h66000 h9900 h4365

[4313B]

A member asked for the raw data curves for the horn/driver combinations in the Everest II, K2-S9900 and 4365 so here they are.

[4313B]

Guido, here's what I got in LEAP substituting in a 1500AL for a 1501AL. These are starting values with optimization set at the knee of the curve so that the proper crossover frequencies are maintained. This results in a slight decrease in voltage drive through the passband, as you mentioned.

[4313B]

The dark curves are the target curves using a 1501AL and the light curves are the result of the above values along with the 1500AL. LEAP wanted to put in less shunt resistance so I put in the values I wanted and took them out of the optimization. The values LEAP wanted took too much of a hit in efficiency. (e.g. 18 ohms on LF2 instead of my chosen value of 33 ohms. Interestingly, when I went over this years ago using BB6P it also wanted to put 12 to 18 ohms across that woofer. My opinion is that's too much of a cut.

[4313B]

Here are the impedance curves. The actual Everest II 1501AL curve is the light one.

[4313B]

Ok, here is a better set of curves. The light gray curve is the 1501AL in the Everest II box. The blue curve is a 1500AL in my personal 5 cu ft (gross) @ 28 Hz box and the red curve is a 1501AL in my personal 5 cu ft (gross) @ 28 Hz box. When I have time I'll run the simulations on these last two curves instead and see what values LEAP comes up with. I don't have time to play with it further tonight.

You probably don't want the dual drivers dipping below four ohms so you might need to use slightly higher DCR in your coils to make sure that doesn't happen?

[Guido]

Thanks for this awesome work Giskard!

I'll have a closer look this weekend and add what I find (or not :o:).
I also work on it but your input is higly appreciated.

[Guido]

We also need to look into the sensitivity thing, right?

The 1500AL will be more efficient and therefore the whole midrange need to be adapted. Or am I wrong here?

[4313B]

I'm taking your drivers to the crate for shipment to Germany this weekend. That's a ten hour day.

There's maybe a half dB sensitivity difference between the drivers. My concern is the dB drop in the passband for both drivers. They might be a dB hot at the crossover frequency. I'll work on it a bit more even though I am very probably not going to end up running the Everest II type bottom end.

I'm contemplating running the first 1500AL in a smallish 2.5 cu ft box, possibly sealed. I'll run the second 1500AL as a dedicated subwoofer as per Greg's 1200 Array system, which I prefer to the Everest II. I might end up replacing the second set of 1500AL's with W1500H's. They take less enclosure volume to maintain the same bandwidth. Basically half the cubic footage requirement of the Everest II/4350/4355/4435.

[Guido]

There's maybe a half dB sensitivity difference between the drivers.

Hmmm, I mean the fact that with a given 2.83V input the 8 ohm 1500AL draw more current than the 12 ohm 1501AL and so the difference is 2-3dB.
I attached your 1500AL / 15001AL T/S comparison

Or am I completely wrong here?

[4313B]

Oh I see what you're asking... do I need to increase the output of the 476BE by something like three dB? I'll figure it all out when I have time.

I do remember that Greg set the optimization curve for the dual 1501AL's at 96.5 dB... I'll check and see what he set the optimization curve to in the K2-S9900 using the 1500AL-1.

[timc]

Guido.

Yes you will have to increase the impedance of the HF driver, thus reducing the voltage sensitivity. You do this by simply adding a resistor in series with the driver.

[4313B]

Guido.

Yes you will have to increase the impedance of the HF driver, thus reducing the voltage sensitivity. You do this by simply adding a resistor in series with the driver.And then reworking the whole high pass network. Let me know how that works out for you guys. :)

[4313B]

I also work on it but your input is higly appreciated.I played around with it a bit more this afternoon. Greg has 1.3 ohms of DCR on that 20 mH coil on the helper woofer. Halving that has an impact on the voltage drive of the helper woofer. Any change in box volume and or tuning is also going to have an impact.

So... basically you can start with the schematic I posted above and then tweak it to your specific drivers/boxes/fiberfill/tuning/coil DCRs.

BTW, here's a run using 1500AL's before Greg and Jerry came up with the 1501AL's instead:

[richluvsound]

D,G

Its wonderful to see a thread of real content .Two gents I admire and respect working on an exciting project and taking the time to post about it .

thanks , Merry Christmas both and all

Rich

[timc]

And then reworking the whole high pass network. Let me know how that works out for you guys. :)


The reworking when increasing the R component only, is purely linear. Meaning you can just increase/decrease the component values by the % amount of the impedance change. Not sure if this is also true for notches.

I'm apparantly no gent :p

[Guido]

Guido.

Yes you will have to increase the impedance of the HF driver, thus reducing the voltage sensitivity. You do this by simply adding a resistor in series with the driver.

Hi Tim,

if we decrease woofer impedance we also need to DECREASE midrange impedance. Not increase.
BUT I think this isn't the right way. I confess that I didn't work on it till now as I have to recone 6 woofers till christmas and finish the cc crossovers for a forum member.
I'll look into at as Giskard does. He or me or we will come up with a solution for sure.

It's a bit of a luxury problem how to substitute 4 1501AL with 4 1500AL in an EV 2 configuration but the task itself should be rather interesting... I guess :o:

To prevent this thread to get too dry I added a pic of the "veneer test" for the EV2 cabinets. Will post more pics.

[Guido]

Its wonderful to see a thread of real content .

Be careful, you might get flamed ;)

[4313B]

The reworking when increasing the R component only, is purely linear. Meaning you can just increase/decrease the component values by the % amount of the impedance change. Not sure if this is also true for notches.I'd prefer to wait until I have time to look it all over in a bit more detail...










Ok... I've looked over everything in a bit more detail ;) and based on the engineering specs it would appear that the original 1500AL is 2 dB more efficient than the 1500AL-1 and 1501AL. Given that, the HF needs to be raised 2 dB. Decrease R8 from 9.1 ohms to 6.8 ohms (or perhaps 6.2 ohms which is worth another half dB) and let me know how that works for you. If you have to go lower than 6.2 ohms then the high pass should technically be reworked. Keep in mind that your coil DCR's are going to affect sensitivity. JBL had 1.3 ohms of DCR on the low 1501AL and 0.8 ohms of DCR on the high 1501AL. You are very probably going to have to use less. I wouldn't be surprised if the HF needs to be reworked a bit, primarily in the 700 to 1 kHz region. I wouldn't be surprised if the LF's end up needing tweaking too. You should be able to end up with a very nice sounding system.

So replacing the 1501AL's with 1500AL's should increase system sensitivity from 96 dB for 2.83V @ 1m to 98 dB for 2.83V @ 1m.

Note that the 1500AL (and K2-S9800) preceeded LEAP. Everest II was the first system that was done with LEAP. The point being that I don't have a LEAP file on the 1500AL/K2-S9800.

Be aware that this is based on the Everest II horn...

Good thing my eighteen month old daughter got me up at five thirty this morning or I wouldn't have had time to look over this today.

[timc]

Hi Tim,

if we decrease woofer impedance we also need to DECREASE midrange impedance. Not increase.

Then i have misread something. But the argument stands. Then you do the opposite, by putting a resistor in paralell :)

[4313B]

Here is the schematic we will be working from:

[timc]

Then i have misread something. But the argument stands. Then you do the opposite, by putting a resistor in paralell :)


Forget it, my head is obviously not working today..... :o:

[4313B]

Removing L14 - 0.5 mH 0.4 ohm shunt inductor

EDIT : Was intended to be the first post in a series following the "What does is all do" post visually showing what each section of the filter affected.

[timc]

Here is the schematic we will be working from:


C7/L8 is a second order highpass section.

C13/R8 is the Constant directivity compensation. This is the R to change to get sensitivity right i think. You will also have to change the capacitor.

L6/R9 i THINK is an addition to the CD compensation

The next three RLC's are notches to flatten the response in the passband.

L14 alters the whole impedance down. Size this to get the same impedance as the woofer in the crossover region. This one only changes things slightly and can not be to small. Then it would alter the response curve in the passband.

[4313B]

C7/L8 is a second order highpass section.

C13/R8 is the Constant directivity compensation. This is the R to change to get sensitivity right i think. You will also have to change the capacitor.

L6/R9 i THINK is an addition to the CD compensation

The next three RLC's are notches to flatten the response in the passband.

L14 alters the whole impedance down. Size this to get the same impedance as the woofer in the crossover region. This one only changes things slightly and can not be to small. Then it would alter the response curve in the passband.Awesome! You just saved me from wasting my life's time graphing all this stuff and posting it for everyone to visualize. :)

[timc]

To be honest, I think the quickest way to do this modification is to start from scratch.

First put in the CD compensation, and tilt the curve as you like. Then add the highpass section. The RL part just after the compensation might be needed, or it might not. It is to adjust the impedance i just realized. The resistor in the CD compensation will to a large degree give you the final nominal impedance.

Then add all the notches you require.

Finally use the last inductor to get the impedance juuuuust right.

[4313B]

To be honest, I think the quickest way to do this modification is to start from scratch.Yep, so you guys stop bothering me with PMs. Figure this out yourselves. I have networks to build and have neither the time or inclination to get involved in your issues. Everytime I get sidetracked with these issues someone sends me a PM asking if their networks are done yet.

[timc]

?????????

What did i do?

[4313B]

?????????

What did i do?I'm sorry. I should thank you. You saved me from feeling compelled to post a dozen or so graphs.

I think you have an excellent handle on the situation so I'd prefer not to be involved.

[timc]

Ok, I see your point.

For all others who might be wondering about the "correct" crossover I'll say this:

There is no such thing really. All topologies, slope charateristics, and whatever, have their strengths and flaws. Whatever floats your boat is correct for YOU.

The key to making a good crossover, without spending lots of $$$$$ is to do a good measurement of the drivers in the intended cabinet. Programs like LEAP lets you estimate the measurements to some degree.

Than smash the data into a simulation program. LSPcad tiral version is nice, and free. Then play around with the choosen topology until you get what you want. It's not really much more to it.

A basic understanding of how things work is, as always, a good thing :)

[4313B]

Ok, I see your point.

For all others who might be wondering about the "correct" crossover I'll say this:

There is no such thing really. All topologies, slope charateristics, and whatever, have their strengths and flaws. Whatever floats your boat is correct for YOU.

The key to making a good crossover, without spending lots of $$$$$ is to do a good measurement of the drivers in the intended cabinet. Programs like LEAP lets you estimate the measurements to some degree.

Than smash the data into a simulation program. LSPcad tiral version is nice, and free. Then play around with the choosen topology until you get what you want. It's not really much more to it.

A basic understanding of how things work is, as always, a good thing :)A perfect quote to end this thread on...

As for trying to maintain some semblence of Everest II voicing? Whatever...

Thanks for disengaging me from this task. :yes:

[timc]

As for trying to maintain some semblence of Everest II voicing? Whatever...

If that is the goal (as in this case), then "transpose" the filter to the impedance needed.

I'll gladly help if I can get the frequency and impedance files.

Edit: My comment about taste, was purly general. I can see no reason why you would want to change a S9900/DD66000 voicing.

[4313B]

I'll gladly help if I can get the frequency and impedance files.Get them from Guido for his specific application.

It's going to be tough to get good measurements down low enough to be of any use. You're dealing with crossover frequencies at ~150 Hz and ~700 Hz. The 150 Hz crossover can be adjusted by ear for the intended room. The 700 Hz crossover frequency could be tough, maybe a ground plane measurement on a tennis court or gymnasium floor. Maybe near-field and far-field measurements and then a minimum phase transform run on the spliced frequency response curve. You'll figure it out I'm sure.

Good luck. It should be fun.

[timc]

Ah...I thought the curves where already available.

I have the facilities to do anechoic measurements down to around 150Hz, but i don't have the drivers.

[Guido]

Why can't we just work it out together?
Why beating starts again?

Giskard, take your time please. I have to finish other works too but in between I play a bit with the EV2-1500AL thing. It is an interesting task and it's fun to play with.

[4313B]

Why can't we just work it out together?
Why beating starts again?

Giskard, take your time please. I have to finish other works too but in between I play a bit with the EV2-1500AL thing. It is an interesting task and it's fun to play with.Starting with post 18 I was going to visually break down the high pass step by step but with every post I made I was being interrupted. I think you get the gist of it though so the exercise was probably moot anyway. You'll figure out if and how you need to manipulate any or all of the notch filters given your specific horn build. If you are using the H4338 horn as your base things will be different. I have no idea how you are building your horn since you haven't divulged that information.

I showed you how to easily affect the sensitivity but there is only roughly 3 dB of leeway in the network, after that other components have to be adjusted for greater sensitivity gain should it be required. If you end up using real low DCR coils on the low pass you might have to adjust for that.

[timc]

Aha...so its me to blame is it?

Well, sorry for showing some enthusiasm. It won't happen again, I promise.

[4313B]

Aha...so its me to blame is it?

Well, sorry for showing some enthusiasm. It won't happen again, I promise.No, it's my fault. This all should have been done between Guido and myself on a private forum. You and I are on two completely different pages with this. You're right from the perspective of building a proper network for a new system. I'm dealing with existing parameters. Building new networks from scratch is neato and all that but it completely destroys the voicing of the system as designed. This isn't going to be right on but it should be pretty close.

And as for interrupting my stride, that's what I get for posting in realtime. I should have had all my posts premade and then dumped them all into the thread at once... I apologize for getting aggravated.

[Guido]

I have no idea how you are building your horn since you haven't divulged that information.

I have a horn that is a combination of conical and exponential design, as the H9800 and H4338 are. I would say it is VERY close to the original EV2 horn. I can't take measurements yet as the horn is part of the cabinet and the cabinet isn't done yet.
At present I try to work out the starting networks with LEAP and your help. When simulations are carefully done with LEAP one don't need to tweak much later, that's my experience.
The networks and drivers will hopepully find their way into the cabinets in spring and then I'll listen, measure, tweak, listen ........ and of course post.

I'll open LEAP now :hyp:

[4313B]

If that is the goal (as in this case), then "transpose" the filter to the impedance needed.The 1500AL (8 ohm driver) engineering standard has a dcr of 5.3 ohms, a minimum impedance of 7.5 ohms and a motional impedance of 130 ohms.

The 1501AL (12 ohm driver) engineering standard has a dcr of 9.2 ohms, a minimum impedance of 11.7 ohms and a motional impedance of 300 ohms.

You can transpose the following schematic if you wish. I believe that our problem is with the motional impedance. I will post a normalized set of impedance curves momentarily to highlight my point.

[richluvsound]

Ah...I thought the curves where already available.

I have the facilities to do anechoic measurements down to around 150Hz, but i don't have the drivers.

If you havent got the drivers then please leave the thread alone .... what you offer is of no PRACTICAL
use to those that are actually doing this ..... YOU MIGHT JUST LEARN SOMETHING . It has been stated so many times that theory only gets you 50 % there .... Simulations are just simulations, like comparing masturbation to a real good session with a hot partner ......:bouncy:

[4313B]

I will post a normalized set of impedance curves momentarily to highlight my point.Here's the graph.

Transposing works well enough in the 100 Hz to 1 kHz range. The delta is 4.6 ohms with the two specific drivers I have. There's more coil inductance in the 1501AL (it does indeed have more coil) and it has a higher motional impedance and those are our differences to deal with.

Now, in Greg's LEAP file the impedance guide curve tops out at 75 ohms, not at nearly 135 ohms as shown here... Reference the gray curve in Post #4 (http://www.audioheritage.org/vbulletin/showthread.php?29853-h66000-h9900-h4365&p=301425&viewfull=1#post301425)

[4313B]

I have a horn that is a combination of conical and exponential design, as the H9800 and H4338 are. I would say it is VERY close to the original EV2 horn. I can't take measurements yet as the horn is part of the cabinet and the cabinet isn't done yet.Ok. I've noticed that the H4338 horn had more bottom end than the Everest II horn so you might have a problem with the stock Everest II network if your horn ends up similar to the H4338. If I remember correctly, the H9800 doesn't have quite as much bottom end as the H4338.

As an aside, the H4365 is a cosmetically altered H4338, they both peform similarly.

[Guido]

Here are the resulting graphs for crossover mod post #2

2-3 dB higher sensitivity with the 1500AL's
Imp goes down to 3.5 Ohms when keeping original coil DCR's

[Guido]

Here is my first hit, open for further refinement.

Impedance goes down to 4 ohms.
Sensitivity isn't that much higher for the 1500AL's in the crossover region. But there is too much juice under 200 Hz.
What to do?
Maybe better use the Giskard mod post #2 and an increased MF sensitivity? But whats with the UHF?

[Guido]

Giskard,

can we get an impedance file for the 476Be on the H66000?
That would be great to modify the MF part.

[Guido]

Not too bad

Pink is LEAP simulation of original EV2 (w/o UHF)
Blue is 1500AL mod giskard post #2 plus MF with 7.5 ohm series resistor and 0.4uF bypass.

What do you think?

[4313B]

Giskard,

can we get an impedance file for the 476Be on the H66000?
That would be great to modify the MF part.Ok.

[4313B]

Not too bad

Pink is LEAP simulation of original EV2 (w/o UHF)
Blue is 1500AL mod giskard post #2 plus MF with 7.5 ohm series resistor and 0.4uF bypass.

What do you think?Same as you, not too bad.

You'll need to sub in your own measurements at some point though.

You also might want to do as Tim suggested and simply transpose the values instead of using LEAP's optimization function like I did. Multiply the capacitors by 1.65 and the inductors and resistors by the inverse. What I did was copy the "system curve" voltage drive generated by the stock network over into the "guide curve" section and then used that curve as the optimization curve after substituting in the 1500AL impedance. I changed the acceptable dB deviation from the default 1 dB to 0.1 dB and obtained the results shown in post #2. In the end it is going to boil down to your specific impedance curves in your specific boxes along with your specific coil DCR's.

Here are my 1500AL and 1501AL impedance curves measured in the box shown. 5.0 cu ft @ 28 Hz

[cooky1257]

Thanks guys.

To the innocent bystander this is still very interesting and educational thread-even at the basic level, the left to right sequential treatment of compensation then notch eq's has for the first time allowed this thick head to understand a xover diagram:).
Of the 3 horns in question the H4365 is one I'm not familiar with-is this the modified one closest to the SAM1?.

Cooky

[4313B]

Used in the new $9,500 monitor Greg did right after the $22,000 K2-S9900. The LF transducer has a cloth surround like in the good old days.

http://www.harman-japan.co.jp/jbl/hifi/4365/



Greg is a serious advocate for the vertical horns and the SAM1/1400 Array horn is that horn of choice.

[4313B]

To the innocent bystander this is still very interesting and educational thread-even at the basic level, the left to right sequential treatment of compensation then notch eq's has for the first time allowed this thick head to understand a xover diagram:).Continuing on from post #22.

Removing the L10 / C9 / R14 notch filter:

[4313B]

Removing the L9 / C8 / R11 notch filter:

[4313B]

Removing the L7 / C6 / R10 notch filter:

[4313B]

Removing R9 / L6:

[cooky1257]

Used in the new $9,500 monitor Greg did right after the $22,000 K2-S9900. The LF transducer has a cloth surround like in the good old days.

http://www.harman-japan.co.jp/jbl/hifi/4365/



Greg is a serious advocate for the vertical horns and the SAM1/1400 Array horn is that horn of choice.

Thanks Giskard.
Cooky

[4313B]

Guido, don't forget about the switchable 120 ohm resistors Greg added in on the 1501AL's for a bit of flexibility.

You'd probably want something on the order of 75 ohms for the 1500AL's.

You also asked about the 045Be... the voltage drive should tell the tell. It's a very nice driver though.

[4313B]

Thanks Giskard.
CookyYou'll notice that the H4338 is horizontal in the photo of the box above? Let's just say that Greg isn't impressed. :rotfl:

But I wanted a couple of "Studio Monitor" boxes with blue baffles and a 1500AL & 476BE/H4338 two-way was a good excuse to make a pair.

I'll do the vertical thing too someday.

[cooky1257]

You'll notice that the H4338 is horizontal in the photo of the box above? Let's just say that Greg isn't impressed. :rotfl:

But I wanted a couple of "Studio Monitor" boxes with blue baffles and a 1500AL & 476BE/H4338 two-way was a good excuse to make a pair.

I'll do the vertical thing too someday.

They do look real nice and compact though, he'd have to concede the vertical horn is a challenging look for some:D

[Guido]

Guido, don't forget about the switchable 120 ohm resistors Greg added in on the 1501AL's for a bit of flexibility.

You'd probably want something on the order of 75 ohms for the 1500AL's.

You also asked about the 045Be... the voltage drive should tell the tell. It's a very nice driver though.

You are bloody fast. I will be able to look into it tomorrow or wednesday. I'll do a bit more "Leap"ing.
Thanks for the real DCR values. These are not in the tech sheet. My guess of the coil DCR for the high woofer and some notches where spot on I see.

I hear the woofers are on the way :D Thanks!

[Guido]

Ok.

Attached Files

http://www.audioheritage.org/vbulletin/images/attach/txt.gif HF_Impedance.txt (http://www.audioheritage.org/vbulletin/attachment.php?attachmentid=49062&d=1292724132) (9.9 KB, 9 views)
http://www.audioheritage.org/vbulletin/images/attach/txt.gif HF_FMP_corrected.txt (http://www.audioheritage.org/vbulletin/attachment.php?attachmentid=49063&d=1292724133) (19.5 KB, 6 views)
http://www.audioheritage.org/vbulletin/images/attach/txt.gif LF_FMP.txt (http://www.audioheritage.org/vbulletin/attachment.php?attachmentid=49064&d=1292724134) (20.4 KB, 6 views)
http://www.audioheritage.org/vbulletin/images/attach/txt.gif LF_Impedance.txt (http://www.audioheritage.org/vbulletin/attachment.php?attachmentid=49065&d=1292724134) (9.9 KB, 5 views)


What is an FMP file??

[pos]

You'll notice that the H4338 is horizontal in the photo of the box above? Let's just say that Greg isn't impressed. :rotfl:
Did he already turn his DD66000s on their side?

[grumpy]

What is an FMP file??
I assume you discovered FMP = Frequency, Magnitude, Phase info in text format ...
(basic info that can often be exchanged between speaker design/measurement programs)

[Guido]

I assume you discovered FMP = Frequency, Magnitude, Phase info in text format ...
(basic info that can often be exchanged between speaker design/measurement programs)

Thank you, I found out meanwhile.
Currently lost into LEAP.
.....will come back

Merry Christmas

[grumpy]

Have fun while you're in there :)
Merry Christmas.

[Guido]

Thanks for this information about the GT active crossover.
I'll have a look at it when I find the time.
The last weeks I tried to refine the passive network in LEAP. I'm done at the moment and can't continue until I have the real thing here. Looks like a 99dB design and impedance is not lower than 3.6 ohms. The implementation of the 045Be isn't that problematic. The UHF will slightly roll off to 3 dB down between 20 kHz and 50 kHz. I'm not really worried about that :p

I'm sure your 1500AL/476Be/H4338 design sound wonderful. I already have a 1x1500AL + 2435 + H4338 design since 3 years now, remember?

This time the 2 woofer setup is a given as the EV2 cabinets are already in production.

[Guido]

I want to refresh this thread as my EV2 cabinets will be here in september/october and I need to lift my a$$ now and build prototype networks to be refined in use with the actual speaker.

Giskard, when converting the schematics from 1501AL to 1500AL, what DCR's did you use?
E.g. the two 10mH subwoofer coils together are 1.3 ohms in the original network.

The mod network use only one 10mH coil. 0.65 ohms? 1.3 ohms?
Same for all othe coils.

What do you think

[bubbleboy76]

A member asked for the raw data curves for the horn/driver combinations in the Everest II, K2-S9900 and 4365 so here they are.

Interesting to see how "unperfect" the crossover slopes looks on these high-end systems.
Is it a limitation in passive design?
Or am I missing something?
The cheapest 4365 looks the best to me here.