Crossover

[noodle_snacks]

I have a crossover design which i am fairly happy with in terms of sound etc, however the impedance graph seems to be all over the place with a rather large peak at approx 1.2khz (22ohms) then two dips, one at about 200hz (4ohms) and one at 3khz or so (3.25). Are these dips and peaks normal or should i be changing something in my crossover design?

http://img146.imageshack.us/my.php?i...003copy7ts.gif

[lndm]

The dips are like the 'normal' level. The 2k peak would be your crossover. The rise at the lower end, your bass resonance and the rise at the top, your tweeters vc inductance.

This seems typical, no need to do anything.

IMO, taming these peaks by compensating the crossover improves sound, but how much depends greatly on your amplifier.

[noodle_snacks]

Out of curiousity how does one tame these peaks in the crossover?

How is it that a 3000hz the tweeter measures 5.39 ohms by itself, but 3.5 ohms after the crossover? Is this to do with the way i have padded the tweeter? Should i adjust my lpad values to compensate for this (i used a crossover optimiser which probably affected it)

[lndm]

If you have components in parallel with the tweeter (I guess you have at least an inductor), then you may find low impedances in places. If it seems extreme or you don't like it, you may have to force the optimiser a little.

A speaker systems impedance is complex. (Complex is a technical term meaning it acts like capacitors and inductors). It is also a complicated mix of driver and crossover interaction, but this doesn't matter when you have an impedance plot showing the end result like you do.

Concentrate first on levelling the crossover peak. Use your optimiser if possible. Create a series RLC circuit and put it across your speaker (in parallel with your amp). The LC should resonate at 1.2k. The resistor affects the amount of levelling, and the actual values of the C and L affect the width of the levelling effect (the Q).

If done right, you should be able to get it level to within half an ohm which is quite good.

When you first throw some values onto the crossover sim, use a resistor value of only 0.1 ohms as this makes it easy to see the effect of the filter. When you get the frequency right, ease up the resistor until the level is right. If the filter seems to be affecting too wide or too narrow an area, try doubling the value of one component and halving the other to see the effect (leave the resistor alone in this step).

Done well, this can relieve a little midrange harshness in some cases.

[noodle_snacks]

i have tried using a RLC filter, i seem to get a result that gives me an impedance of about 3.6 ohms flat in the woofer region, plus the tweeter dip then rise, to me that seems a bit on the low side overall and considering 60uf capacitors don't come cheap i will probably pass on one of those for this design.

However after some tweaking i did find that i could get a better tweeter impedance by using a 8.3uf capacitor and 0.33mH inductor. Since frequency response is pretty close i will try that change and see what it does to the sound out of interest.