View Full Version : Crossover For 123A-1 and LE5-2

10-06-2004, 09:26 PM
I just got a pair of 123A-1 woofers and LE5-2 mids to begin experimenting with for a HT speaker system. I have read through some of the threads regarding the L100 speakers, and know there are some low opinions of them. However, I am trying these drivers because they are inherently shielded, shouldn't require gobs of power (I only have 45W), and should reach 20 Hz in an appropriate box.

I have seen the crossover schematics posted here for the various L100 variants. What I would like to know is if anyone has come up with their own improved crossover and box for these two drivers. Any recommendations for crossover frequencies between the two, as well as how high to roll off the LE5-2?

Also, any recommendations for a shielded tweeter to mate with these?

Thanks in advance

Don C
10-06-2004, 10:16 PM
Have you seen the L-88 schematic?

10-07-2004, 06:52 AM
Originally posted by Giskard
I think it's all in fun on this website. Giskard!! :bash:

The L100's, as you might have noticed from looking at the various schematics, crossed to the LE20 at 7,500Hz. Made for a nice three-way. I don't recall anyone here "re"designing the L100 network, per se - but maybe Earl K or others have some ideas. I'm sure Giskard could "tune-up" any network you might wanna tweak.

One last thing - mind that the 123A is POSITIVE.

Why only 45w? (OK, that is the last thing...)

10-07-2004, 07:14 AM
Hello Ddonnelly

Well your going to need quite a box to get to 20Hz with the 123's. I had L100 for years and they could take a moderate amount of boost because of their x-max but they sounded better without it as it muddied things up a bit. Depends on how you cross them over. Those 2 drivers almost have enough high frequency extension without a tweeter. You could always cross to the LE5-2 at say 800Hz or so and take that up to about 5K or so and bring in an 035ti above that. If you look at a 4312A that is very close except it use the 104H which does not have the rising frequency response like the Le5-2. The 035TI is not shielded but how close will you be I used them within about 3 ft and they were OK with my TV.


10-07-2004, 09:57 AM
Thanks for the replies.

1. I see that all of these speakers let the 123 run full range. I wonder if the driver rolls off smoothly enough for this, or if a simple first order low pass might help things out?

2. 7500Hz seems a little high to run a 4"/4.5" midrange driver. 5000Hz might sound a little better? Maybe even lower? I think I will try 800Hz-5000Hz first.

3. Does anyone have a crossover circuit that would tame the rising frequency response of the LE5-2? Would a zobel help? A shelving circuit?

4. Box size. Modelling gives a box size of around 15 ft3 for 20 Hz. Obviously this is enormous. However, it looks like a 5-6 ft3 box gives you about 25 Hz if vented, or around 35 Hz if sealed. These seem pretty good to me.

5. If I like how the LE5-2 sounds, I plan on using one right on top of the TV, but with different woofers.

6. 45 watts is what my HT receiver is rated.

10-10-2004, 10:19 PM

2.0 mH low impedance coil series, 52 ufd cap parallel, parallel circuit (5.5ohm hi watt resistor, 31ufd cap) wired in phase


13.3ufd cap series, 1.0mH coil series, 3.0mH low impedance coil parallel, 4ufd cap parallel, 8ohm lpad, 2ohm resistor series wired in phase


Typical circuit would be 3 – 4 ufd cap series, .18 to .40 mH coil parallel, lpad, wired out of phase.

Crossover points about 600 and 4000. This design was a demo for a software design program I have in my files somewhere. The tweeter used was the Focal 120K replacing the LE20 (3ufd cap with .35 mH coil). This tweeter in the T90 series might be a good match because dispersion pattern of the concave tweeter is a good fit for a wider baffle needed for the 12” woofer.

David F

10-11-2004, 09:40 PM
David F

Thanks for the crossover design. Did you actually build this crossover, or was it just a computer simulation? If you built it, what were the improvements over the standard L100 crossover?


10-12-2004, 08:58 PM
The author (Bob White, contributor to Speaker Builder Magazine and collaborator with Robert Bullock on speaker modeling material) reviewed the software with the L100 as the patient. As such, he developed a theoretical- “cookbook”- xover design and then modified it using optimization part of the software. Actual response curves were compared to predicted response. The actual response followed predicted quite closely for each separate driver. Curiously, there was no summed-response to indicate overall system response nor was there any mention of off-axis response. A summation of results mentioned better localization in image, more depth, better definition in vocals and tighter drums.

I should also note his polarity was network positive to JBL negative for the woofer and mid (positive pulse-outward cone motion) and the opposite for the tweeter. He also mentions some break-in needed for the mid with the xover point lowered from the stock design.

I have put together this design but not for use on the L100 components so I cannot fairly comment from personal experience.

David F

(those damn Yankees)

10-13-2004, 07:46 AM
I'd be hesitant to lower the crossover point on an LE5-2. That always seemed to be the weak link in an L100- the LE5-2 is almost always the first driver that blows up, IME.

Now, this MAY be that in the L100, with just a cap on the midrange, that the LE5-2 is receiving lots of HF harmonics, which add to its normal load of midrange duty... but the fact that the LE5-2 almost always blows before the LE25... well, that's kinda telling, given that the LE25 is also getting the same harmonics...


10-13-2004, 10:38 AM
Originally posted by GordonW
I'd be hesitant to lower the crossover point on an LE5-2. That always seemed to be the weak link in an L100- the LE5-2 is almost always the first driver that blows up, IME...


Gordon raises a good point about a power-handling weak point on this model, especially considering that rebuild kits are at best scarce for the LE5-2. Yes, there is more energy going to the mid range with the lower xover point but this is higher in the range where the LE5 should be able to handle the load. The stock single-pole xover I assume to be –3 at 1500Hz would theoretically be down –9 at 750, -15 at 375, and –21 around the driver’s resonance point at 187Hz. The double-pole xover on the revised design is –6 at 600Hz, -18 at 300, and –30 around the resonance frequency. The rising resistance approaching resonance will alter these predictions but the point should be that there is potential for better overall power handling at lower frequencies with the revised xover. Also true, the higher frequencies are rolled off under the revised xover design. Given my personal setup having a clean power source with good headroom and use in a moderate sized room at home, I would not anticipate any concerns about using this revised xover. Other scenarios may warrant heed of Gordon’s warning.


10-13-2004, 10:52 AM
One should use the 3.5 mH (7.5 ohm DCR) inductor in parallel with all LE5 transducers with 6 dB/octave crossovers. The 3 dB resistor pad found in so many JBL systems running the LE5 should also be used. The largest single pole capacitor value for the LE5 should probably be limited to 16.5 uF.

The LX30 used a 2-pole filter consisting of a 28 uF series capacitor followed by a 1.7 mH parallel inductor but didn't have the 3 dB resistor pad. A few LE5-2's did get blown in that system. Addition of the 3 dB pad fixed the problem.

Harvey Gerst
10-16-2004, 07:34 AM
It was my understanding that Ed May designed the crossover for the 4310 and George Augspurger designed the crossover for the L100. In Ed's home system (basically 4310's with 2 123-A's per side), they were in 8 cubic foot, ported cabinets and flat down to about 27 Hz. Ed's crossover (at home) consisted of a simple first order network, with the 123 running full out, and a capacitor to roll off the bottom on the LE5 and LE-20. L pads on the mid and tweeter. I don't remember the exact values (maybe 8 mfd, and 30 mfd?).

Ed's system is still running perfectly today, re-built in our Studio A, with a duplicate system going into our new mixing room.

10-16-2004, 07:57 AM
Thanks Harvey :)

I would seriously love to hear your Studio A! :yes:

All I've been able to dig up on the "original" 4310 is that it was called the C533 and it was designed for Bob Fine. The network was an LX30 (not to be confused with the LX30 from the LCS) and consisted of a 3.0 uF capacitor & L-Pad on the LE20-1 and a 13.5 uF capacitor & L-Pad on the LE5-2. The LF was the 123A. Can you please set us straight on this?

10-16-2004, 08:02 AM
Here are some additional schematics we have available to us.

Harvey Gerst
10-16-2004, 08:20 AM
The N100 looks right for Ed's system, but I'm not sure about the values. I think George designed the NX12 for the Century.

10-16-2004, 08:23 AM
Harvey, did you ever get a chance to tinker around with any of the S-70 Alpha or Alpha III systems? The idea of a 123A in 6 cubic feet with a passive radiator has sounded quite interesting to me.

In fact, I've assembled the drivers (123A, LE5-2, LE20) to BUILD a reproduction S-70... as soon as I get around to building some cabinets. The only change from original, is that I'm using PR300s rather than PR12s... I've found that "looser" suspended passive radiators (foam surround) seem to sound better than stiffly suspended ones (fabric surround)... the bass is usually smoother and less 'boomy' even with the same f3 point and same effective mass on the radiator diaprhagm.

I'm probably going to try the C533, the original L100, and the later L100 crossover designs in it, to see how they compare, both listening and testing with LMS/LEAP. I will also be designing a crossover, from scratch, in LMS/LEAP, to see what these drivers can do with "modern" thought in crossover design factored in.


10-16-2004, 08:24 AM
Thanks Harvey!

Very cool Gordon!

10-16-2004, 08:32 AM
Here's a few more networks

07-08-2005, 06:59 AM
Hi, all,

my 5 cents. I am trying now to cross 123a1 with another silk dome tweeter, and measurements (Bruel & Kier high accuracy meter) show 123 is peaky at 5 kHz. It rolls of smoothly up to 2 kHz, and then makes a peak at 5 kHz.

This makes me a doubt how JBL used it in full range.

I can place the curve.

Anybody met the same?


07-08-2005, 11:53 AM
If you would, please do post the curve. My take, is that by 5KHz, the response of the 123A is probably down enough, that the peak doesn't go above the "0 dB" line... ie, even at the top of the peak, it's BELOW the passband level of the driver. In those days, this was probably thought to be good enough... it wouldn't really upset the overall FR curve, though it might wreak a bit of havoc with polar response...

My take, would be to "assist" the rolloff at 2KHz, with first order crossover and a Zobel, and maybe additionally a notch filter at 5Khz. Given that, you should be able to use a GOOD, EFFICIENT silk dome tweeter successfully. Might I suggest the SEAS 27TFFC? It's one of the only silk domes that simultaneously has a) enough low-end excursion- ie, it can handle low frequencies (between 1500 and 2500 Hz) without distorting or sounding strained, and b) enough EFFICIENCY, to actually KEEP UP with a 123A, which is effectively about 94 dB/2.83V. I'd be inclined to put a 2KHz second-order highpass filter with a Q of about .577 on the 27TFFC, and just tweak both the highpass and the inductor value on the woofer, to get the best possible blend...


07-12-2005, 02:23 AM
That's the curve, JBL123-A1 is in free air, so there is normal roll off at low end.

Done with my 2a3 SE amp (it has flat responce, and about 2.5 Ohm output), the measurement is in 1/3 octave pink noise, with Bruel meter.


08-23-2017, 05:58 AM
I have an old black pair of 123a woofers from an early L-88 and have looked at using the S70 Alpha III crossover design but would like to alter the midrange high pass from 7000hz down to 3500hz. Any recommendations would be appreciated. My goal is to avoid the up sloping peak in the LE5-2 midrange.

09-17-2017, 09:03 AM
I have an old black pair of 123a woofers from an early L-88 and have looked at using the S70 Alpha III crossover design but would like to alter the midrange high pass from 7000hz down to 3500hz. Any recommendations would be appreciated. My goal is to avoid the up sloping peak in the LE5-2 midrange.
No thoughts?