New crossover for JBL L100 Century

JANTZEN AUDIO DENMARK

Designed by Troels Gravesen, March 2007 While I'm enjoying the sound from newly renovated JBL L26 Decades, I might as well start writing about the one and only, JBL L100 Century, one of the most common studio monitors of the Seventies plus. A month ago one of my friends threw in these speakers, having bought them from an old fellow for 300 DKK (50 US $) - a total give-away. All drivers in mint condition and the cabs with minor scratches. Tweeter foam had gone but can easily be replaced from eBay sellers. I won't go into historical details on the L100 as so much can be found at: http://audioheritage.org/html/projectmay/pmintro.htm, and in particular about the L100 here: http://audioheritage.org/html/profiles/jbl/l100.htm A rare opportunity to measure what the L100 did back then where I could only dream of owing a pair of these speakers from "over there". I don't recall the retail price, but they were very expensive. Having recently experienced the SEAS 503 kit, I now realise the 503 was actually a better speaker. Quite similar in size and power handling, the 503 had a better midrange and a better crossover and the 33FWK bass drivers may beat the 123A drivers. But the SEAS drivers didn't have the gorgeous looks of the JBL drivers, the whitecoated bass cone and the exquisite finish of the LE5-2 driver. I've always had a week spot for the LE5-2 due to how well it is built. Take a look at how the dust cap is glued to the cone and the fine touch of paint around the edges. It takes great skills to make such fine work. Only sad the LE52 is such a troublesome driver as we shall see later.
integrating well with midrange, actually better than many softdomes in my opinion. I've recently bought some LE tweeters on eBay: LE20, LE25 and LE26 and will make a special report on these some day.
The tweeter foam surround usually deteriorates over time.
Above my own stock drivers for recreating the JBL L100. I have to say it's much cheaper buying an old pair of L100s compared to buying the drivers separately on eBay. On eBay, sellers will completely strip an L100 speaker and sell it bit by bit down to the pins holding the front grille.
Above my friend's JBL L100, some scratches to the cabs, but drivers in mint condition. This first part of the my L100 story will not engage in a new crossover, merely try to find out what the L100 was in terms of driver performance and crossover construction. And we might as well take a look at the original crossover right away, because it's simplicity itself. 8 uF to the mid and 3 uF to the tweeter. Both drivers connected via an L-Pad for attenuation. Mid connected with inverted polarity. Having a 12" bass driver running full range is a bold choice and a lot of work must have gone into designing the 123A driver to produce a smooth roll off. The LE25 is a better tweeter than we might think, maybe not from measuring performance, but well equalised, it can deliver some excellent treble
Cabinet dimensions and driver placement. Cabinet modelling from TS-data reveals the below seen response profile. Basically the 123A driver has a too high Qt to be used in a vented cabinet and the result from a 44 liter
volume is a 3-4 dB increase in response at 70 Hz, which is confirmed by a nearfield reading of the response. Place the classic L100 Century on the floor, close to a wall, and we will have an overall 7 dB response increase from 50 to 100 Hz. Talk about a boom box! So, two options available. Add some acoustic foam to the vent, lift the speaker some 50 cm from the floor, which will raise the bass driver to ~70 cm from the floor, move it some 100 cm (distance to front panel) out from the rear wall, place in on the long side of the room with 1.5-2 meters to side walls, and we will reduce the bass peak to 3 dB above average level. Much nicer. Placing the L100 in corners is a no-no. The LE25 is much better than I had anticipated. It doesn't matter much whether the foam is in place or not. It may change the sound, but not measuring performance. File not shown.
Response from L100 as-is.

Individual response of drivers from original crossover with mid and tweeter attenuation set to zero. Some serious traffic jam around 3-10 kHz - three drivers all trying to make the most of it here. No wonder the original set-up fails utterly on vocal recordings. The 123A bass drivers really shouldn't try to play upper treble. The mid really shouldn't peak at 6.5 kHz and some proper low-end roll-off might improve performance. And the tweeter might improve performance from smoothing overall impedance.
Response in 44 liter closed box (Fb set to 1 = closed box, for the sake of ease)

Driver performance

Individual response of drivers without crossover. Smooth bass roll-off? Well, maybe not from this cabinet, but take a look below where I compare the 123A response from the L100 cab and a 60 cm wide, curved baffle. The LE5-2 is extremely sensitive. If it hadn't been for the dip around 1 kHz, we would have had 96-98 dB/2.8V sensitivity. What is not so nice is the 10 dB peak at 6.5 kHz.
The SPL response of both speakers with "0" attenuation for the tweeter and -3 dB attenuation for the mid to get the terrible peak at 6.5 kHz down to reasonable level. I found this attenuation setting to produce the best overall balanced sound. Nearfield bass response is merged at 350 Hz. Disregard difference in bass response. Probably not correct. Now, listening to some music from this setting may be close to what we heard back then. Quite an experience to hear that JBL L100 sound once again. Good? Hmm. Very impressive, lots of bass, speedy midrange. The sound of acoustic guitars is very good, but vocals? Something is very much wrong in the treble area, the 5-10 kHz range I guess although I'm strongly biased due to the 6.5 kHz peak. The danger of knowing what is going on. As soon as you know how it performs on paper, you can't help instantly having an opinion about what the sound is like - and what is wrong about it. But the 6.5 kHz peak is so severe that it's most likely the cause of some of the major problems of this
simplistic crossover set-up. The peaky nature of the 5-10 kHz range is very much audible and vocals really aren't good. Sibilance is highly emphasized and you can't attenuate the mid to get rid of the trouble without losing tonal balance.
Response of midrange alone with crossover and attenuation set for -3 dB. Severe dip at 1300 Hz and that 6.5 kHz peak.. This looks like a poorly managed tweeter response. A response up to 15 kHz! I mean, a single coil to the mid might at least have produced a more flat response, but then the upper treble may run into trouble from this approach.
Step response showing reverse mid polarity. The strange thing about these drivers is that the red terminal on the 123A bass is correct. The cone moves out when applied positive voltage. The LE5-2 cone in moving inwards when positive voltage is applied to the red terminal. Why would JBL do this? To ensure correct polarity of drivers when plus wires from crossover were connected to drivers? But the LE5 was used in other constructions. Doesn't make sense.

Modelling the original crossover performance
Mid attenuation fixed at -3 dB and tweeter response shown from +3 dB, 0 dB and -3 dB settings.
Modelled response of drivers where the drivers are attenuated with two resistors in order to recreate the measured performance. Not that easy, but this is close. Phase tracking between bass and mid is better that we might expect, but the tweeter is very much living its own life in the upper midrange. When I reverse the polarity of mid and tweeter in the LspCAD, things at least get worse. It's not that bad.
Mid response from attenuation at -3 dB, 0 dB and + 3 dB. Red = summed response from mid attenuation at -3 dB. If you run the L100 from "mid +3 dB" it'll rip off your ears!
A new crossover designed for all variants of LE5-2 drivers is a challenge and fortunately it can be done due to the very high sensitivity of the middriver. Placing a resistor before the driver will increase impedance and a standard crossover will have a deviation in midrange level of +/- 1 dB. This is manageable and for those who want to do some fine-tuning, the mid level may be increased or decreased by the input resistor.
New crossover design criteria
January-March 2007 A lot of work has gone into designing a new crossover and numerous simulations and actual crossover constructions have been tried in order to try taming the LE5-2 middriver and provide an overall balanced sound from the L100. What I wanted was to maintain the basic virtues of the L100, i.e. the 123A bass driver working almost full range, thus a 1st order filter was soon in place. To maintain 1st order filtering as high up in frequency as possible, the LE5-2 had an LCR circuit to flatten the rise in impedance at Fs, allowing the high-pass section to be one single capacitor. To cope with amplitude and phase between mid and tweeter, the mid needs a 2nd order low-pass section and the LE25 needs a 3rd order filter at 5 kHz to get amplitude and phase in place. See graphs below. Three significant things have happened here: First of all, the LE5-2 middriver no longer peaks at 6.5 kHz. Secondly the mid-section was constructed in such a way that the common variation in driver impedance (LE5-2) only produce minor deviation from target response. Last but not least, both the tweeter and mid produce a smooth roll-off below points of crossover. Should you want to increase mid and tweeter levels, you may bypass R2011 (mid) and increase R1061 from 3R3 to 5R6 or 8R2 (tweeter).

The 123A bass in L100 cab, blue, and from a wide curved baffle (red). Now, find a 12" driver today that will perform flat up to 6 kHz and have a smooth roll-off. Remarkable!
The (troublesome) LE5-2 middrivers:
I have two LE5-2 drivers asigned my future L100 reconstruction. So my friend's L100 gave me a rare opportunity to compare four LE5-2 drivers - and my goodness, is there some difference! In the upper midrange we have some 4 dB difference in sensitivity. Basically this is due to differences in voice coil impedance. See below.
Impedance of four LE5-2 drivers. Some difference! From 5 to 7 ohms. Unfortunately both my friend and I have one of each and the cones are so different in colour that we can't swap drivers. If you ever buy LE5-2 drivers on eBay, be sure to have a DCR reading of both drivers before bidding.
New crossover topology. Pay notice to terminal lay-out of JBL drivers.
Installing the new crossover kit
Drawing of crossover board. Full size drawing is included. Remove old terminals including the thin masonite panel. Scrape off residual glue and sand the rear cabinet panel to make a smooth surface.
Prototype of new crossover kit. The kit will include: 1. 180 x 280 x 12 mm MDF board. 2. Terminals. 3. Two resistors for tweeter attenuation will supplied, both can be fixed to the crossover board and a short wire to ground provides the preferred attenuation. 4. Soldering tag strips. 5. All crossover components: Baked non-resonant coils, all polypropylene capacitors, MOX resistors and cables for connecting drivers. Not supplied is silicone glue, screws and optional damping material. As can be seen below, it's highly recommended to replace old damping material. I've used soft 40 mm polyurethane foam and a sheet of MDM3. It changed the overall sound a lot - and for the better.
Kit components for one speaker.

New damping material.

Summed frequency response, red, and response of individual drivers from new crossover. All drivers now display smooth roll-off characteristics.
Six page instruction manual included in kit. Contains basic crossover design with all component values and wiring instruction.
Performance of L100 with new crossover.
The litmus test: Reverse mid polarity. Reading taken a 1 m distance at middriver height. Due to placement of drivers, this reading may somewhat depend on placement of microphone. In my own recreation of the L100 I'll have a standard vertical placement of drivers with mid and tweeter slightly off-set. L100 frequency response from new crossover. The dip at 1.3 kHz is caused by the small front panel. Some of the LE25 tweeter display a peak at 12 kHz. Below is shown one of my LE26 tweeters, where JBL apparently succeeded in taming this peak. However, I must say the treble from the renovated L100 is remarkably smooth. If your L25 tweeters are intact, do not discard. They're excellent tweeters - only need a proper crossover. Final remarks to the L100: I dare say it's quite a different speaker compared to the previous "two-caps" set-up. Classical music is a delight, vocals are handled very well and should you throw a Saturday evening party, give it a 150 wpc amp and let it rock! It runs well from my 20 wpc PSE valve amps, though not at very high levels. One last thing: The L100s need stands, at least 45 cm high to bring the 123A drivers off the floor. Remember, the 123A is handling most of the midrange and a 5 deg. tilt is doing even better. Enjoy!