K22 impedance, Heresy crossover frequencies and other technical questions...

[folkdeath95]

Hello,

Does anybody has ever seen the K22 impedance curve?

The Type E crossover frequency for the woofer is said to be 700Hz and, so it seems that the 2.5mH self has been calculated for a 11 Ohms impedance. But I mesured my two K22 with Limp and the results are a little disturbing.Here is the curve of one of them (the second one is very close):

At 700Hz, the impedance is close to 17 Ohms, so it should need a 3.9mH self...

Can the age of my speakers (30 years) explain such a difference with the theorical impedance value? I've looked for a K22 datasheet with an impedance curve, but I didn't manage to find one....

Does anybody has any clues?

Thanks a lot.

[Guest " "]

you just stumbled across one of the little know facts about klipsch heritage xovers. now think about it. if you wanted to seperate an audio band into low, med and high, with out the use of electronic componets, what would you do.

[folkdeath95]

you just stumbled across one of the little know facts about klipsch heritage xovers.

Excuse my ignorance: which fact are you talking about?

[Guest " "]

ok, the parts used in early heritage crossovers for the most part consists of the same componets. AA, E2, B1, etc, all have inductors and capacitors that are the same. Some rewiring was done and some use of different taps settings was done to make these xovers approximately fit the requirement, but in no case, was it an exact match. This worked out because the horns participated in the bandwidth split thru their hi-pass behavior. So sure, depending on the crossover model you want to use, the crossover topology if done using modern methods would look totally different , and they do. The current xovers are computer modeled and look more like what you would expect thru mathematical calculations.

[folkdeath95]

I didn't know that. I must confess I only studied the Type E crossover as my Heresy are my first Klipsch speakers. So I didn't point the similarities in crossover components. That's a very interesting information. (It could explain why so many KHorn users tweak their crossovers or the crossover upgrades like Al K's offers..) Maybe a little work on the Type E could give good results too...

[WMcD]

Congrats on measuring impedance. It would be nice to know if you did that with the driver in the box or out in free space.

The curve looks good to me. The mountain is the resonance. Your phase measurement shows zero phase at the peak of the mountain, which is normal.

One rule of thumb is to look at were the impedance magnitude hits a low in the freq region somewhat above the peak. That is where the capacitive and inductive effects balance out. Therefore it shows voice coil resistive impedance of about 10 ohms. Again, the phase is zero, on your plot, which makes sense.

Going farther to the right, i.e. higher frequencies, we see the effect of the inductance of the voice coil as it is sitting in the iron magnet. We know it is an inductance or +j because the impedance is rising. The phase is going positive, which shows we're in the +j range.

This inductance in the voice coil acts as built in 1st order crossover. This is true in any voice coil type driver. It is usually equivalent to a 1 mH inductor for large woofers -- though this depends on the driver..

Your observation may be that the crossover inductor seems too small to work at the crossover freq. The reason, probably, is that the voice coil already has 1 mH worth of inductance.

Just FYI. You see tables and calculators for the value of an inductor or even caps for a first order crossover. I observe the following: The calcuated impedance of the inductor or the cap at the cross over freq is equal to the voice coil impedance. X(l) = 2piFL or X©=1/2piFC. IN a second order (Butterworth filter?) the impedance values of cap and inductors are 0.707 of the voice coil impedance -- though the voice coil inductance has to be considered, as before

Early on, PWK became convinced that first order crossovers were more euphonic. But apparently he stepped back from that in later designs.

WMcD

[djk]

" It would be nice to know if you did that with the driver in the box or
out in free space."

Looks like it is in the box.

[folkdeath95]

Looks like it is in the box.

Yes, you're right, the measurement was made with the driver in the box (only the crossover was missing).

[folkdeath95]

This inductance in the voice coil acts as built in 1st order crossover. This is true in any voice coil type driver. It is usually equivalent to a 1 mH inductor for large woofers -- though this depends on the driver..

Your observation may be that the crossover inductor seems too small to work at the crossover freq. The reason, probably, is that the voice coil already has 1 mH worth of inductance.

WMcD

I've never heard of that before. From what I read, I thought the voice coil inductance wasn't included in the crossover calculation. On every source, the formula was: L=Z/(2*PI*Fc). From what you say, it should be read as L=(Z/(2*PI*Fc))-Le, right?

[djk]

Networks that are calculated with simple computer models seldom sound good.

Crossover design is as much an art as it is a science.

[folkdeath95]

Networks that are calculated with simple computer models seldom sound good.

Crossover design is as much an art as it is a science.

I totally agree with you. But the formulas are good starting points, I think. Every art has its basics.[]

[mikebse2a3]

Remember the impedance is just a measurement of one aspect of a speaker's design and can't begin to fully describe how a loudspeaker will sound. Ultimately it's the acoustical crossover between drivers and driver/horns that is the goal which means text book calculations don't always work well with real world loudspeakers and the complex combination of imperfections and consistency of parameters between different sections making up the total system.

It's important to remember PWK called these balancing networks because they did more than just act as a crossover between drivers.

Because the polar response of the various drivers and driver/horns vary with frequency the designer/engineer must make some assumption about the room's acoustics, speaker location and listener location when he sets about to voice(ie: tonally balance) the loudspeaker system. This is easly verified by anyone by just changing the loudspeaker's direction and/or location in a typical listening room to hear the shift in tonal balance of the system.

Another real ear opener is to adjust the frequency(tonal) balance of a loudspeaker system in an anechoic or outdoor enviroment and then take this loudspeaker into a typical listening room and observe the perceived shift in frequency(tonal) balance of the loudspeaker and then observe what adjustments are required to bring the loudspeaker back into an acceptable frequency(tonal) balance in the listening room. This is a very good lesson in how powerfull the effects of how a loudspeaker's polar response variations and our room acoustics play in what the listener will ultimately experience.

It's easy to hear a change in tonal balance of even just .5db(especially in the 500Hz to 4kHz range) over 1 or 2 octaves as a positive or negative improvement. The problem (and challenge) is again these tonal shifts are again often a complex combination of the enviroment's acoustical properties and how the listener and loudspeaker are integrated into this unique enviroment.

mike tn

[WMcD]

I see that I did not address the basic question. But you may have inferred it.

That is a very typical impedance curve for a driver. It does not indicate aging. Curves for new ones look the same.

I was being a bit of a smarty pants in describing it It also reflects the driver being in a box. When you put a driver in a box the box adds a spring effect as the resonant frequency goes up. That is the value of freq at the big peak. The free air resonance of the driver Fs will be lower.

Naturally I agree with Mike SE.

This is getting into the overall problem of describing the "systems" involved. You start with the crossover filter, the driver electrical characteristics, add the box acoustics, add the actual radiation of sound from the front of the box and driver, and then add the effect of the room as the speaker box is placed. Whooo, a lot of stuff.

It is good to know the acoustic output of the speaker in an anechoic (no echo) chamber or outdoors (no echo) because it allows one to examine it without room effects. Room effects can mask a problem you want to look at, which are inherent to the box/speaker.

OTOH,that response does not tell us very much response in a room. Then you try things in a room and location. All well, but it is only "that" room and location. So what is the true measurement? - - - -

This is getting far afield. I'm reminded of the story where a tyke asks his father,: "Were did I come from?" Dad is uncomfortable but goes into the story about how, "Mommy and Daddy loved each other very much and . . . .etc." At the end, the tyke says "I don't really understand all that. My friend at school says he came from Cleavland,and I was just wondering ." - - - - Use the value used by Klipsch and don't worry.

WMcD

[WMcD]

Last night I was using Chrome and the CRs didn't seem to take.

I fixed the CRs with IE tonight.

WMcD

[folkdeath95]

Thank you everybody for your very comprehensive answers.