XSIM scheme + filter response of KEF Reference Model Three

Hello Gents,

Kinda started with crossovers about a week ago so pretty noob to begin with. Studied Allens famous article of course and some online tutorials. I know what capacitors, Inductors and resistors do and understand the math that comes with it but I'm fairly new to audio in general.
Picked up a couple of Kef Reference Model Three's recently and although they sound majestic to my ears I also use them to learn more about crossovers. I've got the actual scheme + vlaues from this forum and started playing with XSIM. I don't have the specific drivers specifications. The only thing I have is:

Frequency Response: 40Hz to 20kHz
Recommended Amplifier: 50 to 300W
Crossover Frequency: 140, 400, 3000Hz
Impedance: 4Ω
Sensitivity: 91dB
Bass: 2 x 200mm B200 pulp cone
Midbass: 1 x 160mm B160 polypropylene cone
Midrange: 1 x 160mm MB160 polypropylene cone
Tweeter: 1 x 25mm NT25 soft dome

Here is the basic scheme and the filter responses. Now the question is whether or not there's room for improvement or does that all come to presonal preferences mainly?
Note: I'm going to replace the bipolar caps and resistor from Bennic with better ones so it might be a good idea to each some values here and there as well.

Comments much appreciated...

regards

Willem

filter-response.png
scheme.PNG
 
From the look of the circuit diagram & that you've only posted electrical transfer functions, I assume you've got static loads (resistors) in there representing nominal driver impedances. The filter will account for the actual on-baffle / in box driver frequency & impedance loads, & without those it's impossible to say whether or not there is any 'room for improvement', so I'd probably leave well-alone as far as that goes, until you're able to measure them accurately -which takes a while to learn, so top tip is don't rush that particular process. 😉 It can throw up unexpected issues even for those with years of experience & most of us will have spent a day or so chasing out tails / pulling our hair out trying to find something only to discover our hardware has decided to start acting the goat (often aka 'we weren't concentrating quite as we should & made a complete ****-up'). As a general note though, knowing KEF, the ESR of the electrolytic caps is likely accounted for in the filter transfer functions, so changing them for MKP types may throw those out from what was designed, in which case you'll either need to add whatever the equivalent series R was for those caps, or replace like-for-like to avoid any chance of such issues.
 
Last edited:
Thx Scottmoose,.....I kinda suspected such an answer. Yes,...I'm in the learning curve, I don't have any specs of the drivers and I've never done an audio measurement just yet🙂.....I only have this effectively..

1 x NT25 tweeter (SP1353)
1 x B160 midrange unit (SP1345)
1 x B160 lower midrange unit (SP1348)
2 x B200 bass unit (SP1343)

I've done some amplifier refurbishments and I can measure the individual components so a like for like replacement including some required changes should be possible I think.

reg

willem
 
I've done some amplifier refurbishments
Then you know that swapping out components is sometimes a bad idea?

The chances are that KEF have done a good enough job of making sure each way is doing it's best. Improvements there will be difficult so it's best to leave those in this case until you feel you can do better.

Until then, you might try experimenting with an equaliser, room placement and room treatments.
 
  • Like
Reactions: witwald
I guess that the next step might be to measure the impedance responses of all the individual drivers while they are mounted in the enclosure. That will get you well on the way to being able to perform further simulations of the electrical responses of the driver filter networks. It would certainly be very interesting and informative to see your results. 🙂
 
  • Thank You
Reactions: AllenB
Yeah thats the plan but I need to learn / know how to measure this exactly. As mentioned I've nver done that but I'm eager to learn. My basic setup is the mentioned speakers driven by a refurbished H/K PM645VXi which gets fed by Volumio on a RPi with Qobuz as source. Thats the starting point I'm at.
I probably need a good mic and some software and I have to take the room into account as well....again I'm a beginner when it coes down to doing such measurements🙂🙂

w
 
  • Like
Reactions: witwald
In addition: If and only if I get the measurments right...could I then convert the measured output into FRD / ZMA files and import them into XSIM allowing me to tweak xover?...or would that not be the way forward...Apoligies a for maybe asking too much but it is exciting stuff.....well for me it is🙂

w
 
Short answer is 'yes' (or an alternative crossover design package of your choice). In passive filter design, we work with the on-baffle / in-box frequency, phase & impedance responses rather than flat-line resistive electrical loads. However, the latter are fine for getting a basic handle on what does what -just remember it won't directly translate over once you've got the actual frequency-variant responses in place.