AA Xover Modification

[Deang]

Too bad DJK bowed out. Together, the two of you fill gaps.

Good thread. I've enjoyed it.

[lynnm]

Early Impressions :-

As mentioned earlier I have moved to a straight Type A configuration and although my Klipschorns sounded very nice indeed prior to this mod I must say that they now sound much nicer overall.

The main change is in the upper midrange and high end which have become more open and spacious.

The soundstage does not seem to have altered.There is still a nice wide but rather shallow image which is hardly suprising given the necessity of the horns being snugged deeply into their corners. My KLF30's actually presented a more realistic stage than my horns.

I tried the speakers with the woofer inductor bypassed and although initially there was a rather nice additional warmth I did percieve an increase in distortion - particularly with saxophone and deep male voices. I have therefore reintroduced the inductor into the woofer circuit and a couple of days listening has convinced me that the system benefits from the woofer being limited to reproducing signals of 400hz and lower.

One unexpected benefit has been that the bass response is ( subjectively ) better since I went to the Type A. I say subjectively as I can see no valid reason that the changes made should have any impact on the bass response and yet the bass sounds better to me.

In sum then I find that with the xovers in the Type AA configuration the Khorns were exceptional and with the xovers configured in the Type A mode the horns are exquisite.

[Al Klappenberger]

Lynnm,

Your experience with bypassing the woofer inductor is interesting. Since the

Khorn woofer simply will not extend much beyond 400 Hz with or without the

inductor, I suspect the difference my be in the effect the woofer driver my

have on the rest of the speaker. The inductor not only keeps highs out of the

woofer it forces them to go to the high frequency drivers. You could say it

isolates the two. The difference you hear may be impedance variations in the

woofer showing up in the squawkers range. That's just a wild guess though. In

any case, I would suggest replacing that iron-core woofer inductor with a #14

AWG air-core inductor. That is, if you haven't already. The iron core

inductor is not linear. The air core inductor is. I think you will hear the

difference.

Al K.

[lynnm]

I was toying with winding my own once I can find an appropriate table of values Re: wire size, # of turns etc. Cost is an issue - I simply don't have a lot of do-re-mi for such adventures and such refinements as pure silver Litz wire etc. ain't gonna fly but I am open to suggestions regarding a low cost method.

Can you suggest a source for inexpensive inductors that would be appropriate substitutes for the factory supplied inductors

[JBryan]

Al K, etc.

Quick question -the Type A xover calls for a 2.5mH so I'm wondering what the effect would be (if any) of using one measuring 2.2mH. Does lowering the mH extend the cut-off of the woofer above 400hz? Also, what do the inductors used in the AK-2 xover measure? Thanks in advance for your help, Bryan

[JBryan]

Sorry - double post.

[John Warren]

c

Edited December 17, 2013 by John Warren

[Al Klappenberger]

Lynnm,

I wouldn't attempt to wind that woofer inductor yourself. It requires a fancy machine to do it as pretty as the ones you get from Solen. You will also find it difficult to get the inductance right. How much wire? How many turns? You really need to do some math to figure all that out. It is a h___ of a lot easier to just go buy it!

There is no reason to use Litz wire in a woofer inductor and silver is a total waste of money. Before going to silver, just go up to one thicker gauge of copper. It will sound the same and save you a pot load of money!

JBryan,

The value of the woover inductor is really not that critical. 2.2 mHy should be fine. Consider that Lynnm actually REMOVED it completely. That's ZERO mHy! How critical could it be!

John Warren,

I have been hoping somebody would post the impedance curve of the Khorn woofer. THANKS! It's interesting how it's a nominal 8 Ohms using a 4 Ohm driver. That's what I wanted to see confirmed. The same driver in the Belle Klipsch is 4 Ohms!

AL K.

P.S. I'm going to try to buy a #561 lamp today to get to the bottom of how that tweeter protection scheme either works or doesn't work.

[Al Klappenberger]

Guys,

I bought a #561 lamp at AutoZone. The price tag was a meir $1.99. The cold

resistance measures 1.27 Ohms. If you put that lamp directly into an 8 Ohms

system you have a very acceptable insertion loss of 1.28 dB.

Next, I connected the lamp in series with an 8 Ohm power resistor to a

(crummy TOA) PA amp. I did the test at 3000 Hz. I measured the voltage going

into the combination and then measured the voltage across the load resistor

using the AC voltmeter portion of a Hewlett Packard 334A distortion analyzer.

The table below starts at .1 Watt and increases by 2 dB steps to 10W on the

left (Voltage, power). The next column is the output level (Voltage, power).

The third column is the compression provided by the lamp at that STATIC

level.

0.894V, 0.100W - 0.79V, 0.078W - 1.08 dB

1.126V, 0.158W - 0.99V, 0.123W - 1.09 dB

1.418V, 0.251W - 1.24V, 0.192W - 1.16 dB

1.785V, 0.398W - 1.54V, 0.296W - 1.29 dB

2.247V, 0.631W - 1.90V, 0.451W - 1.46 dB

2.828V, 1.000W - 2.29V, 0.656W - 1.83 dB

3.561V, 1.585W - 2.72V, 0.925W - 2.34 dB

4.483V, 2.512W - 3.15V, 1.24W - 3.07 dB <-- Lamp just starts to glow

5.643V, 3.980W - 3.60V, 1.62W - 3.90 dB

7.104V, 6.309W - 4.15V, 2.15W - 4.68 dB

8.943V, 10.00W - 4.80V, 2.88W - 5.41 dB

I will leave it up to you to decide if this is serious protection for the

tweeter or not! I will say that I don't think it would cause any noticeable

degradation to the high frequency response of the speaker until you hit about

2 Watts input to the tweeter. At this level the midrange is probably driving

you out of the room on a Khorn! At that level, the 3 dB reduction in highs is

probably welcome even if it is theoretically a form of distortion.

I will try to do some transient tests next. Before doing the table, using a

dual-trace oscilloscope in place of the AC voltmeter to compare input and

output of the system, the lamp action seemed faster than I thought, but I

doubt if it would follow a transient. That remains to be proven though.

Al K.

[Al Klappenberger]

Guys,

Now that we have impedance plots posted of the K33 and the T-35, I figure I would mke one of the K55V too. This is the later K55V with the solder connections and the two-piece phase plug. The vertical axis says V/V 0-2, but it is actually 0-20 Ohms with 4 Ohms / division.

Al K.

[Al Klappenberger]

Guys,

I tried to set up a transient response test of the 561 light bulb protection

scheme that I could document with some sort of screen dump but I couldn't do

it. The only way to display the time it takes for the bulb to settle after an

abrupt change in level was using the dual-trace oscilloscope. I modulated a

3000 Hz sine wave from a function generator with a 0.8 Hz square wave from a

second function generator. The scope was triggered from the second function

generator and swept at .1 Sec / Cm. This showed that the bulb takes about .3

seconds to get to a point that is nearly fully settled. I have no way to

photograph the screen with the trace moving that slowly, so you will have to

take my word for it! This is WAY to slow to be of any use at all to protect a

tweeter from a transient.

If you look at the table of static levels I posted earlier you can see that a

change in level of 20 dB at the input to the bulb is SLOWLY transformed to a

change in level to the tweeter of 15 dB. I am not an expert on tweeters, but

I believe a 5 dB reduction out of 20 dB is NOT good for much. I still say the

way to protect your tweeter is to keep your lips off the volume control! If

you install a protection device that you believe works and count on it to do

so you are more likely to blow the tweeter than if you did without it and

kept the volume down. Time and experience will tell if more tweeters survive

with or without the light bulb scheme.

John Albright suggested that if you can reduce the average power going into

the tweeter that it will survive transients by itself. Maybe, if you can

reduce it enough, but I would really be afraid to count on it! Besides, what

is "enough"? I don't know.

So.. I my opinion (for what it's worth), I still say that the light bulb scheme is a BAD IDEA!

Al K.

[JBryan]

Thanks for the update Al - my you've been busy.

In regard to the inductor, I was really hoping to try out the Goertz 2.2mH and thought if anything, it would allow the woofer to go higher before cut-off. If the woofer naturally runs out at 400dB then I may just try wiring the woofer directly and see how that works out. Maybe some wool batting and/or bracing the cabinet will tighten up the bass enough to forgo the inductor altogether.

[Al Klappenberger]

Guys,

On second thought...

Another way to look at it is that with 10W input to the tweeter, it would see

only 2.55W (the bottom numbers on the table). 10W would probably kill a T-35

but 2.55W probably would not. The question would be for what time duration,

and would the lamp be fast enough to actually provide that level of

protection. I don't know! I guess it couldn't hurt unless you counted on it

too much. The bottom line is probably that you have to decided depending on

how you run your system.

Al K.

[Al Klappenberger]

Jbyran,

Don't wast your money on the foil inductor! They are of no value in a woofer circuit. They have their best performace WAY too high. It will work better than an iron-core inductor but they aren't worth the money. I don't even like them for tweeter filters. Use a simple 12 or 14 AWG solid wire inductor. The folds of a Khorn woofer limit it's upper range. The inductor change won't help its high end a bit.

Al K.

[John Warren]

----------------

On 10/20/2003 3:56:48 PM Al Klappenberger wrote:

The folds of a Khorn woofer limit it's upper range. The inductor change won't help its high end a bit.

Al K.

----------------

The mass cut-off of the woofer limits the high frequency output. According to Keele the mass cut-off is approximately equal to

(2 Fs)/Qes

which, for the K33E at Fs = 31 and Qes (according to Klipsch) = .633 making the mass cut-off about 180Hz.

The folds are a secondary affect.

Look at the plot of the impedance I provided. Note that the impedance begins to increase around 200Hz. With the inductor out, all that changes is the slope of the impedance rise at higher frequencies.

[Al Klappenberger]

John,

If the mass rolloff of the K33 driver alone is responsible for the upper limit of the Khorn woofer response at 400 Hz or so, how come the Belle Klipsch woofer goes very nicely to 700 Hz before it starts to poop out? It uses the same exact driver? I'm crossing my Belle woofer at 700Hz to a JBL driver. The response at 700 is fine.

Al K.

[Al Klappenberger]

Guys,

After some thinking about the tweeter protector problem, I propose a

combination of a ballast lamp and zener diodes as a virtually safe tweeter

protector. The zener diodes are 8.2V. That will allow up to 5W of continuous

power before they start to conduct and cause distortion. At that point, the

lamp should begin to heat up more quickly bringing the level to the diodes

down again. That should tend to reduce the distortion. The action of the lamp

should keep the average power under control with slow increases in level

before the diodes conduct. Any transients will be shorted out by the diodes

and absorbed by the lamp. The lamp has a resistance of at least 1 Ohm which

should be high enough to prevent harm to any power amp when the zeners

conduct and only cause about a 1 dB loss of highs at low level. The 75 Ohm 1W

resistor just brings the impedance of the entire network back to 8 Ohms and

is optional. The lamp alone offers virtually no protection at all to

transients. The zeners alone offer no protection against sustained high

average levels. I think this is a good compromise. I have NOT built or tested

this design and I have no intention to do so. I'm just offering it to anyone

who might be interested.

In case you want to use different diodes:

Maximum voltage = 8.2V + .6V = 8.8V peak times .707 = 6.22V RMS

Maximum clean power = 6.22V squared divided by 7.5 Ohms = 5.16W.

Al K.

[John Warren]

----------------

On 10/21/2003 5:33:30 PM Al Klappenberger wrote:

John,

If the mass rolloff of the K33 driver alone is responsible for the upper limit of the Khorn woofer response at 400 Hz or so, how come the Belle Klipsch woofer goes very nicely to 700 Hz before it starts to poop out? It uses the same exact driver? I'm crossing my Belle woofer at 700Hz to a JBL driver. The response at 700 is fine.

Al K.

----------------

The mass rolloff is responsible for the "horn loaded" output.

Below is the impedance of the K33E both in the horn without inductor (blue) and unbaffled (red). Around 250-300Hz the horn impedance is the same as the unbaffled driver. At 300Hz and higher the driver doesn't know that its in an enclosure, its unloaded. So above 300Hz (or so) I'm listening to a direct radiator that's simply firing into a series of baffles and reflectors. Not a desirable condition from a "purist" perspective. It is a secondary mechanism to achieve HF output from a horn. The LaScala and the Belle have less folding so yes, the high frequencies will get out easier but the fact of the matter is that the mass roll off is still down around 300Hz and both the LaScala and Belle are not horns above 300Hz (or so).

[Al Klappenberger]

John,

That reasoning for the high frequency limit sounds reasonable, but I don't

think it's right. Take a look at the article describing the new Jubilee bass

horn written by PWK and Roy Delgado that is attached. They clearly blame the

"complicated bending" as the reason for the upper limit.

Al K.

[John Warren]

----------------

On 10/21/2003 8:13:49 PM Al Klappenberger wrote:

John,

That reasoning for the high frequency limit sounds reasonable, but I don't

think it's right. Take a look at the article describing the new Jubilee bass

horn written by PWK and Roy Delgado that is attached. They clearly blame the

"complicated bending" as the reason for the upper limit.

Al K.

----------------

I think this is a case of "You can't have your cake and eat it too".

In a straight horn, the HF information above the mass roll-off, passes down the axis of the horn and into the listening space (no problem). In a folded horn anything above the mass rolloff bounces around in the structure and comes out as noise. This "noise" above the mass rolloff is what Delgado refers to as a "response". The Keele relationship is quite valid for the upper limit of the mass cutoff for a horn loaded driver (ref. "Low Frequency Horn Design Using the T/S Parameters" No. 1250, May 1977, AES Convention).

Edgar does not like folded horns and, I believe, it is because of this very problem. Above the mass rolloff it's really just noise. In his words, "folding does funny things to the sound".