3107 crossover question

[gtxdude]

Hi guys! I have been lurking around here for a while and have enjoyed and been amazed at what a wonderful resource this forum is. I worked as a recording engineer back in the late 70’s and early 80’s and began my life long love of JBL products. One of my most lingering memories was a room I used to work in that had a pair of 4350 monitors. You could track for 20 hours straight, at any volume, with those monitors without a hint of hearing fatigue. They were the sweetest, warm sounding monitors I have (to this day) ever heard. I vowed that one-day I would own a pair of those monitors. Then life happened: marriage, kids, career…

(Fast-forward 25 years) I now have the time, money and resources to make that dream come true. I have undertaken the construction of a replica pair of the 4350 monitors. Thanks to you guys and the storehouse of JBL legacy information available here I have found everything needed to complete this task. My eternal thanks to all of you! Please, keep up the good work.

I have secured all of the components necessary to build the monitors on ebay, with one exception: the 3107 crossover networks. They have been very hard to find (for me). After some thought I decided to build a crossover network of my own using modern components which should improve the performance of the monitors too boot.

I have attached a PDF detailing what I have built and the components used. I would invite any comments or constructive criticism on my 3107 knockoff design.

And now, down to my tech question: The mid-range output to the 2440 driver as it is drawn on the original JBL 3107 schematic is electrically out-of-phase with the other outputs. Is this correct? Is the standard polarity marked on the 2440 backwards? (not likely) Is this done for some secret reason that is not obvious? Or, is this just a bone-headed mistake that the drafting technician made when documenting the 3107 network and nobody ever caught it? (not likely either)

The JBL reference document I am using may be found here:

http://www.jblproservice.com/pdf/Net...%20Network.pdf

Can any of you JBL experts shed any light on this?

Thanks again, and I look forward to becoming a useful member of this community.

Dennis Dodd

[speakerdave]

The 3107 is considered obsolete even here, except for those who 1) must have a 4350 instead of a 4355, 2) can find some 3107's--they're around, since JBL sold out a bunch of NOS a few years ago--and, 3) can get real 2440's with good 2440 diaphragms, which are no longer available from JBL. The replacement, the 2441 diaphragm has greater high frequency extension, so the 3107, which lacks low-pass filtering for the treble horn, cannot be used with it.

Member 4313B has created an updated version of the crossover for the 4355. It cannot be replicated in original form because it used a special tapped inductor which is no longer available. I don't think he has done the same thing with the 3107 for the reasons mentioned above. The schematic for the updated 3155 has been posted on these forums. These days I think he would recommend you build a charge-coupled version of it (or let him build one for you).

The 4350 is not buildable unless you can find some 3107's and some 2440 diaphragms. Even trying to replace the 3107 network with your own "equivalent" the speaker will not sound the same, so you might as well use the 4355 network and the 2441 diaphragms. You could do everything else the same as the original, if you want. The adulteration of crossover and treble diaphragm will be hidden and could be your little secret. All reports are that the 4355 combination sounds much better anyway.

Member saeman builds 4350's, among other classic JBL monitors. Maybe he can offer some suggestions


David

[speakerdave]

Just noticed that was your number one post. Welcome to the forum. That is a brave project you've taken on there. I considered it once myself, but my goals changed. I had a pair of those 3107's at one point but sorry to say they are gone. We like pics of work-in-progress (but not every nail), so keep us posted.

David

[4313B]

Member 4313B has created an updated version of the crossover for the 4355.

Yeah, three pairs of biased 4355 networks get shipped this week as well as two pairs of biased 4350 networks. It's a ton of work making that many networks by hand. I also have six pairs of biased 4345's on the bench and six more pairs of biased 4355's to build. Then I evidently have three pairs of biased 4343 networks to do. Everyone has been extremely patient and supportive.

[boputnam]

Then I evidently have three pairs of biased 4343 networks to do. Everyone has been extremely patient and supportive.

That there, is terrific news!

[gtxdude]

thank you david. i am new here, and unsure of protocol. this is my second post on this forum, i DO NOT mean to sound like a jerk here...



but, can someone take a swing at my original technical question:

why is the 2440 output in the 3107 network out of phase with the rest of the outputs?

dennis dodd

[Mr. Widget]

why is the 2440 output in the 3107 network out of phase with the rest of the outputs?

Due to the interplay between the network, the drivers' respective physical locations and the horn length, putting the 2440 out of phase electrically actually is getting it closer to being in phase acoustically.

Phase, like impedance isn't just a simple number... you are rarely really truly "in phase" as with few exceptions loudspeakers' impedances are rarely purely resistive and defined by a single number either.


Widget

[speakerdave]

. . . . can someone take a swing at my original technical question:

why is the 2440 output in the 3107 network out of phase with the rest of the outputs? . . . .

It sounds better that way, sometimes. The electrical phase distortion caused by the high pass and low pass sections of the midcone/treble horn crossover point, combined with the phase distortion caused by the axial offset of the acoustic centers of the drivers as they are mounted in the enclosure MAY be resolved or nearly resolved by connecting the drivers with reversed polarities after the crossover. The connection of the super tweeter may be conditioned by the same considerations. The choice of connection is usually made empirically, however. Finally, some JBL drivers give forward motion with positive connection to the red terminal and some with positive connection to the black.

The woofer should probably also be connected with reversed polarity relative to the midcone. If you use vintage JBL/UREI crossovers and/or amps to biamp you should be aware that the pin polarity in the XLR connections is opposite industry standards. You'll have to sort that all out to get it right.

David

Edit: I see Widget posted while I was responding. I'll leave this up for awhile, at least.

[John]

When did JBL drop the 2440 fram??? I must of been under a Rock or too much time rolling on the interstate system

[Allanvh5150]

Hi GTXDUDE. I think the answer that you are looking for would be this. The bass driver runs through an inductor and as such it runs 90 degrees, lagging, out of phase to the main signal. The mid, 2440, runs through a capacitor which also puts it 90 degrees out of phase, leading, which means the bass driver and the horn are 180 degrees out of phase at the crossover point. To correct this JBL wired the driver backwards. The tweeter is electrically out of phase as well but because it is crossed over much higher and is mounted forward of the 2440, the phase relationship is kept equal. I hope you can understand this. Welcome to the forum

[JBL 4645]

We’re back to this (phasing issue) again nearly all the new digital crossovers today offer (polarity and variable phasing shifts from 0 to 180) of each LF and HF driver with remarkable easy. I think digital crossovers are the word now over the old analogue models, nothing personal.

[subwoof]

I agree - scrap the whole passive ( and old aluminum dia's ) idea and use newer drivers and a full DSP for a crossover. Amplifiers are cheap and you would have all the control to make them fit into your environment.

sub

[gtxdude]

allanvh5150 - that was the answer i was looking for. makes perfectly good since. thanks - dennis dodd

[bldozier]

Hi GTXDUDE. I think the answer that you are looking for would be this. The bass driver runs through an inductor and as such it runs 90 degrees, lagging, out of phase to the main signal. The mid, 2440, runs through a capacitor which also puts it 90 degrees out of phase, leading, which means the bass driver and the horn are 180 degrees out of phase at the crossover point. To correct this JBL wired the driver backwards. The tweeter is electrically out of phase as well but because it is crossed over much higher and is mounted forward of the 2440, the phase relationship is kept equal. I hope you can understand this. Welcome to the forum

im a newbie, in all things inductance, the 90degree inductor is a referral of location & graphical charge, of such specific inductor in the 3107s case the three here in this photo set at 90 degree compared to an upright situation, correct?

[Doctor_Electron]

To properly implement the 3107's response characteristics, any diy or manufactured inductors' DCR should be the same as that of the originals.

I don't know the values for the section controlling the 2405, but the DCR of the 2202 Midbass high-pass and 2440 low-pass 1.8 millihenry coils is 3.5 ohms.

A lower DCR value, 1.8 mh coil can be corrected to the 3.5 ohm value by adding a non-inductive resistor in series to achieve the original DCR. Resistors of the type JBL used in the 3107 (non-inductive, vitreous enamel power resistors) would be a very good candidates to use for this purpose.

Good luck with, and have some fun with this project... Regards, -D_E-