OT: Si, mas, C-MOS run, see C-MOS crap out - tales from the bench

[Mark Hammer]

I've plugged in my soldering iron maybe twice in the last 2 months, once to fix a speaker cable for the speakers on my wife's computer, and last night to fix my own.

The speakers are some decent-for-their-size-and-time IBM computer speakers that I had picked up from a "clearance" table at a store for $10 a while back.  Pretty typical 2-way-cap-for-a-crossover-with-a-4"-woofer speakers, powered by a TDA1552.  The unit has a subwoofer-out jack so I figured there had to be an active filter in there somewhere, but I had never been able to find a trace of an audio output signal from the jack.

Over the past few months the line-in mini phone-jack had been getting very intermittent.  So much so that the cable from the computer had to be angled just a certain way, the wire folded under the cabinet, etc. etc.  I'm sure you've all been in a situation where a fickle connection obliged you to do something ridiculous just so the sound would stay on.  So, I decided that enough was enough, and I would replace the jack or repair it somehow so I could maintain the flow of music. 

I popped the back off the main speaker with the electronics inside, extracted the boards and gave it the once-over (which is how I learned about the TDA1552 - pretty powerful chip for such a tiny speaker - and the LM1036, which it uses for managing processing controls).  Fortunately, I had an old Soundblaster card I didn't need, and extracted one of the 4 phone PC-mount jacks as a perfect replacement for the line-in jack on the speaker.  Dabbed a little Stabilant on the contacts for good measure. 

The subwoofer output jack was a dual switching type with what seemed like 8 or 10 pins crammed into a little space.  Some solder tabs went to "nowhere pads" (little islands of copper).  The subwoofer output was clearly supposed to be mono (as all subwoofers are), but there was a heckuva lot of switching going on there.  Clearly, plugging a phone plug uinto that jack didn't simply make available something that was on all the time.  Rather, plugging into the jack made something happen. 

Looking around the board, I spotted a 4066 quad C-MOS switch chip.  Hmmm.  Ve-ry in-te-resting.  Was there anything in the circuit that required switching?  Nothing that I could see, EXCEPT for whatever the subwoofer jack was doing.  The compactness of the board and lateness of the hour made it difficult to immediately see the physical connection between the jack and chip, but given how many times I've spotted repairs here and on Ampage that began and ended with a bad CMOS chip, I figured maybe that was my target.  Quite likely the contact switching on the sub jack engaged the 4066, which then routed bass normally sent to the onboard speakers to a subwoofer.  IN other words, with nothing plugged in, you get the whole signal coming through the speakers, and by plugging into the subwoofer jack, you engage an active filter normally not in the signal path.

I flowed a bit of extra solder onto the pads of the chip (always good for aiding component removal), put my solder sucker into action, and extracted the chip.  I installed a socket, popped a new 4066 into the socket, re-assembled the speaker, plugged in a subwoofer, fired it up, and bingo - throbbing bass to irritate my wife and scare our bunny.  And the speakers stay on, even when I jiggle the cable.

CMOS:  can't live with it, can't seem to have enough in the parts bin.

[scratch]

very interesting indeed ... I would not have thought that switch chip would sound so 'intermittent' like bad patch cords ... good to keep in the back of the mind!

now, I'm curious, if you were going to chuck the 4066 chip anyway, why not just clip the leads at the body? then removing the lead 'leftovers' via needle-node/solder iron way easier then removing the chip whole ... just a thought.