Crossover puzzle

[pauln]

So would the old 104dB and new 105dB efficiency specs be considered "nominal"?
The actual spec depending on the network with the losses increasing with order?
If the A and AL-4 represent the extremes of the stock network design domain, and the noticable difference is about 1dB, then maybe all the networks fall into a variation range of that 1dB.

Marvel and NOSValves, it is true that I don't listen loud. Anything that peaks into the low 90's is really too loud for me. The A's are amazing.

Eric, about the 2 watts being too much for 45 and SE OTL... I've read that whereas only non-SE amps can have dynamic headroom, the rated output for all amps is an average value and all amps including SETs can hit an instantaneous power level +3dB above their rated power (double rated power). Does that sound correct?

The Wrights as you know will also shift from A1 to A2 (grid current flow) if there appears very high demand peaks. I think A2 allows 8W peaks (maybe only with Sovteks?) if called upon?

[Erik Mandaville]

Paul:

I was just pointing out an example of a couple of amplifiers that in relation to the resistor-based attenuation in crossovers have a very low roof over their heads. I once built a pair of slightly more complex (higher order) networks to compare against types 'A' and 'AA' and happened to have a just finished SE OTL here at the time. The match was uacceptable, with both the A and AA sounding much more extended and open, particularly at lower volume levels.

While I've been surprised at times at just how loud and clean these lower powered amplifiers can sound, I've also heard them run into clipping when listening to surround SACD at louder volume levels than we usually listen. That fact doesn't make me like them any less, however.

The Wright 2A3s are wonderful amplifiers, and have some old and clever ideas built into both the power supply and amp circuit. I preferred them to stock Moondogs (100% subjective), although the Moondogs haven't been 'stock' for a year or so.

I agree, the type 'A' is a great little network.

Erik

[NOSValves]

Eric, about the 2 watts being too much for 45 and SE OTL... I've read that whereas only non-SE amps can have dynamic headroom, the rated output for all amps is an average value and all amps including SETs can hit an instantaneous power level +3dB above their rated power (double rated power). Does that sound correct?

The Wrights as you know will also shift from A1 to A2 (grid current flow) if there appears very high demand peaks. I think A2 allows 8W peaks (maybe only with Sovteks?) if called upon?

Absolutely true just about all amps have dynamic headroom problem is some amps have it along with extreme distortion numbers to go with it. Like 20, 30 or 50% possibly even more. I'd personally would rather never use even a little bit of an amps dynamic headroom no matter what the topology.

Craig

[pauln]

"I'd personally would rather never use even a little bit of an amps dynamic headroom no matter what the topology."

Roger that!

[BEC]

Well got around to running a plot on the AL-4 crossover today.  Still trying to figure out what I am seeing.  Looks very light in the midrange to me as far as output.

First since we were talking about a type A crossover comparison, here is the type A.

Bob Crites

 

[BEC]

Now the AL-4 crossover plot.

Bob Crites

 

[sfogg]

"Think back to the way I said it. A watt into the driver gives a

certain acoustic output. Now what power does the amp have to output to

give that acoustic output?

The answer with an autotransformer

attenuating 3 db is still one watt. (ignoring for the moment very small

losses of the circuit).

The answer for the resistor attenuating the same 3 db is 2 watts."

Huh?

If you feed a driver directly from an amp then put something between the amp that attenuates the signal 3dB to have the same SPL level the setup with the 3dB of attenuation needs 2x the power output from the amp as the unattenuated setup.

Shawn

[sfogg]

" Still trying to figure out what I am seeing. Looks very light in the midrange to me as far as output."

You are only looking at one half of the system. The electrical output of a crossover is only half the picture. What it does when combined with the acoustic output of the drivers is what one actually hears.

Shawn

[BEC]

"  Still trying to figure out what I am seeing.  Looks very light in the midrange to me as far as output."

 

You are only looking at one half of the system. The electrical output of a crossover is only half the picture. What it does when combined with the acoustic output of the drivers is what one actually hears.

 

Shawn 

 Shawn, 

That is why I showed the electrical for both the crossovers he is comparing.

Bob

 

[BEC]

"Think back to the way I said it.  A watt into the driver gives acertain acoustic output.  Now what power does the amp have to output togive that acoustic output? 

The answer with an autotransformerattenuating 3 db is still one watt. (ignoring for the moment very smalllosses of the circuit).

The answer for the resistor attenuating the same 3 db is 2 watts."

 

Huh?

 

If you feed a driver directly from an amp then put something between the amp that attenuates the signal 3dB to have the same SPL level the setup with the 3dB of attenuation needs 2x the power output from the amp as the unattenuated setup.

Shawn

 If you had one watt before then put in an autotransformer that attenuates 3 db, the amp output would go to 1/2 watt.  To get it back to one watt to the driver, you would increase the power of the amp back to 1 watt.

Bob

 

[pauln]

Thanks for the graphs, Bob.

[Erik Mandaville]

Is there a schematic available for the AL-4?

Erik

[BEC]

Erik,

I don't think Klipsch has made that schematic available yet.

Bob

 

[Erik Mandaville]

Bob:

Ah, okay -- thanks, I was wondering if that might be the case. I'm just curious to see what kind of a network it is -- slope/s used, etc.

It uses fixed resistor L-pads though, correct? (just if you or someone happens to know). Since we know the driver impedances and crossover frequencies, it would be straight forward to build something similar to try instead of, for example, a type 'A' sans autoformer. The upper and lower limits are already established, and just the band pass and squawker L-pad needs to be figured out. I can tell right now I'd need a coil and large cap that I don't have, but it would be interesting to try (in addition the trillion other things I want to do).

Other than that, I single Heresy makes a very decent (but still temporary) center channel speaker!

Erik

[garyrc]

AL-4: They really want to get that squawker outta there, before it can cause trouble? ?? Does this not avoid the famous (above bandwidth) glitch in the K-55 V (and F ?)? And maybe they wanted less interference of the K55 with the K77, above the K55's designated range?

[sfogg]

" If you had one watt before then put in an autotransformer that attenuates 3 db, the amp output would go to 1/2 watt. "

Maybe, maybe not. Depends upon how the amplifier reacts to the change in impedance. A low output impedance amp (SS) will act differently in response to the impedance change then say a higher output impedance transformer coupled tube amp and an OTL will act differently yet again.

Shawn

[Erik Mandaville]

"Maybe, maybe not. Depends upon how the amplifier reacts to the change in impedance. A low output impedance amp (SS) will act differently in response to the impedance change then say a higher output impedance transformer coupled tube amp and an OTL will act differently yet again."

I think that's true, and certain kinds of L-pads can also help offset or compensate for impedance variations, making breathing for whatever amplifier a little easier.

Erik

[pauln]

I think the AL-4 is Linkwitz-Riley 6th order (-36dB/oct slopes)?

[Erik Mandaville]

"I think the AL-4 is Linkwitz-Riley 6th order (-36dB/oct slopes)?"

If that's the case, I can see how you and your amps might prefer a more simple network, particularly at lower volume settings. I was going to get the parts to make a pair of ALK ES networks awhile back, but the more I thought about the application, the less I was convinced they would be the right thing here. More than that, also very expensive for a trial and error approach. They're heavy, and impressive looking, particularly the way Al builds them, but I'm not a tiny music signal trying to swim to a loudspeaker. with a harder push than what most SETs are capable of, I'm sure they are excellent networks -- as I'm sure is also the case with the AL-4.

I'm not going to build something with that number of parts, but am curious about the difference between a fixed -3dB L-pad, which is possibly what the AL-4 uses, and the autoformer; and particularly if the output of one is obviously lower than the other, given the same input.

Erik

[Erik Mandaville]

I found the parts I need for the bandpass section of this resistor L-pad, first order parallel network I designed. Based on the values of inductance and capacitance needed for the bandpass, the upper frequency crossover point is set at about 4500Hz, which is what I'm using with Bob's tweeters in the Klipschorns now (using a type 'A' network). The values are close, but to narrow the band pass slightly, one has to add a certain percentage to one value, and subtract from another, which results in final values of L (inductance) and C (capacitance) that are slightly different from the original calculated figures; and those are the values I don't have on hand. What I have will work for this trial, and to see the behavior of the resistor-based L-pad (instead of an autoformer, as seen below. The L-pad, which presents the same reflected impedance as the driver itself, will reduce the output of the squawker by 3 dBs.

*Note: The whole thing here is to test my own curiosity, using some close value parts I had on hand. I spent $4.00 in resistors for the L-pads, and that's all. The rest of the network takes advantage of another, higher order design from a few years ago, using about half of its parts. The new networks will be done later this morning.