Tri-Amp & Time aligned K Horn

[rigma]

I just purchased a Crown DPS processer(IQ-USM 810), which has crossover, time delay & EQ functions for each of 8 outputs. My intention is to Tri-Amp my Klipschorns using Crown DC-300 in mono for each woofer, DC-150 in mono for each Midrange & DC-75 in mono for each Tweeter. I understand the crossover for these is 400Hz & 5000Hz ?? Does anyone know the difference in distance from the woofer to the midrange driver? so I can use the delay function to time align. I know some will question why I am going to this expense, but I currently have all this equipment so just want to hear the results.

[Guest BobG]

Let us know the results. Typically, multi-amp arrangements do not improve the sound of the Khorn. Too much electronics in the way. Also, it's very difficult to create the necessary eq needed for the compression drivers. Compression drivers are not flat throughout their range and must have eq. designed into the crossover network.

Quite a collection of older Crown stuff though. Used to sell/own a bunch myself. Did lots of studio recording on my SX-828. Heavy to drag to live gigs though.

[rigma]

Thanks BobG, By doing the processing in the digital domain I am hoping not to degrade the electronics. The Crown processer has several crossover slopes(shapes) 1st 2nd 3rd 4th-order Butterworth; 2nd 3rd 4th-order Bessel; & 4th-order Linkwitz-Riley. Maybe AL K can help me with this and the EQ. Al are you out there?

[Al Klappenberger]

Rigma,

I wouldn't worry too much about time alignment of drivers. Unless it is so

drastic that sound arrives so much slower form one driver than another that

you hear a echo, you can't hear it. I think this has happened in huge

theater audio systems. I have heard stories that Paul Klipsch performed an

experiment where he moved a driver back and forth in front of an audience

such that the movement could not be seen, from behind a screen, I assume.

The audience could not hear the difference. He did this to prove that the

long path through the Khorn woofer was not a problem compared to the shorter

path from the squawker. The fact is that the relative path time between

drivers changes at different places in the room anyhow. For example, at the

6000 Hz squawker / tweeter crossover frequency, the wavelength is only about

2 inches. Move 2 inches sideways and the timing (phase) changes drastically.

The bottom line is that you can't hear it anyhow! What you CAN hear is

frequency response variations near the crossover caused by interactions

between drivers making the same sound at the same time. I have built a

crossover from my modified Belles that have a slope in excess of 100 dB /

octave at 700 Hz on both drivers. It crosses over smooth and I can hear no

ill effects from the high group delay at the crossover. The idea is to

totally remove any energy adjacent to the crossover frequency both above and

below it so there is no sound energy to interact. It's and old idea and

referred to as an "infinite slope" crossover. You can see it described on my

personal web site. Just click on the little house below this post.

AL K.

[rigma]

Thanks Al, so which crossover shape would you use? By the way I have purchased all the components to build 4 of you Networks, 4 for LaScala (2 center & 2 surrond) you may remember us speaking via phone a couple of weeks ago, thanks so much for sharing design. BobG mentioned EQ built into crossover, is there anything other than the slope (crossover shape)?

The delay feature is built into the Crown, so I figure why not try it, don't think it could hurt.

[TBrennan]

John Hilliard did experiments with time delay at MGM in the early 1930s after the famous Elenor Powell "double-tap" monitoring incident. The monitor was a 2-way WE rig with a 15' differnce between the drivers of the straight basshorn and the treble horn. On playback of tap danceing a double-tap was heard. Hilliard found that with a 500hz crossover there was no audible effect if the path difference was under about 3 feet. The Shearer Horn project at MGM was a result of Hilliard's dissatisfaction with WE speakers, the Shearer horn had a much better distance "matchup" than the WEs, subsequently WE and RCA went to Shearer systems. On the other hand one of the reasons Hilliard designed the VOT a decade later was to get a perfect distance matchup of the bass and treble drivers and instructions for setting-up VOTs were very specific on driver distancing and how to test for it. So evidently this thing still bothered Hilliard despite his earlier experiments, this was certainly the reason Altec never went in for folded basshorns but preferred combination straight horn-reflex cabinets like the 210 and 825. I would be very interested in knowing if time alingnment (trademarked by Ed Long) will improve the sound of KHorns, the distance between the bass and mid drivers is on the ragged edge of Hilliard's findings. IME multi-amping with active crossovers always improves the sound of a speaker; better dynamics and lower distortion are apparent.But as BOB G. noted this can screwup frequency response if the drivers have been passively EQed unless you address that.

[Al Klappenberger]

Rigma,

There are no seperate components in the Khorn networks that are equalizers. There is only filter components. Filters alone contribute dealy and must match the particular speaker. That I way I can't design networks for speakers that I do not have here to evaluate.

AL K.

[rigma]

Thanks TBrennan great info. Do you know the distance diff. of woofer to midrange driver. Al, what about the crossover slope choices I listed for the Crown processor? Which should I use?

[KGeist]

Although I would consider the midrange autoformer on the Khorn network board a form of "equalization", other than that, there are no circuits which are purely intended for that function. On the other hand, it must be remembered that filters come in an infinite variety of alignments (Butterworth, Bessel, etc...) and therefore can be used to equalize a system.

About the only disadvantage of using equalization in this fashion it that the only areas where you have the latitude to equalize is confined to a relatively narrow band (about 1 octave) centered around the crossover point. Surprisingly enough though, this is also the area where drivers usually need the most help because they are typically at the extent of there usable bandwidth.

The problem with using an electronic crossover is that although you may have different alignments programmed in the processor, it is very difficult, without exhaustive testing and listening, to determine which one to use.

Kerry

[JohnA]

Start off by using the original crossover slopes. I don't know which squawker you have, but for the K-55-V and Type AA network, the bass is rolled off at 6 dB/octave above 400 Hz, the squawker is rolled in at 6 dB/octave above 400, has no upper roll off and the tweeter is rolled in at about 18 dB/octave above 6000 Hz. I say about because the high pass filter on the tweeter is not an ideal Butterworth. It has a "bump" in the response of a dB or so at 6k. I believe this is an attemp to EQ the system and make up for the K-55-V's falling response at 6K.

The K-55-M doesn't fall off at 6k like the -V does so I believe the networks for that driver roll off at 12 dB/octave above 6k.

John

[rigma]

John thanks for the info, I have the K-55-M midrange drivers, and will try the slopes you listed.

[slimwoodie]

wow !! just found this forum, .... very smart people here ... your ear can hear 3ms. of time smear .... but, that doesnt mean your drivers a couple inches apart will make a difference .... the k-horn, with that long path, on the surface, would cause concern , i would tune your digital x-ovr to what pleases your ear, after all .... thats the point right ??

[slimwoodie]

by the way, if i had to choose an approach with the k-horn, i would go with the infinite crossover, and a lot of watts ..... i run 1200 crown watts pc, hey, who can hear time smear at 125 db spl ??

[slimwoodie]

i have some programs i use for xover/time constant .... whats the distance alignment of the drivers in the k- horn ?? bass 3 ' or so ?? how about the rest ?

[Al Klappenberger]

Rigma,

I see I am late checking back on this thread!

The big thing to remember with passive crossovers is that the number of

elements (parts) in the network determines the phase between the two drivers

at the crossover. I think and active crossover will act the same. A second

order (12 dB / octave) network, for example, will make the two drivers

exactly out of phase with each other and needs to have one of the drivers

connected backward to compensate! This will make the sound from the two

driver cancel each other out at the crossover frequency if you don't reverse

one of them. A first order (6 db / octave) leaves the two driver 90 degrees

out of phase. This works fine. A horn speaker also has a 90 degree different

output than a direct radiator. Networks need to be paired with a speaker

carefully. The big "infinite slope" network I am using with my modified

Belles came out to have 110 degrees difference between the woofer and

squawker outputs at the crossover. That's pretty close to the 90 degrees of

a first order filter. I also designed the squawker channel to have slightly

more delay then the woofer channel by giving it a sharper slope. According

to a computer program I wrote to add path delay to the computer analysis of

both channels of the network and sum them together algebraicly, the outputs

summed nicely to a flat response at the crossover frequency. The extreme

slopes made interference between drivers away from the crossover impossible

because the energy is gone. The path lengths of the two drivers was

determined mainly by guesswork since you really can't determine it

accurately by using a yardstick!

The bottom line is that you should simply adjust your crossover for what

sounds best unless you have a calibrated mike and plotter! My guess is that

you will be hard pressed to tell the difference between equalizaion

settings.

Al K.

[WMcD]

I may be able to add some numbers.

The Heyster review of the K-Horn in Audio reported that at his mike position about 1 meter away from the front of the speaker:

The tweeter sound arrived after 3.7 mS.

The midrange sound arrived after 5.38 mS.

The woofer sound arrived after 8.4 mS.

Rigma knows the following, I'll certain, but for the benefit of others . . . .

The above delays make sense because we know driver for the tweeter is the source of sound for it and is a few inches behind the front of the cabinet. The driver for midrange is about 22 inches farther back, the midrange horn being about that long. The driver for the woofer is inside the "doghouse". The sound from it travels the path of the folded bass horn.

What to do ?, assuming we want the sounds to arrive at the front of the cabinet at the same time, and then out into the room.

We can't use electronics to make the woofer sound, for example, arrive faster. In all real world / real time systems, we can't make things happen in the future. I.e. we can't put a signal input in and have it come out earlier. However, we can use a delay in the Crown unit to even out the delay caused by the offset distances described above.

We're going to let the woofer get an undelayed signal.

The difference in arrival time between the woofer and tweeter is 8.4 - 3.7 = 4.7 mS. Therefore, the tweeter sound should be delayed by 4.7 mS. Essentially, it is stored in the electronics of the Crown waiting for the woofer sound to arrive at the front of the cabinet.

The delay in the midrange is 5.38 - 3.70 = 1.68 mS. Therefore it must be delayed 1.68 mS less than the tweeter. 4.70 - 1.68 = 3.02 mS.

I did that last calc last night and found an 0.1 mS error. Still can't figure it out. But these are pretty good starting points.

Al speaks with authority about the phase effects of the cross overs in general.

Some years ago I looked at the path length difference between the mid and the woofer. At the crossover of about 400 Hz, I found the path was worth 360 degrees of delay. This means that that in theory, correcting for time, or not correcting for time, is not going to alter the phase at the crossover point.

Also, it seemed to me that the path length difference in the LaScala and Belle of the woofer relative to the midrange cause a 180 degrees of phase shift at 400 Hz. Therefore the polarity of something should be reversed in those designs.

Overall, my guess is that it is no accident that things work out these ways. It is part of PWK's design goals.

Regards,

Gil

[rigma]

Thanks so much William that is the exact info I was looking for.

[Al Klappenberger]

Rigma,

Please let us all know what diference you hear when you do the time corrections form Gil's numbers. I would interested to know.

Al K.

[Chris A]

The difference in arrival time between the woofer and tweeter is 8.4 - 3.7 = 4.7mS. Therefore, the tweeter sound should be delayed by 4.7mS. Essentially, it is stored in the electronics of the Crown waiting for the woofer sound to arrive at the front of the cabinet.

Gil,

I was doing a Google search and found this thread. Note that, for the record, the 8.4ms delay is between the woofer and the tweeter, not the delay between the signal and receiving microphone.

The delay between the midrange and the tweeter is 1.67ms.

Sorry for updating this old thread, but this is just in case someone else finds this like I did using Google. I was actually looking for the Heyser article on the Klipschorn: http://forums.klipsch.com/forums/storage/3/1037187/Richard%20C%20Heyser%20KHorn%20Review.pdf

Chris

[o0O Bill O0o]

The difference in arrival time between the woofer and tweeter is 8.4 - 3.7 = 4.7mS. Therefore, the tweeter sound should be delayed by 4.7mS. Essentially, it is stored in the electronics of the Crown waiting for the woofer sound to arrive at the front of the cabinet.

Gil,

I was doing a Google search and found this thread. Note that, for the record, the 8.4ms delay is between the woofer and the tweeter, not the delay between the signal and receiving microphone.

The delay between the midrange and the tweeter is 1.674ms.

Sorry for updating this old thread, but this is just in case someone else finds this like I did using Google. I was actually looking for the Heyser article on the Klipschorn: http://forums.klipsch.com/forums/storage/3/1037187/Richard%20C%20Heyser%20KHorn%20Review.pdf

Chris

no, no, this was a good resurrection. It re-enforces that PWK Klipschorn design is solid. Tells me all the cash spent in tweaks is best spent on music. :-) P.S. What is the year count on this design?