Klipschorn Jubilee (Babies First Sounds)

[mikebse2a3]

Here are a few examples of the Horizontal Polar Plots that the previous charts represent for the K402.

505Hz to 625Hz

[mikebse2a3]

6485Hz to 7922Hz

My scan of these copies aren't as clear as I would like to show but if you look to the right of the plot you will see the fequency points tested.

Edit: Reposted PictureThanks to Dr Who

[DrWho]

Hey Mike, instead of scanning you can just use the select tool in adobe to copy a portion of the screen to the copy buffer...then go to mspaint and hit paste. Here, I'll show one of them cool planar regions:

[DrWho]

and it's corresponding horizontal

[mikebse2a3]

edit:[]

[mikebse2a3]

Thanks Mike

I'm computer challenged[] anyway I've tried something so lets see if it works.

[Al Klappenberger]

Those polar plots are what I have often seen. The same format is used to show the pattern of radio antennas too. Ploting them on a linear graph fliped me out! I don't know much about that sort of thing but they look good to me.

Al K.

[DrWho]

lol Mike, don't feel bad...I only just learned about that feature last semester (wow, has it been 6 months already?!?). All our profs put their lecture notes in pdf format and they allow us to bring a fixed number of pages of notes to our finals...bet you can't guess what my notes consist of now - remind me to pick up a magnifying glass []

[mikebse2a3]

Thanks again Mike[]

I thought I would show the Vertical plot for the 12907Hz to15966Hz range also. This is very good performance to me in both the Horizontal and Vertical considering these plots are nearly at the extreme top end of the Horns Range!

Roy didn't you design the K402 also?

mike tn[]

[DrWho]

Hey Mike, I got the alert about your phone call at 1am this morning and figured you probably didn't want me calling you back [] Maybe I'll try to catch you tomorrow night - I must say that I'm rather impressed with the polars - not perfect, but holy crap. I was honestly expecting a lot worse...so the real killer is - how much EQ is being implemented on the top-end?

One thing I've always wondered about is the audible effects of EQ at the top-end of a driver's response...every CD horn with CD EQ I've heard has always sounded better in PA applications, but a lot of people say that it sounds bad at home. But obviously you have some level of EQ and are having a blast. I guess the question is "how much is too much?"

[mikebse2a3]

Hey Mike, I got the alert about your phone call at 1am this morning and figured you probably didn't want me calling you back [] Maybe I'll try to catch you tomorrow night - I must say that I'm rather impressed with the polars - not perfect, but holy crap. I was honestly expecting a lot worse...so the real killer is - how much EQ is being implemented on the top-end?

One thing I've always wondered about is the audible effects of EQ at the top-end of a driver's response...every CD horn with CD EQ I've heard has always sounded better in PA applications, but a lot of people say that it sounds bad at home. But obviously you have some level of EQ and are having a blast. I guess the question is "how much is too much?"

Hey Mike

Roy can tell us for sure but from what I've seen the EQ processing is mostly due to the Horns Response Tapering Down and not because of the Driver itself having problems reaching high enough due to mass or other issues. The Highs are very clean and detailed in my opinion and I definitly don't feel a need for a Tweeter.

Call when you get a chance and I will give you some more info!

mike tn[]

[Al Klappenberger]

I would really like to know if that horn is of the CD type. It may explain partly why the passive network Roy posted for it looks like it does. If you totally remove the two R-L-C shunt netwroks from the high section of the filter you get an decrease of attenuation from about 14 dB to zero dB as you move up in frequency to 20 KHz. The increase is not linear though. The two R-L-C networks totally mess it up. These results are based on an 8 Ohm resistive load. I don't know if the driver is 8 Ohms or not. I base that on the analysis programs ability to "tune" on a part value while looking at the response in real time. 4 Ohms simply looked best. I doubt this becasue no instructions were give to Richard (Coytee) to set any CD correction into his active processor. To my knowledge he is not using any.

Al K.

[Deang]

It's not a CD horn, it's a Tractrix. "Advanced" phase plug or not, it starts to dive pretty much around the same place all high quality 2" exit drivers do.

[Al Klappenberger]

Dean,

If it's not a CD horn then the driver has a major mass-roll off problem or the network is wrong. You don't have to guess which theory I believe in! I am no horn expert but 14 dB at 500 Hz makes no sense to me!

Al K.

[DrWho]

lol, the newtork most certainly isn't wrong. And it's not a mass roll-off issue either.

The mass roll-off will be occuring at the top corner of the frequency response - around 18kHz.

The reason for the raw frequency response of the driver has to do with the horn maintaining a flat power response - same thing as avoiding HF beaming. This is actually reality that exists with every kind of driver, horn and every other topology and is a concept that even extends beyond the limited realm of acoustics. You can "never" have both be flat at the same time and it should not be confused as a "flaw" - it is a design choice.

Older large-format compression drivers sucked because the size and shape of the driver forced beaming at the high-frequencies and there was no way around it (even with uber cool horn designs). The beauty of the K-69 driver is that it is implementing a very cool phase plug that eliminates the otherwise natural beaming. The ideas behind the phase plug aren't exactly new, and are far from creative BS. It just wasn't until recently that these plugs could be manufactured on an accurate and large enough scale.

So with new advancements in manufacturing technology, the K-402 + K69 combo is finally able to achieve a true flat power response. And to compensate for the corresponding natural roll-off, the bottom end of the HF section is attenuated to also achieve a flat on-axis response. So now we have both flat frequency AND flat power response at the same time! [H]

It's not an entirely free lunch, but I'll get back to that later (unless someone else chimes in). I gotta run to class.

[Deang]

I'm not sure about Al's comment either. It would be nice if Roy would explain some of the things Al is seeing and doesn't understand -- if not for Al, but for the rest of us. I only have an intermediate understanding of this stuff, but I do know there is more than one way to skin a cat so to speak. It's not unusual to see one person's solution to be seen by another as a kluge -- I see it all the time all over the net. Now, since Al isn't getting any explanations he's forced to draw conclusions based strictly on his own experience -- something I know he wanted to avoid from the beginning. Since none of us here know what Roy's method/thinking/whatever is with those networks -- and can't adequately offer up the explanations on Roy's behalf -- it might be best to avoid any dogmatic opinions either way.

I don't think Al deserves to be slammed any more than Roy does. I do think it's fair to point out that if Al's way of thinking is wrong - it sure begs the question as to why Coytee is so determined to have Al build him a set of ESN's for his Jubilees. He obviously loved them in his Klipschorns. When Mark went down to pick up those same Klipschorns after buying them, and got to hear them before taking them home -- ordered a set of ESN's from Al within two weeks. Do I really need to point out that people around here have been unloading all manner of Klipsch networks for years in favor of his designs. You don't need an anechoic chamber to build a great sounding network.

Al's my friend, but I really don't want to take sides. I'd love to see a collaboration that could possibly advance the state of the art in horn loudspeaker design. I don't think Al will ever admit his networks aren't perfect -- but the thought of hearing them fully optimized using all applicable parameters is, well -- exciting. Because all of the data is intellectual property that belongs to Klipsch, and because Roy works for Klipsch -- I can certainly understand why Roy doesn't want to engage with Al. What a shame.

The mass roll-off will be occuring at the top corner of the frequency response - around 18kHz.

Call me 'Doubting Dean' -- but I don't believe that driver does much beyond 15kHz.

The reason for the raw frequency response of the driver has to do with maintaining a flat power response - same thing as avoiding HF beaming.

That would be a performance parameter determined by the horn, not the driver. Am I really wrong about this?

So with new advancements in manufacturing technology, the K-402 (basically a CD horn) + K69 combo is finally able to achieve a true flat power response.

The K-402 is adverstised as being a Tractrix. Also, it doesn't have the chacteristic slot at the throat section you normally see with CD types.

.http://www.klipsch.com/media/Products/KPT-JUBILEE-535.pdf

And to compensate for the corresponding natural roll-off, the bottom end of the HF section is attenuated to also achieve a flat on-axis response.

This seems to contradict your initial statement, "The mass roll-off will be occuring at the top corner of the frequency response..." IOW's, I'm a dunce and don't understand what you are saying here. It sounds like you are saying the LOWER end of the driver's response is attenuated to maintain flat on-axis response at the TOP of the driver's range.

[WMcD]

I'm basically in Al.K's camp. But I'm not quite sure what is in the camp, entirely.

In my view, the horn is technically pyrimidal. That means the cross section is conical. It does not have a matching section at the throat.

Klipsch calls it tractrix. But that could mean anything. If you play with the tractrix curve you can pick a portion and it will be close to just about any one of the classic curves. Or combinations of them. It is sort of like the curved drawing implement called the French curve.

Back to the issue. A conical horn has a throat impedance which is rising, it doesn't come up to flat level like an exponential. On the other hand, it does have a very good directional characteristic. Like I always say, you have to read Whats So Sacred About Directional Horns. It takes a good matching section to get the CD horn present a relatively flat load.

I have never seen a driver which has a flat response 400 to 15000 Hz on a plane wave tube, which would show its true power output. It is typically an on exponential horn which makes up for some of the roll off, by beaming (controlled directivity).

If you look at the Klispch data for the 402 HF, it is said to be fairly flat output over a wide range. But look at the footnote. That is behind a screen with equalization.

The polar diagrams put up above just show directivity. I'm sure the response has been equalized so all the frequencies look similar in magnitude.

Harping like it do, look at Keele's comments on testing of his CD horns which have good loading because of the matching section. He has to equalize the driver on a plane wave tube to get it (the driver) to have a constant output. When wide band power is put into Keele's horns, wide band power over a constant directivity plot comes out.

In my view, Al's only mis-statement is that the mass roll off is very high. I'll bet it is at about 3000 Hz. He may be right that the upper limit is the point where things really go to heck even with equalization.

Mike is correct in a way to point out that phase plugs help. But I think they only help phase problems.

My guess re equalization is as follows. The driver is going to need a boost at the low end because the horn load is so poor. It may or may not need a boost at the high end. This is because the horn load come up to a level to make the driver happy. This increasing load could balance the mass rolloff. But I'm thinking there could be a need for a boost at the high end too.

Of course this could be accomplished with the passive network to some extent. Passives can't boost output very much. But you can set things up so what needs boost is zero gain and the other points are lossy.

Gil

Gil

[bodcaw boy]

Thanks Roy!

I must confess never having seen such elaborate plots before which makes it hard to compare against other speakers I've heard and seen the measurements for, but after a very quick browse (basically looking at each page for a second) things look rather impressive.

One thing I've always wondered about is why the polars aren't perfectly symmetrical - especially with the rear polars. Is this the product of assymetries in the anechoic chamber or issues with noise floor? Rounding errors with the program? Wobbling of the speaker on the turntable? That kid in the dominian republic? It's more of an academic interest because it shouldn't really effect how one interprets the data.

And how can the on-axis response be lower than just slightly off-axis at some frequencies? That blows my mind - doesn't that imply a planar wave front?

I think we need to start a "get doc some jubilees fund" - or maybe I need to get one of those cardboard signs with strings: "will work for jubilees" []

hey doc,

back in town for a couple of days then out for about 1 week so i am catching up on forum.

the assymetric data at the rear of the speaker is mostly due to noise floor. the best way to know if table and setup are correct is see if the data points at 0 degrees and 360 degrees match up. that was inserted in our programs to make sure that our setups are good.

yes it can imply a plane wave but also when the on axis spl is not the highest spl you can also be on the verge of the wave "wrapping" onto itself. i have a proposed solution for that right now that i am proving out. so far so good and it may provide us with the opportunity to have horns control at wide coverage angles, something like 80 x80 or wider. of course as the data so far looks good, so do the chances of a patent.

i can start a fund and be president of the "doc needs a set of jubs" telethon....

berryboy roy

[bodcaw boy]

Thanks again Mike[]

I thought I would show the Vertical plot for the 12907Hz to15966Hz range also. This is very good performance to me in both the Horizontal and Vertical considering these plots are nearly at the extreme top end of the Horns Range!

Roy didn't you design the K402 also?

mike tn[]

hey mike tn,

yes i did. i am fortunate to have designed most of klipsch's horns. this is the kind of data that drives me to figure out how acoustic waves behave and how they should behave. i can get a little obsessive about it but oh well.

berryboy roy

[bodcaw boy]

Hey Mike, I got the alert about your phone call at 1am this morning and figured you probably didn't want me calling you back [] Maybe I'll try to catch you tomorrow night - I must say that I'm rather impressed with the polars - not perfect, but holy crap. I was honestly expecting a lot worse...so the real killer is - how much EQ is being implemented on the top-end?

One thing I've always wondered about is the audible effects of EQ at the top-end of a driver's response...every CD horn with CD EQ I've heard has always sounded better in PA applications, but a lot of people say that it sounds bad at home. But obviously you have some level of EQ and are having a blast. I guess the question is "how much is too much?"

hey doc,

how much is too much? well it has to be quite a bit. in a theater, you have some high freq transmission loss just cause of the volume of air in a good sized theater and then of course you have the screen. the screen is micro perfed but there is still quite a bit of attenuation. i have seen eq boosts on the order of 6 to 12 db starting at around 5k to 20k. that is why a 3" diap is the preferred choice. at 15k the 3" diap is still moving less than a 1" and the larger voice coil gives you the ability to apply more power.

oh and too much, the diap will let you know. it sounds very slurred and most likely is on the edge of going to compression driver heaven.

berryboy roy