• Please note that there are a few updates and clarifications made in the Audiokarma Rules, mostly relating to advertising and the addition of the new "Paying it Forward" & "Giving back" forums in the AudioKarma Audio Marketplace section.

Al K - Questions regarding ak-3 and AP12

M

meagain

Member
Hi Al, I was over at your website and am confused about the ap12-ak3 vs. 350. I'm not grasping what this page is telling me, what these are exactly, or what this ohm/impedance across woofer means.

"These are generally based on the Klipsch AK-3 network woofer filter which is quite good but is missing the components that equalize the "reactance slope" required to provide a constant impedance resistive load to the amplifier."

I have an ak-3 and wondering what all this means. I think I like this ak-3 or the general concept about it as I'm hearing more detail across the board and there's something 'cleaner' about it. I was wondering where it fell in comparison with your stuff or if there's any similarities between at least a piece of it vs. yours at all. Mostly, I seem to have a need to set the volume higher on this (or perhaps want to or 'can'?). I'm curious about this and I wondered if it had anything to do with either the above quoted paragraph, or ??? perhaps just because it has more 'parts' than I'm used to and it's sucking up energy. :)

Anyway, I was just wondering if you could explain it a bit (in lay terms). Thanks!
 
Register to hide this ad
MEAGAIN,

The idea of the AP12-AK3 is to utilize the parts of the AK-3 network that are already in side the woofer hatch of a Khorn that came equiped with the AK-3 network. As it happened, Klipsch lucked out with that network and actually got the woofer filter correct, or at least half of it anyhow! What the AP12-AK3 network does is utilize the part of the AK-3 that is correct and provides a correction to the other half that is wrong. This is, the squawker lower limit filter that is the wrong impedance. The Ap12-350 simply provides the parts that are inside the woofer hatch of the AK-3 that are NOT provided with older (and newer) Khorn networks. The AP12-AK3 is simply a way for AK-3 Khorn owners to save a few bucks and the hassle of opening the woofer hatch.

Richard: How's that for simple-minded English?

Al k.
 
So the AP12-AK3 uses the goodness of the woofer x-over but fixes the bad part and that bad part is a mismanagement/wrong impedance. The AP12-350.... gah! Let me just grasp the AK3 for now :). The AP12-AK3 makes for a constant impedance of 6 ohm.

If the stock AK-3's impedance is wrong, how is it wrong? Is it a fluctuating impedance situation or is it... 8 ohms or something? I'm wondering what negative effect this has and does this translate to something 'audible'? I'm not sure how it would benefit me in real life. Can one use these as a stand-alone device or must it pair with the top part of an ES network?
 
Meagain,

I think you got the idea. Bacially, A crossover is a combination of a highpass filter and a lowpass filter. These are mirror images of each other with respect to their nominal impedance and impedance slope with frequency. The AK-3 has a 6 Ohm lowpass grafted to a 30 Ohm highpass. They are also unequal orders. The lowpass is 3rd (the woofer voice coil inductance counts as one) and the squawker filter is 2nd. This will only work correctly with computer optimization. I have provided a 6 Ohm highpass that is also 2nd order but is optimized to yield a constant 6 Ohms. This is considered "correct". Doing it wrong can have numerous advers consiquences. What they are depends on your amp and other factors. My position on this is simple. Why define what goes wrong when it's done porely. Just do it right and all the basses are covered! There is simply no good reason to do it wrong unless you just don't know how to do it right.

Al K.
 
Interesting... Do all the standard heritage type crossovers do this as well? Or worse? Like the AA, A, etc? And... I have Craigs VRDs set at 8 ohm... What would this mean for me having this stock ak-3 doing this impedance dance?

There's something about these AK-3s that I like over say, an AA. Caps aside.... Mostly, they are dealing with a long-standing issue I've had at the 6k area. They are doing something 'right' there for me over other networks and - would you be able to tell me why? Cuz I need to latch on to whatever that is. From what I can see, these frequencies are being managed differently. In my favor.

Coytee! Guess what I got today? :)
 
Meagain,

Yes, all the Heritage sereis networks are similar. All go up to A LEAST 30 Ohms in the midrange. The Cornwall and Heresy go much higher. This was done by PWK to reduce the value of the cap needed to set the squawkers lower frequency limit. In has day, large value caps required at an 8 ohm level were not available. That is no longer the case, so there is no reason to let the impedance go up so high. As to setting the amp at 8 Ohms, Yes, why bother when it not only seeing 30 Ohms in the midrange but a reactive load as well! The speaker looks like a capacitor at some frequencys and like an inductor at others. A correctly designed crossover looks like a resistor of some constant value at all frequencies. That value is dictated by the resistive part of the woofers' impedance. In the case of the K33, that's about 6 Ohms. Look at the specks on you amp as see what it tells you about power output and general performance into 30 Ohms. You won't find much! No amp wants to see 30 Ohms, especially not a tube amp and certainly not a one-lunger (SET) amp!

The AK-3 does have on good thing going for it besides a smooth woofer filter and that's an extreme-slope tweeter filter. Again though, it's only half there! The squawker upper limit is determines COMPLETELY by the K55 squawker pooping out at about 6 KHz. It is not filtered AT ALL! That means the K55 and the k77 tweeter are connected together from 6 KHz up! The mirror image filter to work with the tweeter filter is not there at all! The filter that is there though is very similar to half my ES5800 network.

Al k.
 
Interesting. About this impedance/ohm issue.... Is this something audible or just technically correct on paper? Am I hearing negative effects of this with an AA, A, AK-3 without realizing it (I've never heard your networks).... or no?

The AK-3 DOES have an extreme-slope tweeter filter. Interesting. But the AK-3's version allows the squawk to run off naturally instead of being cut off?
Like, basically you're washing a car whindshield clean, but forget to rinse?
Does the AK-3 at least clean up a little better up there over the say... AA? Or at least removes a few more mosquitoes? Does this ak-3 tweeter filter or 1/2 an ES tweeter filter have any effect on the squawker end or does it merely come in differently and not overlap so much?
 
Meagain,

Here's a set of computer generated plots on the AK-3.

All of the Klipsch heritage sereis networks look similar to this including the AA. The 27 Ohm peak at the center of the midrange is in effect "unloading" the amp to some degree. Would you run a tube amp with no load on it? It would not be advisable. Any amp will operate better when loaded with the impedance it was designed for. I can't say exactly what this will sound like, but it centainly won't be a good thing.

The lack of the mirror image filter on the squawker high limit, besides providing a bad load to the amp, allows it to pass a lot of higher frequencys that should be handeled only by the tweeter. The natural rolloff of the K55 is by no means extreme-slope. It's very gradual compared to the slope of the tweeter filter. This will casue a rough frequency reponse in the tweeters range caused by comb filtering due to interference between the two sounds where both the squawker and tweeter are making the same sound. It also messes up the stereo image. Multiple sources of the same sound, especially when they are not in time with each other because of propogation time differences will confuse the brains ability to determin the location of a sound. It also causes a smearing becasue you are hearing everything TWICE. That is becasue you hear the sound from the tweeter first, then again from the squawker becasue it's farther away from you by a foot or so. All this adds up to a much better sound from a correctly designed extreme-slope network. You WILL hear a difference!

BTW: The tweeter filter in the AA is nothing more than a jury rigged 3rd order filter expanded from the 1st order filter of the "A" network. It has a very rough response becasue it is symmetrical. It was designed using very outmoded "Constant K" equations. These equations are not appropriate for crossovers and don't work well even for normal "stand-alone" filters. The tweeter filter in all the later networks from the AK, AB and AL and newer were correctly designed, probably by cut-and-try methods. I say this becasue computer design of the tweeter filter in the AK networks requires careful choice of passband ripple to allow a particular part to be eliminated by forcing it to go through zero to a neative number. At the point of transtion from negative to positive it can be eliminated. I seriously doubt Klipsh has the software necessary to do this. The latest ones (xx-4) are again poorly designed. The attempt at a true crossover at 4500 Hz not depending on the K55s rolloff was beyond them!

Al K.
 

Attachments

  • Ak3.gif
    Ak3.gif
    44 KB · Views: 36
Hi Al,
Can you please give the cap value and tap for autoformer that would correct the Forte crossover.
Also, get off your duff and do a complete Forte crossover...Just kidding... No really, please.
 
Alteklipsch,

It will take more than a cap and transformer change to correct the Forte crossover. The schematic I have is for the Foret II. It looks similar to the AA but with an extra cap across the woofer. It would take a redesign from the ground up. I would actually need a Forte sitting here to use for tests to actually do that.

Al k.
 
Al

Here is the schematic for the Forte, just a little different than the Forte II. They are both awesome speakers and worthy of one of your designs. I wish I could send you a Forte for the measurements needed to do a proper crossover.
 
Alteklipsch,

I'll add that schematic to my collection..

That network is typical Klipsch:

* Doubly terminated tweeter filter when it should be singly terminated.
* Squawker lower limit filter impedance way to high.
* No squawker upper frequency limit at all.
* Parallel connected highpass filters (a total no-no).

Al k.
 
I don't really see a problem with relying on mass rolloff for the squawker. I have no low pass on my 2470's with your Trachorns and it sounds outstanding.

I don't remember the problem with parallel connected high pass filters, but it's fairly common in audio, so it's not like Klipsch is only company doing it that way.

The impedance issue is complex (ha ha). Since current goes down as impedance goes up it doesn't mean there will be an increase in acoustic output at those frequencies.

Something that's becoming increasingly popular is raising the impedance of a speaker. http://www.zeroimpedance.com/TechnicalInfo.html Of course, "crappy" Klipsch networks already have this feature built into them.:D
 
Dean,

The problem with natural roll off of the squawker is that it is unpredictable and very slow. It's only done to save parts!

Two highpass filters in parallel is simly poor design. The impedance simply can't be constant that way unless one high pass zo is extremely high. That what the Klipsch networks do, but it's not high enough to not cause interferance. It's simply a major corner-cut! Audio lets you get away with doing a lot of stuff that's not correct. That doesn't make it right. I just makes it cheap. I build upgrades. They wouldn't be upgrades if I cut corners.

Al k.
 
I half expected a klappenrant after that last post. :)

Meagain (Lisa) likes the sound she's getting from her tweeters with the AK-3 filter section. She's asked if there is anything I can do with her Jr's to get a similar response. If I modify the network and change the squawker low pass to that of the Universal Type A -- she can run the ES5800's, right?
 
I have offered before that if someone in the southeast wants to lend me their Forte for one hour....I can get the info AL needs to make a Forte network.
 
You must log in or register to reply here.
Back
Top Bottom