|
Register | FAQ | Members List | Calendar | Gallery | Today's Posts | Mark Forums Read | Search |
Loudspeakers Your DIY Speaker designs |
|
Thread Tools | Display Modes |
#1
|
|||
|
|||
wad kls3 new xover
I have acquired a pair of kls3s and done a theorectical xover for it, not implemented yet, but will do.
Its not a textbook formula, but properly implemented based upon the slopes of the drive units. bass uses 4.7mH and 27uf capacitor mid has 1.5mh and 12uf in series, then 4.7mh and 8u2 shunt treble is 3r3 resistor, 3u9 capacitor and 0.3mh the original used 3rd order butterworth positive phase target slopes with the top around 2700hz. mine is using a linkwitz riley to give perfect phase implementation, with asymmetrical slopes on the mid, the mid was very difficult to get right. I went through a good few iterations, many of which more complex, this was the simplest and best that worked out. please feel free to experiment and use, I would be flattered if someone tried it and liked it. this is the best xover I can currently do, after days hard work and its very flat frequency, impedance is less smooth than before, but still a kind load, frequenices are 400hz and 4khz, it may even be worth taking the top upto 6khz, but that'll mean a reahash and more theoretical work bass polarity is again inverted, as per the original, rest are regular polarity if you have changed your tweeter, or are thinking of doing, I can do a xover for you free as a starting point, you may then wish to commission peter if you think improvements may be had. its important to note its not based upon real drivers in cabinets, but theory, however, I did the original as done by noel/dominic baker/audax and it comes up absolutely identical, so fair to say its at least as reliable. commercially as well, most mass runs aren't measured individually and a different xover done for each drive unit, they just measure once. provided manufacturers data is correct it will be very close And if anyone is doing a speaker with just drive units, I can do a xover for them, no charge, as long as I can get data, it will be at least 95% correct finally, it may be worth commissioning peter as a group project and sharing costs for kls3? I see precisely why peter dislikes 3 ways, they are fiendish, due to midrange interaction between components and other things, too. attachment drawn on linear technologies free switchercad Last edited by Ianm2; 22nd May 2007 at 05:40 PM. |