypropylene PB Fast Cap 150 uF 400 V Madisound, US
  Vishay MKP 1837 0.01 uF 100 V RS Components
         
R1 Mills wire-wound 3.3 R 12 W Sonic Craft, US
R1 Mills wire-wound 3.3 R 12 W Sonic Craft, US
R2 Mills wire-wound 0.56 R 12 W Sonic Craft, US
R3 PEC enamel coated wire-wound 0.22 R 7 W See Note 1
R4 PEC enamel coated wire-wound 0.22 R 7 W See Note 2
         
L1 Solen Hepta-litz air core inductor L16 0.2 mH 0.180 dcr  
L2 Sledgehammer 15AWG steel laminate 1.5 mH 0.140 dcr MarkAudio, HK
L3 Sledgehammer 15AWG steel laminate 1.0 mH 0.100 dcr MarkAudio, HK
L4 Sledgehammer 15AWG steel laminate 2.0 mH 0.156 dcr Madisound, US
L5 Sledgehammer 15AWG steel laminate 2.0 mH 0.156 dcr Madisound, US
         
  Acrylic mounting board (210 x 170 x 3mm)      
         
  Gotham 2 x 1 sq.mm (HF, MR drivers, board wiring)     DIYhifiSupply, HK
  Gotham 4 x 2.5 sq.mm (LF drivers, posts)     DIYhifiSupply, HK
Note 1: Resistor R3 was added after the parallel R1 resistors and before the C3 capacitor. This removed the increased volume and some harshness in the midrange. The same result could have been achieved better by buying the Mills 3.7 ohm resistors for R1.

Note 2: Resistor R4 was added between resistor R2 and capacitor C5. This added more bass. The same result could have been achieved better by buying the Mills 0.75 ohm resistor for R2.

Update - October 2008

I was contacted by a René in the Netherlands who had used the parts listed above to replace the crossovers in his Matrix 802s3 speakers. He had seen a review by Tony Gee of the Vishay MKP 1837 used as a bypass capacitor, and he found them to be a great improvement.

So I tried them on mine, at first adding them in parallel to caps C1, C2, C3 and C4. But although this did make the sound clearer, for me it made the treble too pronounced with the tweeter really grabbing the attention. So after a few days I tried removing the bypass on C1. This was a definite improvement as some of the clarity was kept but leaving the speakers sounding more balanced.