JBL L26 3-way
Copyright 2006 Troels Gravesen


The JBL L26 Decade has been in the attic for some time and a mail from the US about a possible second round on the L26 got me started on the final chapter on this vintage JBL speaker.
From the beginning it was clear that this speaker might end up as a 3-way. A large 10" bass driver with problems in the 2 kHz region doesn't exactly call for a 2-way despite the 2-way crossover working reasonably well. To make a 2-way from a 10" midbass takes a driver with a very smooth roll-off profile, and this JBL hasn't.

As discussed before, restoring old speakers always leaves you with a lot of dilemmas. What should be left "as-is" and what should be modified to possibly enhance performance. You easily get into modifying everything, leaving little of the original design and end up with a modern speaker made from highly tweaked old components with little resemblance of the real thing. The L26 may not be one of the most precious gems from the JBL company and I have tried taking the middle road and added bracing to the cabinet, done only little to the drivers and most importantly, kept the LE25 for treble. This is really not a bad tweeter, but it needs some equalisation to deliver the best. I could very easily have taken the SEAS MCA12 for mid and the 27TFFC tweeter for treble - but then it definitely wouldn't be a JBL speaker any more.

Having a pair of JBL LE5-12 middrivers, these might be a possible option for a suitable midrange relieving the bass of handling the troublesome upper mid/lower treble range, but these LE5 drivers can be major trouble and I almost gave up in the process. The LE5 drivers are found in a number of variants and I have a pair of LE5-12 and a pair of LE5-2 (the latter the alnico version from the L100 and other JBL speakers). The LE5-2 is worse than the LE5-12 but I'll get back to this later.

The one and only alnico LE5-2 to the left and the "modern" ceramic magnet version LE5-12 to the right.

Seen above (left) is the response of the LE5-2 (blue) and LE5-12 (red) on a 20 cm wide baffle. Overall sensitivity appear to be almost the same. The ceramic magnet on the LE5-12 is doing very well and from this presentation the LE5-12 seems like a much better driver compared to the LE5-2, having a smooth response in the upper mid and lower treble.
When I got my LE5-12 drivers, one of them had a severely bumped alu dome and there's no way you can restore the very thin alu that forms the dust cap. So off it went and I added an inverted paper dome like the one on the LE5-2 - and I shouldn't have done that. The frequency response turned out just as bad as the LE5-2. Hmm.... I really thought this alu dust cap was crap and then it turns out to be much better than the paper "bowl" of the LE5-2. So what to do? I didn't have new alu domes and no suitable paper domes as well. But I had a pair of rubber domes from an old pair of Vifa P13WH drivers. A rubber dome on a JBL driver??

Right: The response of the LE5-12 with alu dome (red) and with rubber dome (blue). This really doesn't look bad. The driver now has a steadily rising response up to 6 kHz, a minor peak at 8 kHz and it's gotten rid of all the rubbish above 10 kHz. The sensitivity seems a little reduced, but there's plenty for an L26 3-way.
So in went the rubber dome as seen on the photos below.

LE5-12 with alu dome removed to the left (felt material added the polepiece for damping) - and LE5-12 with rubber dome to the right.
Should you want to follow this route I have no idea where you can get two rubber domes, but I'll try making a crossover version suitable for the alu dome just in case you have these in good condition. Even if they have minor bumps I recommend leaving them in place. I didn't measure any difference from my bumped and non-bumped drivers.
So this all looks very nice, but the trouble doesn't stop here. The LE5-12-RD (rubber dome) may perform well on a 20 cm baffle, but the L26 cab is much wider.

Left: Above the response of the two LE5-12-RD on the L26 baffle after the rubber dome glue has settled. Not bad at all.
As soon as the LE5 mids sees a wide baffle, the response below 800 Hz rises some 4 dB, the LE5-2 even worse than seen here. One of the reasons for this is the very high Fs of the drivers as seen below.

Right: Above the impedance of my two LE5-12-RD drivers in free air. These LE5 mids really behave more like giant tweeters than middrivers. They would fit in well with the old Wharfedale open baffle speaker with its up-firing 3" "tweeter".
Numerous hours were spend with the LspCAD to make 2nd, 3rd or even a 4th order filters work, but it didn't - or it turned out so elaborate that it was simply too much.
The solution to this very high Fs came from the recent
C17 work and the High Efficiency Speakers, Part II. Working on 1st order filters often takes a notch filter to flatten the impedance and a LCR circuit is actually used to create the acoustic roll-off needed for the LE5-12.

Another thing about the LE5 mids - and this is the last one - is consistency. As can be seen on the graph above, the impedance of these two drivers are not exactly the same. 1 ohm difference in basic impedance and again - this is getting boring - the LE5-2s are even worse. I have no idea why JBL wasn't able to make voice coils with the same DC resistance. It doesn't make sense. I mean, you wind so and so many turns one way and so and so many turns back again to make a two layer voice coil. How tough can this be? Nevertheless, I've had three pairs of LE5 drivers in my time and none of these have neither measured the same, nor sounded the same. I guess it's difficult to make a very rigid suspension the same all the time, in particular when the stiffness of the suspension is partly determined by a manually applied layer of coating. A fabric spider and a rubber or foam outer suspension is likely to produce the same compliance all the time.

Seen above is the very simple 2nd order crossover for the LE5-12-RD middriver. There's not much we can do about the peaks at 6 and 8 kHz but they don't appear to ruin the treble response.
Whether the points of crossovers are going to be 450 Hz and 3.5 kHz is another question, but at least this approach makes the troublesome LE5 suitable for a 3-way. And the good thing about the LE5 drivers in general is that they are fast as lightning. Very lightweight paper cones and very strong motor systems.
Before we go to the crossover fine-tuning, let's take a look at the cabinets.

The Cabinets

Restoring these old cabs takes at least the same time as making a new pair. Quite some work!

The old front panel was cut off by a jigsaw and the remains removed with the router (router bit with a wheel).

The middriver takes a cabinet and this one is 140 x 140 x 96 mm made from
12 mm side panels and 16 mm bottom panel producing 1.1 litre internal volume.

Bracing made from 12 mm MDF was added to the cabinet as seen on the photos.
And I had to cut out for the mid enclosure. Fillets were added to support the front panel.

The LE25 cut-out needs "ears" for the terminals and these are not 45 deg to the horizontal alignment. Put the tweeter into the hole and mark the placement of the terminals to route for the "ears". The middriver really needs serious chamfering to allow free air ventilation. Route 45 down to 5 mm from the front as seen on the photo to the right. I can't stress the importance of this enough. The LE5 mids have almost all the ventilation to the sides and need a lot of chamfering to breathe and not sound congested.

The rear panel is attached to the side panels by a notch in a groove and can actually vibrate a lot. Loads of PVA glue was added and I left it overnight to dry. The hole for the terminals were chamfered as seen on the photo to the right and a 19 mm MDF block was glued to the inside for the new terminals.

I like the original oak veneer and this was one reason for keeping the olds cabs. Adding bracing and make the best of it. The cabinet panels are made from chipboard and I see chipboard coming back in a few constructions (Living Voice). The veneer was sanded gently and is ready for a new oil rub.

The vent?
You may already have wondered where the vent has gone. Well, the 125A bass driver has a high Qt and is really not suited for bass reflex designs. So I've decided to try an acoustic vent in this construction. These are available from Scan-Speak and I'll add a single vent on the rear panel behind the bass driver. If you want to know more about an acoustic vent, I suggest you read the 2006 isssues of HiFiWorld where the Dynaco A25 has been revived with the use of an acoustic vent. Usually an acoustic vent provides a more amplifier friendly impedance profile. We'll see how it performs.
If it doesn't work I can always remove the vent an route a panel with a vent to be inserted in the rather large hole that's needed for the acoustic vent.
Should you decide to stay with the vent, here are the dimensions of the original:

Front panel driver layout and port

Left: As always: The placement of the drivers have a major influence on the response of the drivers
and I suggest you follow this layout to make the crossover work properly. In particular this is
important for the middriver as seen above. The front panel was made from 22 mm MDF.

Right: The JBL L26 vent made from cardboard tube with an internal diameter of 68 mm..
The bent tube is very suitable for long vents.
The cuts have to be 90/4 = 22.5

More pics from cabinet rebuild:

Gluing front panels in place.

All internal panels except rear panel were added 10 mm polyester foam.
Rear panel had 30 mm "wave" polyester. Additional MDM3 poly/wool will be used for damping.
The crossover will be on small MDF blocks over the terminals in the middle of the cabinet.

Rear routing for the acoustic vent. Right behind bass driver. Centre 120 mm from bottom.
To the right: Front panels after adding laquer for saturating the MDF to allow the use of water based paint.

The cabs were finished during the summer holidays and the sanded veneer was given a fresh oil rub. Front and rear panels had a dark grey colour and to put a long story short, I'm very pleased with the result. And it's very much time to finalise the crossover after a couple of months' break. Gut feeling tells me the 800-3500 crossover option is the most likely to succeed. Gut and gut.....a friend came by with a pair of JBL L100 Century speakers a month ago and I spend quite some hours recording what they were doing. This will be reported later, but what was apparent was the point of crossover between bass and mid. The LE5 drivers really aren't meant to go very low due to high Fs and much of the JBL sound is derived from the bass handling lower and middle midrange. The LE5s are more like mid-treble hybrids and should be treated accordingly.

Due to the bracing, the cabs appear non-resonant and I can't wait to hear how the drivers will perform under these new circumstances.



Left: In all of the midrange and up to 10 kHz a smooth response is achieved and minimum phase tells it all seems to work fine.

Right: SPL including merged nearfield reading of bass response. Disregards the 200-300 Hz region - all things can be achieved when doing nearfield measurements at 1 mm distance to the cone.
Blue in including the tweeter notch filter.

Left: Mid, tweeter and summed response. Green = reverse polarity of tweeter displaying nice phase tracking between the two drivers.

Right: Inpact on frequency response of RC circuit bypassing bass series coil. It's impotant that the huge 2-3 kHz peak from the bass drivers is reduced.

Impedance of system. An overall easy load on the amplifier.

That JBL sound!

Connecting the first speaker from the test crossover set-up, my first reaction was: Well, I'll be damned - or something similar in Danish. My surprise was due to the treble. My expectations from the LE25 tweeter had been so and so. This tweeter is much better than measured performance would suggest. Sibilance is handled gracefully and there's an overall ease in presentation. Hmm... a large, rigid paper cone... It takes a minor equalisation to get the balance right and there's not much we can do about the 12 kHz not

ch - except for lowering the response in the 10-15 kHz range a little. You can increase tweeter level by 1-2 dB and yes, there's too much treble, but it doesn't become bothersome in the way most dome tweeters do. Strange. Does this tweeter really have such low distortion? Whatever it does, it points to the fact that we should take great care with what a tweeter does at 3-10 kHz and what's beyond may be less interesting. Oh yes, the 10-20 kHz provides the airiness to the overall sound, but whether it has to be ruler flat is a good question. Supertweeters? Honestly, forget about supertweeters and let's talk about the things that matter.

My most troublesome, sibilant female vocal recordings? Probably the best speakers I've had with regard to this particular problem. Disturbing!
Having both speakers running, some further mid and tweeter attenuation adjustments were done, but nothing suggested a second round of crossover tweaking necessary. Mid attenuation may be further improved, currently running from a 4R7 series resistor giving a flat upper midrange with lots of detail. The LE5 mid appear well tamed and it's presenting upper mid and lower treble with delicacy. Not without some JBL colouration, but this was expected.

This speaker can play loud, really loud. Almost as loud as the old SEAS 503 kit - with it's 13" bass driver. There's nothing like a big bass driver! Due to the acoustic vent, the bass is firm, solid and up in level all the way down to 50 Hz, from where it starts rolling off.
Initially I played the speakers from a 6W SET amp, and at average listening level everything went well - actually with an amazing level of transparency. Cranking up the volume, the poor SET started clipping. No wonder, the JBL-3W may have 89-90 dB sensitivity, but it can handle a lot of power. The 50 wpc Copland power amp was a more adequate driver and with an old 250 wpc Parasound power amp the bass became even more rock solid. You need a decent amplifier for these speakers, not because of inefficiency, but because you can't help playing them loud.

Detailed descriptions of sound is boring and I'll stop here. If you love that JBL sound and want to hear what a modern filter can do to the JBL sound, you may try what's been described here. It's some journey and some of the effort going into such a project has to come from the satisfaction of restoring vintage equipment. I'm certain some of the JBL drivers hold qualities worth the effort and the end product can easily compete with modern equipment if not sometimes leave it way behind.

Crossover Components

Crossover kit available from Jantzen Audio, contact@jantzen-audio.com

Please state where you live for full quotation incl. shipping.

All kit and component prices may be subject to change and are always to be confirmed by Jantzen Audio Denmark.

Download Kit Sales presentations here  (pdf file)

All technical questions to me at troels.gravesen@hotmail.com


All crossover boards.

Bass section.