Vertical Bi-amping and a Different Idea

[greg928gts]

Does anyone out there have input about this article on vertical bi-amping.

This is the link to the article:

http://www.caryaudio.com/articles/bi_amping.shtml

In particular, this is the part of the article that I'm most interested in learning more about:

"The woofer (or multiple woofers) will exhibit lower impedance as they go below their respective self-resonance (Fs) frequency. The resonant frequency will be considered the lowest usable frequency of the woofer in the system for this discussion. As the bass notes go deeper below Fs, the woofer will tend to drop lower in impedance. These dips can call upon the amplifier to deliver more current in the lower registers. If the driving amplifier needs only to produce current in these registers it will have more control and usable power since the high frequencies going through the amplifier are looking into a very high impedance, many times on the order of 20 to 30 ohms. The amplifier will simply throw out voltage in these out of band frequencies without the need to deliver current and in turn power.

The same occurrence takes place with the midrange tweeter crossover system. As low frequencies enter a tweeter midrange crossover network, the amplifier looks into a very high impedance that will not pull down the amplifier. In simplistic terms, the amplifier channel that is feeding the midrange tweeter network will only be delivering any appreciable current when the frequencies are in the mid to high frequency range."

If this is true, I could use this same theory with a conventional bi-amp setup, without using an active crossover, AND without putting undue strain on my SET amp to produce bass frequencies when hooked up to just the mids and tweeters on my K-horns.

Does anyone follow what I'm gettin at here?

SET amp on my mids and tweets, and my big SS on the woofers, without the "coloration" of an active electronic crossover in line. Traditionally, the thinking is that the SET is still going to be working hard to reproduce the full spectrum, including bass frequencies. So in order to take the "load" off the SET, the bass frequencies must be rolled out before the signal gets to the amp.

But if what this article implies is true, then this really isn't the case, because due to higher impedences, the amp really isn't "seeing" a big load when it's hooked up to just the mids and tweets.

Greg

[djk]

If you have a 'normal' tube amp, bi-amping the Klipschorn is easy.

Remove the A/AA/whatever network, and store on the shelf.

Hook the K55 straight to the 8 ohm tap, and the K77 to the 16 ohm tap through a 2µF cap. You can 'borrow' a 2µF from your old network, or buy something exotic.

This duplicates the type A crossover for level and frequency for the tweeter.

If you don't have a 'normal' tube amp with taps, just unhook the woofer from your type A/AA network and run your amp straight to the K55. This cannot be done with networks newer than the A/AA (but it works with the old 500/5K).

Change the input coupling cap and/or the input resistor (also called the grid-leak resistor) so that it rolls off below 400hz, a 0.01µF + 39K = 400hz.

This duplicates the type A crossover for the midrange.

Buy a cheap car stereo crossover and a 12V wall plug for the bass. Most of these are 12dB/oct and have level controls for the amplifier. Pyramid/Radio Shack make ones for about $30 that work OK. Change two resistors and/or caps per channel to make it 400hz.

Later you can make a 'real' crossover with good parts if you like, I usually just change out the opamps and such with better and am done with it.

Off-the-shelf electronic crossovers don't cut it.

[Deang]

Cool. Always interesting info from you -- my favorites bar is piling up with the bookmarks.

Dennis' notes on vertical biamping presuppose one is still using the passive networks in the speaker.

[greg928gts]

Thanks djk for the thoughtful response. I wish a few others would chime in on this Cary article though.

Did you get a chance to read the Cary article? If so, do you agree with the statements made that the mid/tweet amp will not be working hard to reproduce the bass notes due to high impedences even though it's being fed a full range signal? This is a key element to what I'm trying to do, because I already am getting great sound quality from the 3 wpc SET amp, but not enough kick. If I'm going to still load up the amp in a bi-amp situation, then I'm really not going to get more power to just the mid/tweet, which is what I'm trying to do. If the amp is busy making bass and then it's wasted as heat at the crossover, I'm still using the amp harder than I want to. I only want it to be working hard to produce 400hz and up. I just don't know if this is possible without an active crossover. The Cary article, if true and accurate leads me to believe that it is possible.

I'm not sure I understand what you're suggesting with hooking up to the K55. My amp doesn't have taps. do you mean hook the amp directly to the mid driver, or on the AA crossover? You said "change" the cap, not "add" a cap, which leads me to believe you mean change the one on the crossover. In this case I assume you mean to hook the amp to the crossover, but that's not exactly what you said, you said "run your amp directly to the K55". I'm finding the wording confusing. Also, if I just hook to the K55 with the cap, I'll have no tweeter.

Sorry, I'm not familiar enough with these crossovers and electronic terminology, and need to have things spelled out very clearly, otherwise I ask a lot of detailed questions.

Greg

[djk]

I'm tired of discussing 'fools' bi-amping, it seems to be too technical for lay people.

If you do not use a filter(crossover)to remove the out-of-band signals from feeding the amplifier, they still feed the amplifier.

If they are still being fed into the amplifier, the amplifier is still amplifing them.

When you run the input signal up past a certain point the amplifier will clip.

If you remove the load(say the woofer)from the amplifier it no longer delivers any power at the frequency the load was at.

But the amplifier will still be clipping.

Because the input signal is still driving it into clipping.

Changing the input coupling cap/resistor to 400hz removes some of the input signal, about half of the voltage in a musical signal is below this point.

The gain of the preamp may now be increased about 6dB before the power amp goes into clipping again.

Most tube amps have taps for different impedance speakers. The amplifier puts out the most power into a matched impedance.

Since your amp does not have taps you will have to use your AA network to run the tweeter.

In the normal uno-amp mode(not bi-amping)the amplifier signal goes into the crossover. From there the bass half goes through a 2.5mH inductor to the woofer. The woofer has been un-hooked and run to another amp, so we don't have to worry about it. The mid/hf half of the signal goes through a 13µF cap to tap 5 on the autoformer, the hf part of this signal goes through the rest of the crossover to the tweeter, the mid part comes out on tap 4 to the midrange driver. This gives a 400hz crossover and lowers the level of the midrange driver about 3dB.

If we hook the amp straight to the mid driver we loose the cap and the attenution, neither of which we needed. The tweeter is still driven with the AA network and gets its proper signal. We have now picked up another 3dB, for a total of 9dB better than uno-amping. This is good as it takes about 10dB to sound twice as loud. We have effectively made the input for the AA crosover tap 4 on the autoformer rather than tap 5.

One thing to try at this point is to use tap 3 on the autoformer as the input for the AA network. Using tap 4, straight into the mid driver, gives a 16 ohm load to the amplifier. Using tap 3 gives an 8 ohm load to the amplifier in the mids, dropping down above 6Khz where the tweeter comes in. There is very little power above 6Khz on most program material, the amplifier won't care(the reverse is not always true. The woofer drops down to about 5 ohms in the region where there is a lot of power, some amps don't like this and need to have the Klipschorn on their 4 ohm taps, many work well on the 8 ohm taps. It is just something that must be tried both ways to see what sounds the best).

If you choose to use tap 3 as the amplifier input you will pick up another 3dB in volume. All reference to tap numbers are for the 'hot' or + lead of the amplifier, the 'ground' or - lead is connected to the 0.

Many people are happy with the type A network which feeds the tweeter from a single 2µF cap. If you have a normal tube amp with 4/8/16 ohm taps you don't need the Klipsch network at all. Run the mid off the 8 ohm taps and hook the tweeter to the 16 ohm tap through a 2µF cap. You are using the output transformer taps to balance the tweeter, the amplifier transformer is of a better quality for this purpose than the one in the Klipsch network.

The low-pass filter section of the electronic crossover can be put together as mentioned with a car stereo crossover. Ideally we want a 12dB slope at 400hz with a Q=1, and I always sneak a 12dB high-pass in at 32hz with a Q=2.

The Q=1 crossover was used for the Klipschorn very early, this version of the Klipschorn was also bi-wired. PWK's remarks about it may be read at US patent #2,612,558. The crossover was part of the amplifier, a push-pull 2A3 design.

PWK's patent #4,237,340 also uses a Q=1 to smooth out the midbass on the K33E. All networks after the AA for the Klipschorn/LaScala/Belle have a Q=1 on the woofer for the same reason.

The Q=2 at 32hz makes the Klipschorn go lower, and protects it from large sub-sonic signals below horn cut-off.

[greg928gts]

djk, this is really good information. I'm reading and re-reading to absorb it. Some of this info is just a little bit out of my league though.

Just to clarify, when you speak of a input coupling cap/resistor, is that the same thing as what you're referring to making with a car stereo crossover? I assume this is in line before the mid/tweet amplifier input?

I'm still confused by the hookup directly to the mid driver. If the speaker wires go directly to the mid driver, how does the signal go through the AA crossover to the tweeter? Jumpers?

What is your reference to "fools" biamping about?

Thanks,

Greg

[djk]

"Just to clarify, when you speak of a input coupling cap/resistor, is that the same thing as what you're referring to making with a car stereo crossover? I assume this is in line before the mid/tweet amplifier input?"

The input coupling cap is the input coupling cap of the power amplifier. It is inside the power amplifier. This becomes the high-pass portion of the crossover. The car stereo thing is for the low-pass portion of the crossover.

"I'm still confused by the hookup directly to the mid driver. If the speaker wires go directly to the mid driver, how does the signal go through the AA crossover to the tweeter? Jumpers?"

The woofer is unhooked fom the AA network and is hooked to its own amp and filter. The tweeter and mid are left connected to the AA in the normal fashion. The amplifier is hooked straight to the mid. The wire from the AA to the mid carries the signal from the mid back into the AA and through the tweeter section of the AA to the tweeter.

"What is your reference to "fools" biamping about?"

The Cary article, any paralleling of amplifier inputs without the use of filters ahead of the amplifiers and thinking you are 'bi-amping'.

[mike stehr]

What if the power amplifier only had resistors for input coupling to grid without a coupling cap?

Could you still configure resistor values to come up with a particular chosen frequency for input in a crossover application?

(I guess you could always add the coupling cap.)

Cool idea, a little line level passive crossover network right before the input grid.

So they bi-wired a Klipschorn using OPT impedance taps from a Brook 2A3 Push-Pull?

Was the particular output transformer taps wound with the Khorn impedances in mind?

Or they just try it to see what happens?

Sorry. Those are interesting posts.

[djk]

"What if the power amplifier only had resistors for input coupling to grid without a coupling cap?

Could you still configure resistor values to come up with a particular chosen frequency for input in a crossover application?

(I guess you could always add the coupling cap.)"

Just add an input cap and juggle that and the input resistor to get the frequency you want.

The Dahlquist electronic crossover was this way, as was the JBL BX63, and others.

"Cool idea, a little line level passive crossover network right before the input grid."

Easier than a seperate box, jacks, and extra cable.

"So they bi-wired a Klipschorn using OPT impedance taps from a Brook 2A3 Push-Pull?

Was the particular output transformer taps wound with the Khorn impedances in mind?

Or they just try it to see what happens?"

PWK used two different output transfomers at the same time with only one pair of output tubes. Each transformer was optimized for the frequency range and load it had to drive. At the time PWK did this(1946)caps were a bit pricey and hard to get. By keeping the caps at the high plate impedance his values were less than 0.1µF for the 400 hz crossover point vs 13µF at the speaker level.

Also, it is more expensive to make a wide band output transformer for the amplifier and another autoformer for the speaker than to make two inexpensive narrow band transformers.

Unfortunately, I don't think this idea had any real interest at the time.

Yamaha pro sound electronic crossovers have transformer balanced outputs that are different sized for the bass/mid/treble, a bit along the same line.

[mike stehr]

I think some of the old organ amplifiers used various output transformers for specific bandwidths.

I'm pretty sure of it, in mono applications of course.

Thanks for the reply, Dennis!

[greg928gts]

See, this is where I start to get lost, but I like how it's starting to sound. So there are things called input coupling caps huh? Do all amps have these? Is it difficult to change to one that will high pass 400hz and up? I would have a friend of mine who is an electronic repair man do it for me.

What a great idea, just a simple component instead of an electronic crossover with all the controls, plugs, cords, etc... I love the sound of this.

Is there any reason why I can't cross the Khorn midrange lower than 400 hz? 250 hz would sound so much nicer to my ears.

Thanks djk

Greg

[djk]

"Do all amps have these? Is it difficult to change to one that will high pass 400hz and up? I would have a friend of mine who is an electronic repair man do it for me."

99% of most amplifiers have them, if not you can add them. 15 minute job for a tech to make a 400hz hi-pass out of the input circuit.

"What a great idea, just a simple component instead of an electronic crossover with all the controls, plugs, cords, etc... I love the sound of this."

Works pretty good, that's why I suggested it. These kind of filters don't work well for the low-pass, that's why we go active here.

http://www.arduman.com/aa/Sayfalar/slagle/slagle.htm

Excerpt from above:

"but the damage done to the higher frequencies by the X-over was awful, the entire system was unlistenable. I also played briefly with the ultracurve, but again the sonic signature was not small and made the system very tiring to listen to... So neither will ever be allowed to get in the way of the music again.

I don't mind the digital X-over on the bass, since below 150hz, there isn't really much damage it can do... and it makes tweeking the points really easy... So I will keep it until I work on a few different basshorn ideas,"

"ultimately I want to build the X-overs into the amps and do away with any unneeded stuff in the signal path."

"Is there any reason why I can't cross the Khorn midrange lower than 400 hz? 250 hz would sound so much nicer to my ears."

From your previous description of how you had it hooked up I think you still had the sound going through the passive crossover. The frequency and slope would not be what the knob said. I think if you try it my way you will be happy with the sound. You could cross lower, but that would require a steeper crossover to sound the best.

A good reference on this subject:

http://www.passlabs.com/pdf/phasecrx.pdf

[greg928gts]

Thanks again djk for your thoughtful response to my questions.

I have put together a little schematic drawing to help me understand how this all works and have a couple more questions.

How is the input signal split to go to the two different amplifiers? I have a McIntosh C31V preamp that has main outputs and a couple of other outputs labeled "switched 1 and 2" and a tape loop output. How do I know if these signal outputs are adequate to feed the second amplifier in my bi-amp setup? What if I had a preamp with only one output, is it o.k. to just use "Y" splitters?

In your previous posts, you mentioned tap numbers on the AA crossovers that do not seem to correspond to the numbers on MY AA crossovers. My numbers are as follows: input is 1 black and 2 red, woofer is 3 black and 4 red, squawker is 5 black and 6 red, and tweeter is 7 black and 8 red.

For instance, you said "the mid part comes out on tap 4 to the midrange driver", but as I see it, tap 4 is the positive to the woofer. Am I not understanding this correctly?

Any idea where I can find schematics to build (or have built) a very nice quality active crossover for the bass instead of using a car stereo crossover?

Thanks,

Greg