L112/96 Network suggestions

[Nightbrace]

Just got a set of L112's today, they sounded good, but overall I actually preferred my L36's.

Has anyone tried using a L100A network in these speakers? What about the network from the 4313?

In a direct A/B compariosn with my L36's I noticed a few inadequacies worth noting. Overall the bass was a bit better in the L112's/ And despite having virtually the same midrange and much better tweeters, they are not as detailed in the midrange, and are nearly equal in the highs, although the L112 did go a bit higher, overall it wasn't as pleasing and sounded a bit muffled. I don't think the efficiency of the drivers is that much different, so it leads me to believe the crossover is the weak-link. There is a lot of potential with the drivers in these speakers. Any advice would be appreciated.

[4313B]

so it leads me to believe the crossover is the weak-link.D'oh! That would be your opinion as opposed to fact.


What about the network from the 4313?
You've been reading my old posts - I ended up running the 3113B in every L96 I ended up owning.

[Nightbrace]

J, it leads me to believe the crossover is the weak-link. .

Again, I am mislead, I should start reading some of your old posts, thats what I was considering too, but where to find schematics for the 3113B crossovers? Would these work just as well in the L112?

[4313B]

I should start reading some of your old posts, thats what I was considering tooWhy would you consider the 4313 network? What do you know about the 4313? I find it hard to believe anyone would just happen to think a 4313 network would work, especially if they don't even know where to find the schematic.

http://www.jblproservice.com/navigation/Network%20Schematics.html

[Nightbrace]

Its was just a thought as the 4313B has the same mids and tweeters as the L112's, and generally I have been more impressed with the pro-sound crossovers. Plus I was considering these for my EN3's till I abandoned them, figured it would be cheaper to get a set of L112's as I was planning on using all the drivers anyways :)/

[Regis]

D'oh! That would be your opinion as opposed to fact.


You've been reading my old posts - I ended up running the 3113B in every L96 I ended up owning.

Last Saturday, I just installed the 3113b in my bud's L-150's (the early ones, not the "A's" with the same crossover as the L-112). The sound was dramatically improved over the stock crossover. Better detail, separation and the bass was less muddy. Good stuff! It isn't hard to convert the L-112 to the 3113b. I used the 3.0 uF GR research cap in parallel to modify the big cap to (I believe) a 16.5 uF value. Do the bypass mod. Easy if you have basic soldering skills.

[DavidF]

Again, I am mislead, I should start reading some of your old posts, thats what I was considering too, but where to find schematics for the 3113B crossovers? Would these work just as well in the L112?

I have heard a stock L112 xover and a mock-up of the 4313 using the same L112 drivers. I suspect you will find the overall response characteristics similar. So, if you prefer the stock L36 over the stock L112, you will not find the conversion to the 4313 more in line with your preferences.



One long term idea for the fun of it is to "force" yourself to live with the L112s for good, long, while. Get accustomed to the different frequency balance of the L112. Then, switch back to the L36. If you still feel that the L112 is inadequate, then you know for sure your preference is for more energy in the mid-to-brilliant spectrum. If so, sell the L112s , then go look for some of the pro model 4312 models as an upgrade to the L36.



David F

[Nightbrace]

I'm not a preference really, its just that the L112's sound a bit tame to me. I loved the 4312's I heard, but I got the L36's for $150 in MINT condition, so it was a no-brainer. The difference wasn't that much, but the bass is a bit lacking in the L36, similiar to the XPL160 I recently heard. The drivers are BETTER in the L112, so how can they be not as detailed as the L36's? Thats what I don't understand. The crossovers can't take away that much. Was the crossover design a measure taken by JBL to tame the drivers down, esp. the tweeter down a bit?

[Nightbrace]

your preference is for more energy in the mid-to-brilliant spectrum. David F

Of course thats what I want, shouldn't everybody? I also want more detailed and more dynamic bass which the L36's don't deliver. Getting L100T's soon, yippee, they will be MUCH better than both, just thought the L112's would be worth a try. After reading a thread saying that these were the best speakers ever made by JBL a few weeks ago, I got myself a pair on the Bay and they finally got here., I wasn't horribly disappointed, but I was expecting a vast improvement over my L36's/

[Nightbrace]

BTW, another thing that I don't understand is how the L50 could sound so much different than the L36. The drivers are virtually the same. I liked the linear design of the drivers, but was overall not too impressed.

[briang]

I'm glad this thread came up. I'm considering reworking the crossovers in my L112s and L96s.:applaud:

[Nightbrace]

There's so much potential with these drivers, the 3113B seems like the way to go.

[4313B]

I'm considering reworking the crossovers in my L112s and L96s.:applaud:Biased (charge coupled) 3113B networks in the L96. Accept no substitute. You'll need to tweek the conjugate across the 128H for the L112.

[4313B]

BTW, another thing that I don't understand is how the L50 could sound so much different than the L36. The drivers are virtually the same. I liked the linear design of the drivers, but was overall not too impressed.It's all in the networks. I guess you like the earlier stuff.

[briang]

Biased (charge coupled) 3113B networks in the L96. Accept no substitute. You'll need to tweek the conjugate across the 128H for the L112.
Looking at the N96 vs the 3113B network, the difference appears to be the crossover point to the LE5-12 (16.5uf vs. 13.5uf) and the LE10H-1 gets a 2nd order (12dB) roll off instead of first order (6dB).

I'm really new at this and my electronic engineering class I took in college is now 14 years old (i.e. forgotten nearly everything), so feel free to correct me and instruct me (be nice).

Regarding the crossover point and your conjugate comments for the 128H, I assume that is because the 128H rolls off sooner than the LE10H-1, yes? No?

(I need to send Giskard a 6 pack of his favorite beer :bouncy: )

[Nightbrace]

Giskard, I like what sounds the best.

[4313B]

Looking at the N96 vs the 3113B network, the difference appears to be the crossover point to the LE5-12 (16.5uf vs. 13.5uf) and the LE10H-1 gets a 2nd order (12dB) roll off instead of first order (6dB).

Regarding the crossover point and your conjugate comments for the 128H, I assume that is because the 128H rolls off sooner than the LE10H-1, yes? No?
Just run an impedance curve on the 128H and compare it to the LE10H. Get the impedance of the 128H similar to the LE10H using the conjugate or else change the value of the series inductor.

Also note that the 4313B used the edgewound copper ribbon wire LE5-odd midrange while the L96, L112, and L150A used the round copper wire LE5-even. Since this is DIY, try whatever you have and see how it works for you. In this specific instance I think you might find that it doesn't much matter.

[4313B]

Giskard, I like what sounds the best.To you. And that's not a problem. I just don't want people going away from the forum thinking the L36 is "better" than the L50. Mark thought the L50 sounded better and that's why he designed it the way he did.

[Nightbrace]

Glad I'm not the only one who needs a class in Crossover Design 101. It just makes me wonder what JBL was thinking, especially with the L50.

[Nightbrace]

So you think the L50 sounds better? They are 2 completey different speakers sonically.

[4313B]

Yes, I preferred the smoother L50. I've always preferred the smoother JBL systems. A buddy of mine referred to them as "boring" - L50, L110, L212 - they were all boring to him. We both agreed though that the 4313 was killer. Whatever...

[Nightbrace]

At least we can both agree that they are completely different speakers sonically, most would not think so given the fact that the drivers are virtually identical.

The only thing I can think of as to why JBL messed so much with the crossovers, is that I recognize a slight resonance around 240 Hz with the L36, was this a reason why they changed the crossover?

[Nightbrace]

So its more of a preference you are saying, but aren't we all after accurate sound reproduction that is as true to the music as possible? To me the L36 is more "true" compared to the L50, but thats just my 2 cents, and I agree that we should not try to mislead people. The XPL160's I heard were even more true the mid-range flat out blew away everything I have ever heard. I am thinking of buying a set of L100T's, do you think I'll like these speakers at least as much as the L36's?

Has anyone tried the N36 crossovers in the L50?

[Nightbrace]

What are your thoughts about the L166, L200 or L65?

[John B]

This may be too obvious, but I’ll ask anyway. Have you tried turning up the mid and/or tweeter levels on the L112s?

[Nightbrace]

yep, but all it does is change the volume, it does nothing to change the crossover points. But to do a proper comparison they should both be at 0. I actually prefer my L36's at about +1-+2.

[duaneage]

Looking at the N96 vs the 3113B network, the difference appears to be the crossover point to the LE5-12 (16.5uf vs. 13.5uf) and the LE10H-1 gets a 2nd order (12dB) roll off instead of first order (6dB).

I'm really new at this and my electronic engineering class I took in college is now 14 years old (i.e. forgotten nearly everything), so feel free to correct me and instruct me (be nice).

Regarding the crossover point and your conjugate comments for the 128H, I assume that is because the 128H rolls off sooner than the LE10H-1, yes? No?

(I need to send Giskard a 6 pack of his favorite beer :bouncy: )

The N50 is doing a bit more than that. Compared to the N36, the LFRolls off a lot lower than the N36. The midrange gets a real bandpass instead of a simple highpass and there is a RC trap on the output of the filter to further roll off the driver. The midrange xover also has attenuation fixed in addition to the lpad control. This is a really thought out bandpass and very specialized to the LE-odd midrange charactoristics.

The 128H has a large voice coil and a steep rise in impedance above 2000Hz. taming that rising peak with a RC network will yield better results from the midrange since it won't interact as much. Predictable driver impedance that is flat and stable is easier to design for than drivers with swooping impedance curves.

[briang]

The N50 is doing a bit more than that. Compared to the N36, the LFRolls off a lot lower than the N36. The midrange gets a real bandpass instead of a simple highpass and there is a RC trap on the output of the filter to further roll off the driver. The midrange xover also has attenuation fixed in addition to the lpad control. This is a really thought out bandpass and very specialized to the LE-odd midrange charactoristics.

The 128H has a large voice coil and a steep rise in impedance above 2000Hz. taming that rising peak with a RC network will yield better results from the midrange since it won't interact as much. Predictable driver impedance that is flat and stable is easier to design for than drivers with swooping impedance curves.
Looking at the N36 vs. the N50, I see what you mean about the bandpass on the midrange for the N50.

However, I've only got the L96 and L112s...so I'm still ignorant on the changes needed to the LE10H-1 vs 128H woofer conjugate to the midrange's high pass filter. Can you help? :)

It seems it's time for me to pick up some books on passive loudspeaker crossover design, theory and application.:blink:

[Nightbrace]

The N50 is doing a bit more than that. Compared to the N36, the LFRolls off a lot lower than the N36. The midrange gets a real bandpass instead of a simple highpass and there is a RC trap on the output of the filter to further roll off the driver. The midrange xover also has attenuation fixed in addition to the lpad control. This is a really thought out bandpass and very specialized to the LE-odd midrange charactoristics.


I was asking WHY JBL decided to change the crossover so radically from the L36 to the L50. The only thing I can find wrong with the L36, besides its poor response under 40 Hz, is its resonance at around 240 Hz, not a reason to completely change the values for the mid and tweeter is it? Last I checked they have nothing to do with frequencies this low. They are 2 compeltely different animals and have virtually the same cabinet size porting and drivers.
Does the vertical alignment really call for this much change?

And anyways my question was if anyone ever tried using N36 networks in a L50, thats all. But thanks for the education.


To get back on Topic, finally, What are we talking about here? I forget, oh yeah L96/112 crossovers.

I think Giskard already summed up what to do with the L96 crossover, USE THE 3113B, the link is posted here. All is said is that you have to tweak it a bit for the L112 to suit your tastes. Or you could just leave it alone, as is stated here, there will not be MUCH of an improvement vs. the N112 crossovers.

[briang]

To get back on Topic, finally, What are we talking about here? I forget, oh yeah L96/112 crossovers.

I think Giskard already summed up what to do with the L96 crossover, USE THE 3113B, the link is posted here. All is said is that you have to tweak it a bit for the L112 to suit your tastes. Or you could just leave it alone, as is stated here, there will not be MUCH of an improvement vs. the N112 crossovers.:) Shame on me for hijacking. Bad, bad Brian! ;)

[Nightbrace]

Don't worry about it, happy someone else thinks the L112 needs some tweaking. I am selling mine soon.

[duaneage]

I was asking WHY JBL decided to change the crossover so radically from the L36 to the L50. The only thing I can find wrong with the L36, besides its poor response under 40 Hz, is its resonance at around 240 Hz, not a reason to completely change the values for the mid and tweeter is it? Last I checked they have nothing to do with frequencies this low. They are 2 compeltely different animals and have virtually the same cabinet size porting and drivers.
Does the vertical alignment really call for this much change?



To answer the last question first, yes.
Physical distance between drivers greatly affects how they sound at the crossover frequency. Increasing the complexity of the network makes each driver perform only in it'sintended range, rather than overlapping and contributing too much, or in some cases detracting from, the sound at crossover. When two drivers produce even a little bit of the same sound there is likely to be trouble.

240 hz just so happens to be close to the resonant frequency of most Le series 5 inch midranges. A simple, single capacitor network like the N36 has would let more energy go to the midrange below crossover, and since the driver impedance rises at resonance, it would hamper the capacitor from doing it's job. if the small enclosure for the mid was lacking fiberglass, it would resonate more.

So the reason JBL radically changed the crossover seems to be to clean up the midrange performance, increase the power handling of the midrange, and make the crossover behave better. Remember, the midrange is where the music is.

[Nightbrace]

Thanks for clarifying, I understand that positioning the drivers differently will change the sound enough to warrant reworking the crossovers, but my question is why would JBL deviate so much from the final result of the L36? As I said they sound completely different. Correct me if I'm wrong, but if the crossovers allow LESS energy to the mids and tweeters, wouldn't it stand to reason that there will be less information sent to the mids and tweeters?

Yes, the newer versions of the LE25-4 tweeters (033,066 etc.) are MUCH better, and I guess you could argue that they don't need as much energy, but as you said the midrange is where the music is and the midranges are virtually the same in all three of these speakers and there is a noticable lack of detail in the midrange of the L50 and L112 as opposed to the L36. So I am confused as to why this is better. If its to reduce the resonance its not a worthy tradeoff, at least I don't think.

I can't remember if the resonance I hear in the L36's was there in the L50's. It sounds like its an issue with the the construction of the enclosures and the lack of fiberglass than the crossovers, after all these were amoungst the cheapest speakers made by JBL in the 70's/

[Nightbrace]

BTW, Free Air Resonance is 160Hz for the LE5-6, and I was estimating, I don't have any fancy equipment unfortunately to test what I am hearing :(

I will try adding some insulation, and see if I notice a difference.

[4313B]

Yes, the newer versions of the LE25-4's in the L36's are MUCH better (033,066 etc.)What?

[Nightbrace]

I think you misunderstood, I mean the (033,066 etc.) are MUCH better than the LE25-4's..., but thanks for pointing that out, I changed it,. it should be a little clearer now :D .

[duaneage]

Like everything else , speakers work best in their intended range. Most 5 inch JBL mids I tested resonated between 220 and 275,but I have not tested them all.

One school of thought goes that a simple network is best because the phase of the crossover network is simple and unobtrusive. What can be easier than a single cap for the HF driver and leaving it at that?

But if a midrange is receiving information below and above it's best operating range then it may be damaged by bass and smear imaging for the highs. Midranges don't make good tweeters typically.

Cutting off the bass and highs abruptly is a noble goal but if done incorrectly or overdone it causes more problems. Complex networks are used to stop information right at the crossover frequency but they also change the phase of the signal as they do their job. Smart designing actually takes advantage of this to build a matching crossover circuit for the tweeter so when one driver is a little out of phase, another driver is coming into phase. These phase delays can also correct physical distance problems that cause response at an angle to be uneven. Look at the way 4411 or 4311 monitors have the
tweeters and mids mounted and you see an attempt to minimize distance.

I feel much of the success of JBL speakers has to do with the clever use of crossover networks over the years and the partnership between the driver engineers and the system designers who saw the big picture.

[Nightbrace]

Again thanks so much, I am learning, the crossovers are the heart of the speakers, this I know, but I know very little about them, just using my common sense.. And its good to know for me, that I am hearing resonance in the 230-260 Hz range as well, the spec sheet shows it at 160, but that may have been in a different enclosure.

My common sense is telling me that quick efficient shifts from one driver to the next will be best, but obviously this is not always the case.
I obviously need an education, as great as the majority of the JBL designs are, I know there is room for improvement. I'm not saying I'll be the guy to break the code so to speak, but it would be a skill I'd like to have.

Would upgrading the caps I have in the L36's keep the energy flowing while reducing the amount of resonant frequencies? I have spme experience with Coaxial drivers, but the cap values only change the crossover points and since its coming from the same voice coil, there is no worry about resonance. I am considering going back to my LN3's if not for anything, to learn about crossover design and proper driver loading, they would be great for this.

Would you recommend any resource material to get me started learning about crossover design and theory?

[duaneage]

If your really interested in how filters work you need to learn some of the math and concepts behind these three factors: impedance, reactance, and resistance. Speaker crossovers are AC devices that have active parts. capacitors and inductors react to AC voltage unlike resistors that do not change in value. I learned what I know about reactance years ago in electronic schools, reading ARRL manuals, and studying numerous books on electronic design, math, and project applications.

A good book for the money on all things speaker is the Loudspeaker Design Cookbook by Vance Dickinson. While not the end all of the science it is a damn good starting point. Vance explains more than you can imagine on how speakers work and how the drivers act in boxes. I have his 4th, 5th and 6th edition, the 7th is out I believe.

You can get a copy from Old Colony Sound Lab for about 39.95, probably money well spent. If you spend a few nights reading through it you'll know more about what we talk about and how things work.

There are other books I am sure but the LDC is a decent book well written and affordable.

[Nightbrace]

Thanks, I am an engineer and understand things pretty well, but am new to speaker design, but am learning quicky. I'll go to the library and see if they have that book, Care to explain reactance? Not a term I remember from my Physics 212 class.

[duaneage]

I doubt the library will have that one unless you go to the library of congress. This link explains it pretty well

http://www.electronics-tutorials.com/basics/reactance.htm

[Nightbrace]

UIUC has virtually every book every written, I am sure there is something I can find :).

[Pioneer727]

You can get it at PartsExpress.



http://www.partsexpress.com/pe/pshowdetl.cfm?&PartNumber=500-035&DID=7

Ron