DHAxovers for my '75 Khorns are done

[Kudret]

Here
is the picture of one of the DHAxovers I built this past weekend for my
'75 Khorns. Inductors and Bennic caps (as a temporary measure[])
are from solen.ca and 3636 from BEC. I did not put in the auto bulb and
the 65ohms resistor in the tweeter section as per Dean's recommendation
- thanks Dean for being so patient with me and answering my questions.

My
'75 Khorns originally came with AAs. I had first replaced the caps in
AAs, then converted them to type As. Vocals and everything else
sounded cleaner in type As and these were the ones I had in my Khorns
before putting in the DHAxovers.

I did not have much time to
A/B type As with DHAxovers yet but my first impression of DHAxovers is
positive. I noticed some improvement in bass and midrange
dynamics. They also seemed to widen the sweet-spot. Having said that, I
also like type As and, even though my audio memory is not very good, I
think I prefer the vocals on type As better. I wish there was an easier
way to swap crossovers to compare.

Time will tell which ones are eventually going to stay in my Khorns...

Kudret

[formica]

I honestly believe majority of people have limited audio memory (myself included) but the difference is some just don't realise it. That's one thing I find funny about some discussions on miniscule differences between certain tweeks when auditions were done months apart, in different rooms, using different source material... etc... []

How about A/B them by using Left and Right channels and a mono signal? Just a thought...
Rob

[tofu]

what was the total cost for parts?

[Frzninvt]

Trust me when I tell you there is no struggling or guessing that you are hearing a difference it is immediate and it is apparent! It will bring the speaker to a level of detail that you would not think possible. I have them in my '77 Khorn's and in my '78 La Scala center. The one in my center is built like Kudret's, without the bulb and resistor.

So much so that I pulled the beautiful DeanG Jensen PIO "AA" network in my center La Scala and put a DHA2Xover in it and I could tell the difference immediately, I did not to sit and play identical sonic passages to compare the "before" and "after" it was that obvious.

It will leave you with that "Whoa" look on your face.

[Kudret]

Charles, I was kind of disapointed that I did not have that "whao"

response. May be its due to the cheap caps I used in them. Did you try

Type As in your Khorns?

Kudret

[Deang]

How about A/B them by using Left and Right channels and a mono signal?

That's how I do it. I also have a couple of mono-recordings that I use. It's not very hard to hear the differences between networks.

[formica]

Charles, I was kind of disapointed that I did not have that "whao" response.

Do you guys think that maybe the "amount" of difference was due to the amplification... SS versus PP? That is... it was more noticable with solid state versus tube amp?

[Deang]

Charles doesn't impress easy -- he's spoiled rotten.

[Kudret]

Thanks for the recommendation to use mono recordings to test the networks.

I went through the pile of old records I have and found a mono record titled "The Singing Nun" that I used for testing. Not a scientific test but rather the testing involved walking back and forth between the speakers. What I hear is DHAxovers producing tighter bass and more refined treble. Vocals are also cleaner with DHAxover but perhaps a little dry. So, I'll give it to type As for the warmth of the female vocals.

I am sure both networks will benefit a lot from better caps, but they don't come cheap. I am planning to use better caps in the tweeter section first.

Kudret

[Kudret]

Tofu, the major components (including shipping and taxes) will cost about $250. EDIT: This is with Bennic caps.

Kudret

[Deang]

What are you using for tweeter caps? When you buy more caps, try a 2.5uF value. This will alleviate some energy from the lower part of the tweeter's output without adversely affecting the crossover point. 3.3uF is technically correct, but 2.5uF might sound better to you. This is one of the cool things about a first order filter. There is a lot play room, and just trying a few simple things can bring that thing right into perfect alignment with your ears.

[Kudret]

Dean, The picture I posted is with some 3 uF/150V ARPM caps. I prefer

these to the 3.3uF Bennic caps I tried first. I am really curious to

try some Thetas in that position. How much do they cost again?[]

Kudret

[Deang]

Not too awful price wise, but did you know Solen makes film and foils in small values? Every bit as good as Thetas for less.

[Kudret]

Actually I saw those caps but I didn't know if they were any good. I'll get some 2.7uF and try. Thanks, Dean.

kudret

[Deang]

Do you guys think that maybe the "amount" of difference was due to the amplification... SS versus PP? That is... it was more noticable with solid state versus tube amp?

There are several factors that determine to what extent the perceived difference will be when flipping networks. For example: Most of Charles' upgrades involved badly out spec Cornwall networks. Obviously he's going to hear a dramatic difference here. Also, I P-trapped his verticles which squelches the squawkers at 9kHz. Again -- clearly audible. His center LaScala had some old AA's too, so once more -- a pretty significant difference going to Jensen AA's. Then he drops those in favor of some decked out DHA's -- which use a first order tweeter filter similiar to the Type A. Anyone comparing the difference between the A and AA immediately discern the difference in the treble. Users of Bob's convertible network might chime in here. I've talked to several who were really surprised by the amount of difference between the two circuits. So, going from terribly out of spec filters to some that are back in line with where they should be is very audible. Ranking up with that is a change in the circuit: AA to A (or visa versa) -- either one of those to the ALK is pretty dramatic -- which leaves this DHA thing. Where does it fall?

It's best to think of the DHA as a constant impedance Type A. A solid state user may not hear a difference between it and a Type A at lower volume levels. However, When they turn it up -- they will definitely notice a difference. What does volume have to do with it? The Type A has no bandpass on the squawker, and as energy is dumped into the K-55 the throat of the K-400/401 becomes a choke-point. Rear reflected sound off the mouth of the horn traverses back down into the throat creating high orders of IMD -- which we hear as harshness. I just wrote about this on page one of this thread:

http://forums.klipsch.com/forums/1/602933/ShowPost.aspx#602933

You can also read comments by John Warren about it here:

http://home.comcast.net/~wooferboy/Page_3.html

In short, roll off the squawker. I admit the first order bandpass of the DHA doesn't do it fast enough or steep enough, especially since the natural roll-off of the squawker is closer to 18dB/octave. However, the electrical roll-off does eventually catch up and pass the acoustic roll-off -- and this at least puts a cut-off on the upper response of the squawker's output which does a nice job of cleaning it up. In my mind, it's towing the line between the ALK and the Type A. Because it's a first order network, it has a higher degree of immediacy than the ALK, and because it has a bandpass on the squawker, its output is cleaner than the Type A's at higher volume levels. So, this where solid state users get the most benefit.

So, why not just put a bandpass on the Type A? Sure, why not, and this is exactly what I did with the 300Hz networks I built for my Trachorns. It works, but it's a double edged sword, here's why: On any of the older stock Klipsch networks, the reflected impedance for the squawker is roughly 30 ohms, and this is where we get the 13uF value [159155 / (400Hz * 30 ohms)]. When impedance is high, the cap values are relatively small. However, the inverse applies with inductors -- they get bigger. The inductor for a low-pass section of a Type A bandpass is .82mH (159.155 * 30 ohms / 6000). Now, in the DHA (or the ALK), where the squawker's reflected impedance is about 7.5 ohms, you end up with a cap value about four times bigger, BUT an inductor four times smaller.

From the perspective of losses, there isn't all that much difference between a 13uF cap and a 40uF cap, but the loss factor between a .20mH and .80mH inductor is a boat-load. An .80mH inductor is like putting a .25 ohm resistor in series with the squawker. I know it doesn't sound like much -- but it "hears" like much! In the case of my networks, I've removed the inductor and I'm running the Trachorns full out until I can afford to buy the four 25uF Auricaps I need to build the networks right. So, a bandpass on the Type A makes the squawker sound a bit repressed and closed in to me. However, do it with the DHA and it sounds great. BTW, if you do a patent search on PK's patents, you will see that he was not completely opposed to the idea of a bandpass circuit.

Tube amps -- especially of the type that Kudret prefers will notice a difference regardless of volume. Low to zero feedback triodes will definitely track impedance. The last thing these amps need to see is a reactive load with impedance swinging all over the place. Incidently, Jeff Lessard is using a "Cornwall DHA" in his Cornwalls which he much prefers over the Type B's he built when he first got them. If you want flat response out of these type amps then constant impedance is the way to go. I think tube amps in general prefer resistive loads as opposed to reactive ones. On my system, the constant impedance filters always sound smoother (less peaky).

Tomorrow, when I have some time. I can shoot my mouth off about attenuation and some stuff detractors have to say about these filters -- if anyone is interested.

[BEC]

Dean,

You are quoting John Warren and using a swamping resistor?

Heresy!!!

http://home.comcast.net/~wooferboy/Page_2.html

Bob

[Kudret]

I've been listening to DHAxovers all night long, both with TT and CDP.

I played everything I play when I demo my system (Rob will know some of

them). I now realize that there is somethng very special about these

networks. They give me all the detail I was longing for. I used to lean

towards the speakers (especially with AAs, and to a lesser extend with

As) trying to hear more details. There is no need for that anymore. I

am hearing excellent bass and treble, vocals are not too bad either. I

guess these networks are growing on me.

I've done another thing to my SETH 2A3 amps. I finally put in the

ultrapath capacitors in them. This also must have contributed to the

overall sound I am hearing. I haven't been so happy with my system for

a long time. I still have a little way to go, a few tweaks to try but I

am almost there. Goodnight...

kudret

[Deang]

LOL, yes, it is Heresy. Let's see if I can drag the old wizard back in here.

According to Klipsch, the AA certainly wouldn't be the "best" network for anything using the K-55-M. In fact, they would tell you it's the wrong filter! At any rate, I think he's wrong -- everyone knows the Type A sounds better than the AA. The AA is for volume junkies.

I'm gonna post tomorrow and make you slug it out with Gil. He likes the resistor, as all sensible men do.

[formica]

Thanks Dean for that overview... I should really take

the time one night with a paper and pad to backward calculate what each

of these circuits are doing... but it would take me a while given my

BasicCircuits is rusty.

Let's see if I can drag the old wizard back in here.

Would make for interesting discussion...

[Kudret]

I've had alittle more time to listen to various CD s with the DHAxovers last week.

As I said earlier, I like the detail I am hearing with these

crossovers. Bass is also very tight. I feel that vocals are a bit on

the thin side and not as smooth as I would like them to be. Also, there

is all that detail but I am not hearing a full sound overall. I

would like to retain the details while smoothing and fattening a little

(if this is the right word) the overall sound.

For starters, I am getting some better quality 2.7uF caps

for the tweeter section (I am using 3uF at the moment). Second,

increase the attenuation from 3 to 4 or 5 dB in squawker.

Finally, get better caps for the squawker section. Am I on the right

track, or do I need to consider more substantial changes, such as Beyma

CP25 tweeters or tachorns, etc.?

Kudret