PDA

View Full Version : jBL c60 sovereign crossover upgrades/replacement



aaronz28
05-09-2014, 06:55 AM
Hey gang-
i've been going back and forth with these old boys as to whether to outboard the 375 and 075 with custom wood horns, or just elevate them to ear level etc, whether to run them with a single tube amp, or actively crossover/bi-amp etc.

i think i've finally come to the realization that leaving them stock appearing is going to be the most realistic for me - even if I can't improve on the potato masher/lens setup - I still might replace the 075 with 077 down the road - but for now, i'll leave them stock..

what I am wondering is if anything can be done to improve the stock crossover network... i'd be ok with building/buying a new unit for each (leaving the stock set in place, but just running new wire in to a modern crossover network -
something that i could single amp now, and eventually bi-amp down the road.

i'm on the home stretch of rebuilding an estate sale find McIntosh C22 and MC240 that will live with these guys - and might eventually put something a little "stouter" on the big 15" - but I really do not want to introduce any additional active circuitry between the preamp and power amp(s)

any thoughts here? I know there was a guy who redesigned the crossover for the L-100/43xx series, but i don't know if anyone every improved on the stock network for the S8R setup...

Thanks

Mr. Widget
05-10-2014, 09:09 AM
I would replace the potato masher with the short foghorn and slant plate lens... it shouldn't be too difficult to adapt the cut out. The 2405 or 077 will drop in the same cutout as the 075 so that's easy and very worthwhile.

On the network, a new ground up passive network will certainly be a step up... those early autoformer based networks aren't terribly good in my opinion.


Widget

aaronz28
05-10-2014, 09:13 AM
I would replace the potato masher with the short foghorn and slant plate lens... it shouldn't be too difficult to adapt the cut out. The 2405 or 077 will drop in the same cutout as the 075 so that's easy and very worthwhile.

On the network, a new ground up passive network will certainly be a step up... those early autoformer based networks aren't terribly good in my opinion.


Widget

These speakers have a slant lense attached/ or are u referring to another one?

who would be a good contact for the ground up crossover redesign?
i can build it, but id need someone to design and recommend the best caps/coils
thx

Mr. Widget
05-10-2014, 10:08 AM
You mentioned potato masher in your first post... if you have the slant plate then adding an 077/2405 and networks will move you forward.


As for who can help you with the networks... there are a number of people here who can help... the problem is most of us just don't have the time.


Widget

aaronz28
05-10-2014, 11:51 AM
You mentioned potato masher in your first post... if you have the slant plate then adding an 077/2405 and networks will move you forward.


As for who can help you with the networks... there are a number of people here who can help... the problem is most of us just don't have the time.


Widget

yea - sorry - some reason I assumed that these had potato mashers behind the lens - here is the setup
http://i593.photobucket.com/albums/tt20/aaronzl1/DSC_8861_zpsfd43dcfc.jpg (http://s593.photobucket.com/user/aaronzl1/media/DSC_8861_zpsfd43dcfc.jpg.html)
http://i593.photobucket.com/albums/tt20/aaronzl1/DSC_8869_zpsba7dc4f4.jpg (http://s593.photobucket.com/user/aaronzl1/media/DSC_8869_zpsba7dc4f4.jpg.html)

BUT - I'm wondering if I could modify and go to the L89 lens/setup - i'm just not sure that would 1) be possible, or 2) improve them that much....
is this lens a direct fit, or do I have to change the "horn" as well -
http://i593.photobucket.com/albums/tt20/aaronzl1/_57_zpsc5f968dc.jpg (http://s593.photobucket.com/user/aaronzl1/media/_57_zpsc5f968dc.jpg.html)

going to the 077 = gotcha -

as for the crossover - i'm kinda surprised that no one has improved on the LX5 network - wasn't these the standard crossover for the S8R?
at some point - i bypassed one of the switches and replaced some of the caps - which was an improvement (suggests by a user here - whose name i can't remember :D) - but i still think there is room for improvement in a passive setup -

if anyone else can provide assistance with a design, or an existing network that i could build - i'd greatly appreciate it.

cheers

Aaron

mbeards
05-16-2014, 09:18 AM
Interestingly enough I am currently working on modeling the LX5 and N7000 for my home built Hartsfield.

With some simple simulations I have found that we can do much better on the crossover designs. I have been using Texas Instruments free TINA spice simulation software to model and analyze the transfer function of the crossovers. Note that this is does not include driver response.


Here is the stock crossover:

62158


Using Ideal loads the response doesn’t look too bad:
(This shows the 3 possible attenuations through the autoformer with the top curves displaying the sum of the three possible positions.)


Here is 16 ohm woofer/ 16 ohm midrange horn / and 8 ohm tweeter:

62159



Next is 8ohm/16ohm/8ohm

62160


However things get interesting when you add the actual load models for the drivers I plan to use. I am using the E145, 2441, and 075 setup. Using my DATS (Dayton Audio Test System form Parts-Express) I have measured the impedance curves of the drivers I plan to use and came up with models.

The 2441 driver with Hartsfield horn and lens:

62161


The 075 driver:

62162

As you can see the 2441 looks like an average load of 16 ohms at the first crossover of 500Hz. However, at 7KHz the driver is very inductive. To accurately model the crossovers, I used an ideal 16 ohm load for the first crossover point and for the second added a resistive and inductive load. You must switch these loads models depending on which crossover point you want to accurately view. The final full transfer function will be a estimated sum of the LF and HF models.

So LF model : 16ohms

HF model: 7.47ohms + .178mH L

mbeards
05-16-2014, 09:21 AM
Here I varied the HF model impedance of the series resistor and inductor combination to see how well this match the DATS system:
(as you can see it is a very good HF model)


62163



For the 075 the load is roughly an 8 ohm average load through an L-pad and is modeled as such.

The E145 was measured with and Re = 5.5ohms and Le = .820mH


The first step was to model the LF point with the LF models:

62164


Next the HF 2441 Model was added to model the 7KHz crossover point:


62165


This looks very nasty the inductance of the 2441 at high frequencies kills the crossover point and also adds a nasty resonate peak! A zobel on the 2441 is absolutely necessary when going to a 3-way system with a tweeter. JBL never thought of this in the 1960’s and 70’s


Here is the impedance plot of the 2441 with a zobel of Rz = 9.5ohm and Cz = 2uf:


62166


The Impedance at HF is now relatively constant.


I also modified the original LX5 zobel to better suite the E145


Here is my final crossover design:


62167

mbeards
05-16-2014, 09:22 AM
Final LF Response:

62168


Final HF Response:


62169



My goal was to flatten the HF response as much as possible. The dip and then subsequent bump at around 500Hz for the LF model was hard to compensate for. I wanted to keep the Autoformer in my designs for selectable attenuation and therefore, the inductance was fixed so my option were limited.

However, looking at the published driver efficiency curves, this dump and then bump may be beneficial. Since the E145 will be gaining efficiency and thus output right up to the crossover point, the dip can help flatten the overall bass response by making lower frequencies relatively more powerful. As for the bump, the 2441 hasn’t reached full efficiency until 2KHz or so. This bump I feel will actually help make the transition to the 2441 smother and flatter as until it reaches full efficiency.


I hope someone finds this useful.


Matt

Mr. Widget
05-16-2014, 12:15 PM
With some simple simulations I have found that we can do much better on the crossover designs. You needed a computer to tell you that?:D

My ears told me that decades ago.;)



Widget

Mr. Widget
05-16-2014, 12:19 PM
I hope someone finds this useful. It is certainly interesting... thanks for sharing.


Widget

aaronz28
05-17-2014, 02:30 PM
that all looks very interesting -
however, I have no idea how to extrapolate the info shown, and build a crossover suitable for my speakers LOL

other than simply recapping the original parts - which i'd rather not do - -

:D

A

mbeards
05-17-2014, 07:34 PM
So basically what I have found is that the LX5 is an "Ok" crossover.

It has some peaks and dips in the frequency response that with some trial and error can be improved. However, when using two crossovers for a three way system, say the LX5 cascaded into an N7000 as used so popularly by JBL in systmes such as the paragon, the stock design falls flat on its face if the midrange driver as any appreciable voice coil inductance.

This inductance causes a rising driver impedance as frequency increases, thus killing the midrange attenuation slope at the crossover point between the midrange and tweeter. Basically, the drivers rising impedance cancels out the attenuation of the crossover network.

The VC inductance also adds a nasty resonant peak making the response even worse.

Generally, the midrange driver will continue to play at higher frequencies than expected at higher output that the tweeter. It will therefore mask the tweeter and play over it. The tweeter is useless in this case. Worst of all, if the midrange driver is not designed to play at frequencies this high, Its distortion will be significantly higher.

Therefore, a zobel is needed across the midrange attenuation network that will cancel out the rising inductance of the midrange driver.
To pick the right zobel values, use one of the many calculators online. However you need the driver VC resistance and inductance

This is by far the easiest and best place to start modifying.

Happy Listening,

Matt

tube-it
02-10-2015, 05:37 PM
What is M1, N1 and N2?

I assume L3 is .3mH (millihenry) ? I see a "u" that is often used in place of µ (microhenry). Am I Correct?

Were you able to try this out? How did it sound?

ivica
02-11-2015, 04:29 AM
Final LF Response:


However, looking at the published driver efficiency curves, this dump and then bump may be beneficial. Since the E145 will be gaining efficiency and thus output right up to the crossover point, the dip can help flatten the overall bass response by making lower frequencies relatively more powerful. As for the bump, the 2441 hasn’t reached full efficiency until 2KHz or so. This bump I feel will actually help make the transition to the 2441 smother and flatter as until it reaches full efficiency.


I hope someone finds this useful.


Matt

Hi mbeards,
Thank You for the info.
As I have understood, You have removed L-pads ( 16 Ohm for 2441, and 8 ohm for 2402) what can be seen on some JBL networks, and have removed 16~20 Ohms resistor paralleled with the 2441J driver, and almost the same for 2402...

regards
Ivica