Tuesday, May 31, 2016

Audio breakup on the JBL 4315B speaker

Over the past several weeks I noticed a problem with one of my JBL 4315B 4-way speakers:  It sounded as though the woofer was "breaking up" - that is, there was occurring what sounded like some sort of distortion related to the travel of the cone in response to "heavier" bass content.

Figure 1:
The "business end" of the 4315B

Some background:

A year or so ago I had a similar problem with this same woofer:  The audio was breaking up very badly, so I put it on the workbench, connected it to the amplifier, and observed the same problem.  Having nothing to lose, I carefully removed the dust cover - a job made easier with the careful application of a heat gun set to "low" - and quickly saw that an aluminum stiffening ring (constructed with an open gap in its perimeter) just "outboard" of the voice coil (e.g. toward the "front" of the woofer) had broken loose from its original adhesive.

With this loose ring - or at least a portion of it being loose - the speaker's excursions via the force of the voice coil allowed a bit of physical distortion, causing the voice coil to rotate slightly away from its axis, permitting it to hit the magnet assembly.

To fix this I laid the woofer on its face on the edge of a workbench, clamping it down to prevent it from falling off (and on to me!) and working from below to prevent debris from entering the magnet gap I removed the aluminum stiffening ring and scraped away the old, brittle adhesive from the inside of the voice coil assembly.  After this I righted the speaker and stuffed pieces of paper towel into the voice coil opening to prevent additional debris and uncured epoxy from getting into the coil-magnet gap.  To reinstall the metal ring, which I'd also cleaned of old adhesive, I used metal-filled epoxy ("J.B. Weld") making sure that the ring was straight and the adhesive uniformly distributed.  To assure that the ring would bond properly I jammed a small screwdriver in the gap in the ring to widen and wedge it tightly into place while the epoxy cured.

Figure 2:
More or less what the resistors looked like
when the cover was removed.
Click on the image for a larger version.
After allowing 48 hours for the epoxy to fully cure I again inverted the speaker, removing the piece of paper towel that I'd put in place to prevent debris and epoxy from fouling the voice coil.  Subsequent testing indicated that the woofer was, again, working normally so I replaced the dust cover with a new one, identical in size to the old, that I'd ordered previously, placing the speaker back into service.

Unfortunately it did not occur to me to take pictures of this repair until I was well into it, hence the "thousand words".

Back to present day:

Everything was working fine until recently and this recent 'bout of distortion initially let me to believe that the stiffening ring has broken loose again, so I removed the woofer from the enclosure and the distortion suddenly went away.  This immediately indicated to me that it was unlikely that the stiffening ring had come lose as the speaker now in free air rather than in the tuned cabinet moved more freely that before outside its enclosure.


Suspecting that gravity may have caused the cone assembly to sag over time - a problem not too uncommon with larger drivers with foam surrounds - I rotated it 180 degrees, placed the woofer back into the speaker and the distortion reappeared immediately.  While the distortion was occurring I pressed gently at different places around the edge of the cone to see if I could cause it to get worse but I didn't find anything obvious:  Doing this - and finding one spot on which when pressed causes sudden, severe distortion - can be a helpful diagnostic to determine if the cone is off-center, either due to gravity-related sagging or some sort of damage to or alignment of the "spider", voice coil assembly and/or the surround.

Figure 3: 
This picture shows where the resistors' leads broke,
right at the body of the resistors.
At the top of the picture are some of the plastic
capacitors used in the crossover which seemed
not to be affected by the vibration.
Click on the image for a larger version.
Puzzling over this problem for a moment, with the woofer laying on the floor next to the speaker, I happened to notice that one of the plastic grommets emerging from the metal crossover box inside the speaker had popped out of place so I snapped it back into the hole - an action that was coincident with a sudden bout of distortion in the woofer.  On a hunch I started whacking the grille cover of the crossover with the handle of a screwdriver and observed that every time I did so I could hear a sort of "clicking" in the woofer.

Inside the '4315, just behind the woofer, a grille covers a series of power resistors and plastic capacitors comprising the crossover (additional components like the inductors are located in a separate compartment behind) and upon removing the cover I noticed that at least one of the 10 watt resistors was at an odd angle.  When I touched this resistor, the problem was obvious:  One of its leads had broken away from its body due to fatigue and the vibration of the woofer and was causing it to make intermittent contact.  Clearly these rather heavy 10 watt power resistors had been vibrating for years, eventually causing the connecting wire to break.

Inspecting the other resistors, the leads of two others broke away from the slightest touch meaning that I simply had to replace them all.

Figure 4: 
The new resistors in place.  The two 10
ohm resistors (brown, tubular) did not
have leads so heavy-gauge wire was
used to connect and support them.
Click on the image for a larger version.
Fortunately, these resistors were nothing special - just plain, ordinary 10 watt, 10% tolerance wirewound resistors of the values 2.4 ohms (1), 10 ohms (2), 20 ohms (2) and 50 ohms (1).  A quick rummage through my collection of power resistors indicated that while I could have "kludged" a repair that minute, I decided that I wanted to replace all similar resistors in both speakers and I made a shopping list for the next time I happened to visit the local electronics place.

A few days later I found myself at the local electronics store,  rummaging through their resistor collection.  I didn't find exactly what I wanted as they didn't have any of the rectangular "sand" 10 ohm, 10 watt units, but they had some of the same in ceramic tubular which would work with attached leads.  They also didn't seem to have any 47 or 50 ohm 10 watt units, but I found that they had plenty of 100 ohm, 6 watt resistors so I obtained twice as many of those so that I could parallel them to 50 ohm at 12 watts.  For the 2.4 ohm resistors I had my choice between 2 ohm and 2.7 ohm and chose the former since the lead length was better for mounting:  Since this resistor was simply in series with the upper midrange ("HF") driver, and there is "T" pad to adjust its level - I figured that this departure in value from the original 2.4 ohm 10% part to a 2 ohm 5% part was not of consequence.  (Various diagrams show slightly different factory values, anyway - none of which matched the factory-installed 2.4 ohm resistor, anyway!)

The repair job was pretty straightforward.  In order to get the slightly larger, ceramic-tubular 10 ohm resistors to work I soldered to them some #12 AWG connecting leads (to provide rigid support and prevent future vibration-related breakage) and folded over the tabs along with paralleling the two 100 ohm units such that the effective lead length was increased (see picture).

Figure 5:
The cover for the crossover in place.
Some "blobs" of black RTV were added
previously to suppress a buzzing caused by
a mechanical resonance of this cover at
certain musical note frequencies.
Click on the image for a larger version.
I quick test revealed that the repair was successful with all of the drivers on the speaker working as they should.

When I replaced the grille cover I was reminded another "fix" that I'd done many years ago:  Some blobs of RTV (silicone) sealant along the edges of the grille cover forming damping pads to eliminate an annoying buzzing noise that it would make when it resonated with particular bass notes.


With the repair of one speaker done I quickly moved on to its mate.  In it, I found no broken resistor leads, but one or two felt very "weak" indicating that metal fatigue had already been at work.

Finishing the resistor replacements I now had both speakers working properly, using "matched" components.

Final notes:

After a quick search I was able to find several different diagrams for the crossovers themselves, no doubt due to changes made during the production.  It was interesting to note that the precise values of some of the resistors in my speakers weren't exactly those on any of the versions of the diagrams which indicates that, perhaps, JBL didn't consider them to be extremely critical and used what was available from their suppliers.  (If one looks at the diagram one can tell that +/-10-15% isn't really likely to make much of a difference in the properties, anyway - at least not one that couldn't be compensated for by the adjustable pads...)

[End]

This page stolen from "ka7oei.blogspot.com".