Subwoofer amp repair - power and signal but no sound

[crxguy52]

I replaced the capacitors in my subwoofer amp per this excellent thread. I verified the capacitors I was replacing were the correct value and voltage rating, confirming the information in that thread. I re-assembled everything and plugged it in, but I'm not getting any sound output. It's an improvement, as before I was getting the typical high pitched screech.

Using a lab supply and signal generator (800 Hz, 1V pk-pk sine wave) I probed around the amp while it was powered up. I verified that the input signal is reaching pins 1 and 2 of the left vertical IC. I also verified that the FETs are receiving a gate drive signal and are switching at ~140 kHz, but seem to be stuck at 50% duty cycle regardless of amp input. This seems to indicate that either one of the ICs are bad, or a trace connecting the 2 ICs got destroyed when the electrolyte leaked out of one of the capacitors. I repaired one trace, visible under the large inductor in the pictures below, with a fly wire on the bottom of the board.

1. What is a typical input signal magnitude to the amps? I don't want to overdrive the input and blow it out during testing
2. Does anyone (@Briank [MENTION=25737]Kaz-kzukNA1[/MENTION]) have pictures of the top of a bare board so I can verify continuity without pulling everything off? I'm thinking I may have missed a destroyed trace under one of the inductors or ICs
3. Does anyone know the signal path from the first IC to the second IC? If either IC is dead, I may be looking for a new amp all together.

 

[Old Guy]

The following may provide some useful information for you. Its specific to Bose amps on the Corvette; but, the Bose amps all seem to be pretty close in construction. The document defines the signal path. Most useful are the pin numbers on the ICs which you can use for checking for the presence of the input signal.

http://www.zr1netregistry.com/Portals/0/Bose2.pdf

I suspect that you are probably aware that this is a class D amp with the Bose 134499 chip driving the 4 FETs in an H bridge configuration. Being an H bridge you will need to connect a speaker or a suitable load resistor across the audio output if you want to test for an output signal.

Edit - you commented about the FET gates receiving the 140 kHz signal PWM signal at 50% modulation. Are all 4 gates receiving the gate signal simultaneously or are just the diagonally opposite pairs in the H bridge being gated at the same time? If all 4 FETS are being gated simultaneously or any two FETS on one side of the bridge are being gated simultaneously that will result in a direct short from +12v to ground and will almost assuredly result in destruction of the FETS. That could also be an indication of a significant problem with the PWM driver circuit. Gating of the diagonally opposite FETS would be sort of consistent with correct operation; but, why they are being modulated at 50% is a mystery.

 

[crxguy52]

Excellent reference, thanks! I spent maybe 10 minutes looking at it today with the sub connected, and it seems like the signal is leaving the equalizer chip and entering the FET driver as expected. I checked the gate drive timing, and they're switching in diagonal pairs as you might expect - no shoot through. It makes sense that they would be 50% duty cycle at zero output, it would be equivalent to 6V on either side of the speaker, and thus zero potential across the speaker. This would allow for bidirectional current through the speaker with only a single ended supply.

I'm at somewhat of a loss since everything "looks" right. I'll read through that documentation and poke around a bit more, but I'm not sure which direction to go in at this point.

 

[Old Guy]

I am a bit out of my league on class D amplifier details. My experience is more with PWM motor control. However, your comments about both bridges being gated when the amp was quiescent got me curious because it seemed odd for a simple H type bridge configuration. Turns out that the Bose amp is likely a type of class D amp that utilizes filter-less modulation. That type of design would explain why both sides of the bridge are getting gate signals causing them to sit at the same voltage when the amp is quiescent. It also explains why the amp has no explicit low pass filter between the bridge output and the speaker although it does have the two inductors that appear to be in series with the speaker to alter the inherent cross over frequency of the . This technical document from Maxim provides a rather brief overview (about 1/2 way through) of the filter-less class D amp topology which looks remarkably like the Bose amp.

https://www.maximintegrated.com/en/app-notes/index.mvp/id/3977

The only reason any of this may matter is that the filter-less topology involves output feedback. I haven't spent enough time looking at the details to be sure; but, I am thinking if you don't have any output load connected you may not see any change in the gating signals in response to an input signal. I know that you originally tested with everything put back together and got no output. It might be worthwhile to confirm that the speaker works or that there wasn't a bad connection. If you are bench testing the amp try doing that with the speaker connected and confirm that the connection is good. Right now, I am kind of thinking that if you are measuring your 800 hz test signal at the input to the 134499 chip and both sides of the bridge are being gated to give equal voltage (0 volts across the speaker), something has happened to the modulator function in the 134499. There don't appear to be any external signal loops for the 134499 that could have been damaged (unlike the 121661) so that suggests an internal problem in which case I think you are dead in the water. Perhaps carefully inspect the solder pads for the 134499 to make sure that you have not lost a connection due to corrosion damage from leaking electrolyte. Beyond that, I don't really have anything useful to add.

I did the Hugo window kit upgrade on my car last summer and pulled one of the door speaker enclosures apart to see if it was showing any evidence of leaking capacitors. Everything looked fine so I decided to let sleeping dogs lay. I am going back into the doors this spring to replace a number of the trim pieces around the edge of the interior door panel because somebody had managed to snap off most of the hold down tabs when they messing around in the interior of the door (explains some mystery rattles). I am thinking that I should probably just bite the bullet and replace the capacitors. If I do that, I will try to remember to go in and measure the signal path and operating voltages on the operating amp so that there is a base line which people can use as a diagnostic tool to determine whether either of the ICs are dead.

 

[crxguy52]

Old Guy,

Thanks for the input! I removed the subwoofer from the enclosure and hooked it up to the amp, then drove with the 800 Hz sine wave. Unfortunately I experienced the same result, no output on the subwoofer. I was able to drive the sub directly with the signal generator, so I believe the subwoofer is functional. I'll take a look for output feedback to the 134499 and see if I'm missing something. I may remove the caps again and look for any broken traces underneath the output inductors - I may have missed something there. If there isn't, the 134499 chip may be dead. I'll post back with my findings.

 

[Briank]

Good try on fixing it, sent you an email reply so you can get it to me for repair.