Impedance curve with AA xover

[UncleRobb]

Hi,

I'm the proud owner of a pair of old La Scalas with AA xover. Does anybody have the impedance curve (impedance vs. frequency)? I would also appreciate the schematics. Btw, is the AA one of the best xovers that has been designed for the Heritage series?

Thanks, Robert

[JohnA]

try this!

The Type AA is my favorite and is one of the best, if not the best. It really shines if rebuilt with premium components.

I do not have a graph, but the impedance curve runs about 6 ohms up to 400 Hz, except for the bass resonance. At 400 it rises to a little over 32 ohms and drops to near 8 ohms above 6000 Hz.

[UncleRobb]

John,

thanks a lot for your post that I found very helpful.

I did not expect the impedance to vary that much with the frequency. Is this normal for 3-way systems? A buddy of mine believes indeed that my La Scalas pose a difficult load to my flea powered (~1.5 W) tube amp (Decware SE84C). Any opinions on that?

Thanks, Robert

[JohnA]

Most speakers do not have such an impedance curve. I believe it will indeed be difficult for your power amp, but you'll have to try and see.

If you want a crossover with a flat impedance curve look at the ALK crossovers

http://www.alkeng.com/klipsch.html

or this one: http://forums.klipsch.com/idealbb/files/DHA2.jpg

[Al Klappenberger]

Rob,

Here is the computer simulation of the AA network input impedance. The entire speaker curve will look very similar. The big peak in the middle of the squawkers range is the autoformer in action.

The top line of the curve is 35 Ohms and each division is 5 Ohms.

AL K.

[UncleRobb]

John, Al,

thanks a lot for your posts. So if such an impedance curve is rather unusual and if the La Scalas/AA xcover are a difficult load for the amp, why is then the AA one of the best xovers?

What property must a good amp have in order to drive a speaker with such an impedance curve? Should the output power be independent of the impedance? The Decware SE84C delivers slightly more power as the impedance drops. Does that help? In any case, to me the setup sounds just fine.

To my understanding, the squawker is more efficient than the tweeter and bass horns and thus it has to be attenuated. Would a resistor network not serve this purpose better than an autotransformer?

Question over questions....

Thanks, Robert

[Al Klappenberger]

Rob,

If a pair of resistors costing a total of about $1 was as good as a $30 transformer, PWK would have used the resistors in a minute! The transformer sounds better. The reason is not clear, but my personal theory is that the transformer does not isolate the driver from the amps damping factor like a resistor network does.

Al K.

[sfogg]

Al,

"Here is the computer simulation of the AA network input impedance. "

Do you have one for your network?

BTW, just waiting on the transformers and I'll be starting to build five of them. All the other parts are ready and waiting to go.

I'm doing it slightly differently though.

First I'm making the boards a little longer and wider so I can put the terminal strip on the short side so when it is installed in a La Scala the terminals will be facing the opening on the speaker. The little bit extra size means I can move the coils a little further apart while still keeping the same orientation.

Second, I'm going to add the poly switch tweeter protection and maybe make that switchable in and out of the circuit. At least on one so I can easily hear how much of an audible effect that will have.

Third, I may wire the autoformer on one up to a 2 pole rotary selector switch so I can quickly switch between taps to see how I prefer the midrange setting. The poly switch has about 1/2 ohm of resistance so that is going to drop the tweeter level a little, so dropping the midrange slightly more might balance that out better. With your swamping resistor on the squaker crossover do you think it would be a problem using a break before make rotary switch to change taps while the amp is on/playing?

BTW, thanks again for sharing the plans for this for us DIYers.

Shawn

[JohnA]

Robert,

Though I'm getting to the limit of my understanding, SET amps in particular, and ones without feedback, have high output impedance. This makes their frequency response track the impedance curve of the speaker. The cure is lowering the output impedance. At 1.5 Watts, you cannot afford more feedback or other mods that will reduce the amps output.

The La Scala/Belle/K-horn impedance curves, while odd are not that extreme. Some panel, ribbon or electrostatic speakers have loads as low as 1 ohm that is mostly capacitive. That a terrible load for an amp.

Mr. Paul always said the K/L/B impedance curve had never been a problem. I'd say it was not a bad enough problem to injure the sound of the speaker.

[Al Klappenberger]

Shawn,

Do I have an impedance plot of my network? Man.. I thought you'd never ask!

The top line is 9 Ohms. 1 Ohm per division. The reduced impdeance at 20 Hz is because a 4 Ohm woofer driver is assumed. The same 4 Ohm load is assumed in the plot of the AA I posted earlier. The production units actually measure 8 Ohams +-1 Ohm with a phase angle of less than +-15 Degrees form pure resistive impedance (that's "polar" impedance).

The modes you propose sound ok to me. Fliping the proposed squawker level switch with the amp on shouldn't casueany problems.

Al K

[sfogg]

" The production units actually measure 8 Ohams +-1 Ohm with a phase angle of less than +-15 Degrees form pure resistive impedance (that's "polar" impedance)."

That is really flat.

Have you ever modeled the difference in FR a user with a highish output impedance amp (tubes) would see with the stock crossover compared against yours? Yours would obviously be flatter through the 1-6k range.

"The modes you propose sound ok to me."

Thanks, glad to hear it.

"Fliping the proposed squawker level switch with the amp on shouldn't casueany problems."

Cool, that should make it easier to find the setting I like. I don't think I want to wire up five of those switches.

Shawn

[Al Klappenberger]

Shawn,

No, every amp is different. Some don't care what load you connect to them. Others do. I am not an amp expert and I don't profess to know what the total effect of strange loads might be on SET amps. I do beleive that a flat 8 Ohm resistive load is what amps are tested with and would like to see. A load of 30 Ohms can't be a good thing though. I suspect the frequency response is not the factor that would be upset by a strange load. My guess is that it would more likely lead to stability problems. Vacuum tube amps don't like to operate into an open circuit. That could cause extreme voltage to develop at the plates becasue of inductive kick in the output transformer primary (like a automobile ignition spark coil). I figure 30 Ohms is on the way toward an open circuit. Again, a big deal to some amps, and not so big a deal to others.

Al K.

[sfogg]

Al,

"I suspect the frequency response is not the factor that would be upset by a strange load."

Actually it is exactly what gets effected if the speakers impedance varies by frequency and the amp itself has a highish output impedance.

See:

http://www.transcendentsound.com/amplifier_output_impedance.htm

and

http://www.stereophile.com/showarchives.cgi?810

"No, every amp is different. Some don't care what load you connect to them. Others do"

Exactly, and part of what that is from is the output impedance of the amp. E. Brad Meyer and someone else (can't recall the name) wrote an article for Audio magazine explaining this in the late '80s or early '90s. One of the tests they did was double blind ABX tests of two different amps (one tube with a high output impedance, one SS with a low impedance) on two different speakers. One speaker had a benign/flat impedance curve, the others varied widly. On the benign speakers the amps were indistinguishable from each other when level matched and blind folded. On the speakers with the widly moving impedance the amps were easily distinguished due to the differences in FR caused from the interaction of the tube amps output impedance and the speakers impedance compared against the much smaller changes in FR with the low output impedance SS amp.

Bob Carver says this is where a large portion of the 'tube' sound comes from. In some of his amps he literally offers an output connection which basically has a large 1ohm power resistor in series with the output to cause the same FR type deviations.

Shawn

[UncleRobb]

Al, how can I now sleep well in the night without having your Xover in my La Scalas? ;-)

The Decware SE84C seems to have a quite low output impedance for a tube amp, though. Anyway, I made a new thread in the Decware Forum: Klipsch LaScala/SE84C-a match from heaven or hell?

Thanks, Robert

PS: Al, how does the phase angle depend on the frequency for the AA and for your Xovers?

[Randy Bey]

I am far from a technician on this subject, but what I understand is that impedance, far from the Stereophile level of significance, is not the stumbling block that the typical thought process dictates.

The reason is: when the speakers impedance gets harder to drive, decreasing the efficiency of the power amp, the speaker simultaneously gets proportionately more efficient, thus requiring much less power to drive.

If it were not so, most every speaker in the world would be unlistenable.

So, my advice is not to get hung up on particular electrical characteristics, and just go with what sounds best. The next time you get hung up on how much capacitance or inductance your speaker circuits have, lie down with a cool compress on your forehead until the feeling goes away.

[sfogg]

"If it were not so, most every speaker in the world would be unlistenable."

If you amp has a low output impedance it doesn't interact with your speakers impedance to nearly the same degree.

Back in the days of tube power only companies like AR specifically added impedance compensation to their networks because of this. In the days of SS power amps it isn't as much of an issue.

The FR changes aren't theoretical. The Audio article I mentioned earlier measured the changes to the same speakers frequency response based on the interaction with the amp that was driving it. When the amps were measured by themselves into a non reactive load their FR didn't show the same changes as seen when plugged into the speakers. They interacted with each other.

No one here claimed not to go with what sounds best. I was just curious about the modeling Al has done with his networks.

Shawn

[Al Klappenberger]

Hi guys,

I think I agree about the frequency response being affected by the load

impedance if the amps source impedance is high. There's a common

specification for that, it's called "damping factor". A low damping factor is

a high impedance. The question is how low can the damping factor get before

the load becomes a significant part of ratio of source / load. That's where

my experience ends! The old Marantz 8B I used to have could be adjusted to

lower the damping factor as well as set for triode operation. I never tried

either option so I have no experience with it. The 8B got traded in on a

McIntosh MC250 SS years ago.

As to phase differences between my network and the AA, well, to be honest,

there is virtually no difference between the two and it doesn't matter

anyhow! I believe that you can't hear phase and time errors within a complex

waveform. If you could, the Khorn would sound like crap because the three

drivers are WAY out of time alignment. Your ear CAN detect phase difference

between sounds from two different sources however. That's part of the stereo

image we all look for. I have an analyzer that can display the transfer

function of any device displaying amplitude and insertion phase (Spectral

Dynamics SD375). I connected it between an amp and a calibrated microphone to

display the transfer function of a speaker (the "Heresy on steroids" that I

have in the den). The phase shift looked like a saw blade! Worrying about the

few degrees contributed by the second or third order filter in a crossover

network is silly!

An opinionated character aren't I ?

Al K

[UncleRobb]

The reason why I asked for plots of the phase angle vs. frequency is that Id like to know how difficult the load is on an amp. Do I understand it right that in the AA xover, the impedance peek at 2kHz insures that the more sensitive squawker gets lesser power and that in your (ALK) xover, you avoid this peek by destroying power in the 10 Ohms resistor parallel to the autoformer?

From what Al says, I would expect that his and the AA xover sound virtually the same. Is this right?

Btw, one argument that I hear from multi-range, single-driver speaker users is that the xovers in multi-way speakers not only mess up the phase information (we already discussed that), but also lower the efficiency when music is played. I thought that speaker efficiencies are measured INCLUDING the xover. It is not clear to me, whether these measurements are done with a sine wave, or close to, of say 1kHz or some pink noise. Is it possible that two speakers with say 10dB difference in efficiency generate about the same sound pressure level with music? Im thinking of the La Scalas compared to lesser efficient single-driver speakers.

Thanks, Robert

[sfogg]

Robert,

"but also lower the efficiency when music is played. I thought that speaker efficiencies are measured INCLUDING the xover."

It is including the xover. What they may be talking about is the inherent efficiency of the drivers themselves. On a single amped system if you have a three way system your end efficiency will basically be no better then around the point of your lowest efficency driver. Reason being is the xover has to scale back the other two drivers to have a balanced system. For example your tweeter is 100dB/w/m, midrange is 105dB/w/m but your woofer is only 85dB/w/m. For these drivers in a system with a relatively flat FR at best you are going to have a speaker that is 85dB/w/m efficent.

This is what the transformer is doing on the squaker in the La Scalas. The squaker is more efficent then the tweeter or the woofer so to balance that out you have to in effect lower the efficiency of the squaker. This is an advantage of a multi-amped system that moves the crossover ahead of the amplifiers and has the amps directly connected to the drivers. Now all your drivers can run at their normal efficiency and to balance the system out you just lower the volume going into the amp that runs the more efficent driver.

On a single driver speaker you only have the one speaker so there is no balancing needed. As such it runs at whatever its efficiency is. Of course a single driver speaker has all sorts of other problems like IM distortion, very uneven dispersion (which is a function of wavelength and. driver size on conventional driver non-horn systems) and so on.

Shawn

[Randy Bey]

Robert,

I have had direct experience with the Decware Zen amps and Klischorns; I'm glad it had a 30 day money back deal becauseI sent it back after 29 days.

You should try it yourself. This amp is truly sweet, a wonderful sounding little beauty.

It's only fault (besides a hum issue which may not have been the amp) was that it was nowhere near the 5 or 2.5 watts that Steve D. claims. In fact, I question if it could make it to 1 watt going downhill with a tail wind.

I heard it routinely clip on relatively easy music at nominal listening levels. You would have to have a very small room and go with near-field listening to enjoy them.

Even then, you might need to constrain yourself to girl-with-guitar music.

But try it, you may well like it! Steve is the consummate professional and took it back, no questions asked.