Active conversion of Kef 104/2

edavey

Wammer
Wammer
Jul 12, 2010
196
3
0
About 18 months ago I bought a pair of Kef 104/2 speakers from "Spider". They are an early model.

I played about quite a bit with their placement and with use of the included KUBE. Also measuring them in comparison to my B&W DM2a speakers in the same room. Here's the thread from mid-2017:

https://hifiwigwam.com/forum/topic/125677-measurement-of-kef-1042

I enjoyed the Kefs for about a year but this summer I moved the B&W back -- and despite the fact that they have a less flat response than the Kefs, there's something about their presentation that I just seem to prefer, though it's a very close thing.

I've decided to have a go at converting the Kef 104/2 to active crossovers. If I get it done and it's a success I might move on and convert the B&W DM2a.

I've done some reading here:

- starting with Serge's B&W 801 project
  https://hifiwigwam.com/forum/topic/47853-a-winters-project/

- general discussion of crossovers
  https://hifiwigwam.com/forum/topic/124334-crossover-design/

- candidates for active conversion
  https://hifiwigwam.com/forum/topic/125091-what-speakers-to-turn-active/

- "MF 1000"'s recent 104/2 conversion
  https://hifiwigwam.com/forum/topic/129572-kef-1042-restoration-conversion-to-3-way-active-build/

and abroad:

- MiniDSP's intro to active crossovers
  https://www.minidsp.com/applications/digital-crossovers/digital-crossover-basics

- Active Crossover Design toolkit in Excel
  http://audio.claub.net/software/ACD/ACD.html

- Measuring driver timing offsets
  https://app.box.com/s/ouxjjsx0m8bs00cil5iq

Based on this reading, never having tried one of these conversions before, my plan so far is:

1) measure them unchanged with their original passive crossovers, both in the room and outdoors (for a near-anechoic meaurement) so that I know what I'm trying to improve on.

2) remove crossovers and measure each set of drivers (tweeter, pair of mids, pair of woofers) individually outdoors and in the room so that I have the characteristics of each driver and also their relative timings

3) load this response and timing information into miniDSP

4) programme the miniDSP to reproduce the original passive crossovers (frequencies, slopes) with a view to measuring the anticipated superiority of the active crossover implementation

5) explore improved crossover designs to implement in the miniDSP software

Crossover

It seems like the best way for me to implement the crossovers will be using a miniDSP 4x10-hd (https://www.minidsp.com/products/minidsp-in-a-box/minidsp-4x10-hd) which takes analogue input and has up to 8 analogue outputs. 

Amplification

Whilst it might be fun to put together two 3-channel 'backplate' style amps using ready-made D-class boards, this sounds like unwise complexity for a first attempt. I will have enough on my plate getting the crossover, timing  and polarity issues solved without introducing the factor of a possibly flaky DIY amp. So in terms of using a pre-built amp I think the choice is either:

- 3 x stereo amp, e.g. like Serge's trio of Behringer A500 which he noted had the benefit of being dual-mono design, i.e. with very little chance of cross-channel distortion

- a multi-channel amp like the Rotel RB-976 which has 3 pairs of inputs and outputs and is rated at 60W per channel in 8 ohms. Should be more than adequate given the 104/2 is a sensitive speaker and removal of the passive crossovers is going to make it still easier to drive?

One piece of information which I don't have -- and perhaps someone here can help with is: what is the precise design of the original passive crossover in the 104/2? At what frequencies are they set? What type of filters are they? (e.g. Linkwitz-Riley, Butterworth etc)

Apart from getting an answer to this specific question, I will be most grateful for general comments and advice on the above.

Many thanks

 

MF 1000

Extreme Full Range 😱
Wammer Plus
Apr 5, 2011
6,172
4,732
193
Caego
AKA
Keith
HiFi Trade?
  1. No
I'm running the 104/2 with my transmission line sub/bass speakers so currently the bottom end of the 104/2 starts at 120hz run by a behringer a500 up to around 1200 hz after which the b110's take over up to around 3500 hz (fed by my Ming da 845s) after which the tweeters take over run by a cheap t amp

The std passive xover is quite a complex affair to reduce the load on the supplying amp 

 

edavey

Wammer
Wammer
Jul 12, 2010
196
3
0
Thanks Keith, that's an interesting configuration and I've noted your crossover points.

I'll want to handle all the frequencies within the 3-way Kef speakers.  

You don't happen to have any info on exactly what frequencies and filters are implemented in the original passive crossovers do you?

 

bobovox

Wammer Plus
Wammer Plus
Nov 21, 2010
1,283
1,325
193
HiFi Trade?
  1. No
A bit of info here:

http://www.kef.com/uploads/files/en/museum_pdf/80s/Reference_Series_Model_104_2_r.pdf

States 150Hz for the bass / mid crossover.

This is *claimed* to be the cross-over schematic here:

https://www.scribd.com/doc/266612771/KEF-104-2-Loudspeaker-Crossover-Circuit-Diagram

Does this look anything like the crossover in your speakers?

It's relatively complicated but the high pass on the tweeter looks to be 4th order. The low pass on the basses looks to be 2nd order electrically, with some LCR notch filters but I understand there will be a 2nd order roll-off due to the coupled cavity bass enclosure both at the top and bottom end - TBH I don't know enough to understand where that top end roll-off will start so the acoustic slope may or may not be higher than 2nd order.

You could measure the acoustic response of each set of drivers, with and without the crossover to see what the transfer function is. I would suggest using the passive crossover as a starting point rather than necessarily seeking to match it exactly.

 
  • Upvote
Reactions: Tony_J

Forum statistics

Threads
113,444
Messages
2,451,263
Members
70,783
Latest member
reg66

Latest Articles