SOLVED - Need help with Focal Solo6 Be repair?

[chilidawg]

On a side note, while I was checking all the components near those two problematic transistors, I found a rectifier diode in parallel with a ferrite bead that tested NULL/SHORTED in both continuity and diode mode. A similar rectifier diode in the reference working board, in the same place, also yields the same result. Both diodes are not dead, because I'm sure if I lift one of their legs, they will check out ok. Wonder what's the purpose of it being there.

 

[chilidawg]

UC3842 is a power supply controller. With that big toroid the amp definitely does not have a switching power supply, so don't see how a UC3842 could be used in this amp.



I couldn't figure out what you were saying at first, but from what I gathered elsewhere there is something like a switching power supply which supplies a varying supply voltage to the linear amplifier to improve efficiency of the amp. If something happened to that supply it could cause problems for the amp. You should be able to see by looking at the power supply rails for the bass amp if the voltage is stable, if the switching frequency looks reasonable and stable, etc. I would think that with no input the supply voltage should be pretty low and stable, and as you give an input signal the supply rail should increase as the input signal amplitude increases. Would have to go look for the BASH patents to be sure, but the description Focal gives indicates that the supply voltage follows the input signal pretty closely. I'm not sure if the UC3842 can track reference voltage changes that quickly, but the picture you showed is definitely an ST part and not a UC3842, so if you thought you were taking a picture of something with a UC3842 on it I can't see it.



So you completely disconnected the input circuitry with crossover from the amplifiers? Your wording was a little big vague on whether you actually were able to disconnect enough to make absolutely sure the noise was from the amplifier and not the input/crossover circuit.

The input and crossover circuit is connected to the amp board that has tweeter amp and a woofer amp via a 6 pins cable, so yes when I was looking where the noise is coming from, it was completely disconnected.

I tested the tweeter amp first, and it was quiet. Turned it off, desoldered the tweeter speaker cables, and I soldered back the woofer speaker cables to where they're supposed to be. After I've done that, and I turned it on, et voila! the hissing and crackling noise comes out of the woofer. I can also see the woofer cone moving like that automatic breathing oxygen pump you see in the hospital.

That part is definitely a UC3842. I used a microscope to see the part number, despite their effort in trying to obscure it. That little board is a proprietary switching controller. Has a few SOT-23 parts on it, with the laser marking on the surface also obscured, so you won't be able to tell if they're diodes or transistors, because INDIGO also removed the part designation from the PCB design.

All BASH amplifiers have something similar in it. They all numbered with 6600 in front. Feel free to ask INDIGO about it, but I doubt they will tell you anything.

 

[chilidawg]

If the IC isn't UC3842B, then I don't know what it is, because I also ran the search for 38428, EZ052, E7052 in STMicrolectronics' parts database but came up with nothing.

A high res bird's eye view of both mainboards. Don't quite get it why they put thru hole transistors and fast switching zener diodes when they literally have the SMD version populating everywhere else on the board (better thermal performance?)

Anyway, all the components located nearby the transistors that overheated checked out ok and are of the same values when compared with the reference working board. I'm probably gonna replace that 22uF/35V SMD electrolytic close to the toroid inductor because it was exposed to the heat coming from the TIP112 heatsink, and it was very hot.

 

[ccaudle]

I can also see the woofer cone moving like that automatic breathing oxygen pump

That is an important clue, because it implies very low frequencies or changing DC levels somewhere in the amp, in addition to the audible noise you hear.

I found a rectifier diode in parallel with a ferrite bead that tested NULL/SHORTED in both continuity and diode mode.

Shorted at DC. Usually when a diode is in parallel to an inductor it is because there is a switching element in the path somewhere, and when the switch turns off, the current flowing in the inductor has to go somewhere. If there is not a diode path for the current, the voltage keeps increasing across the inductor until it eventually causes a breakdown somewhere so the current can flow. You will see them variously referred to as "catch" diodes, "flywheel" diode, "freewheel" diode, basically all informal terms trying to indicate that the diode is there as a fall through path for the current once the original path is cut off.

That part is definitely a UC3842

OK, interesting use of that part. Still, it seems like without any audio input for the power supply to follow, it should simplify to a class AB amplifier stage, with a switching power supply providing the power, so you can check the supply rails to see if they are stable, and check the amplifier to see if the bias is stable, if the transistors appear to be functioning properly, etc. With a switching supply involved you will definitely need a 'scope and not just a DMM.

Interesting that one of the PCB's is marked as Sonavox Canada, but Sonavox does not appear to have a current Canadian location. Probably out of luck trying to get any info from them at this point.

This thread on DIYAudio indicates that a lot of the BASH amps had problems with one of the adhesives used to hold down the heavy components becoming conductive over time:
BASH repair thread on diyaudio

No idea if it is involved in your problem, but not hard to find threads in different places that are all variants of "my BASH amp stopped working."

 

[chilidawg]

That is an important clue, because it implies very low frequencies or changing DC levels somewhere in the amp, in addition to the audible noise you hear.



Shorted at DC. Usually when a diode is in parallel to an inductor it is because there is a switching element in the path somewhere, and when the switch turns off, the current flowing in the inductor has to go somewhere. If there is not a diode path for the current, the voltage keeps increasing across the inductor until it eventually causes a breakdown somewhere so the current can flow. You will see them variously referred to as "catch" diodes, "flywheel" diode, "freewheel" diode, basically all informal terms trying to indicate that the diode is there as a fall through path for the current once the original path is cut off.



OK, interesting use of that part. Still, it seems like without any audio input for the power supply to follow, it should simplify to a class AB amplifier stage, with a switching power supply providing the power, so you can check the supply rails to see if they are stable, and check the amplifier to see if the bias is stable, if the transistors appear to be functioning properly, etc. With a switching supply involved you will definitely need a 'scope and not just a DMM.

Interesting that one of the PCB's is marked as Sonavox Canada, but Sonavox does not appear to have a current Canadian location. Probably out of luck trying to get any info from them at this point.

This thread on DIYAudio indicates that a lot of the BASH amps had problems with one of the adhesives used to hold down the heavy components becoming conductive over time:
BASH repair thread on diyaudio

No idea if it is involved in your problem, but not hard to find threads in different places that are all variants of "my BASH amp stopped working."

Thanks for the explanation about flywheel diode.

From what I gathered, apparently Sonavox is NOW a Chinese company, which is not surprising at all, because my first impression when I opened it up to change the main rocker switch at the time was "the expensive price tag doesn't justify the cheap construction." heck, even the PCB is of low quality, as proven by how easily I destroyed a plated thru hole in just a matter of seconds, and my soldering tip wasn't even that blazingly hot, and I used a lot of flux with low melting point solder to get the factory solder flowing. The $5-$10 prototype PCBs that I ordered from JLCPCB.com and PCBWay.com have a much tougher plated thru holes.

Not sure about a certain adhesive becoming conductive, all the adhesives they used on this board are what you will normally find in switching power supplies and adaptors (thermal silicone glue and a white colored glue that's effin' tough to remove or scrape which holds large electrolytic capacitors/toroid inductors in place)

 

[chilidawg]

Successfully repaired the monitor.
Hissing and crackling noise issue with the woofer amp is gone.
IRF540 and TIP112 replacement aren't even warm

 

[ccaudle]

Successfully repaired the monitor.
Hissing and crackling noise issue with the woofer amp is gone.
IRF540 and TIP112 replacement aren't even warm

So damaged transistors? Replacing the MOSFET and the bipolar was all that was needed?

 

[chilidawg]

So damaged transistors? Replacing the MOSFET and the bipolar was all that was needed?

Actually, a few hours after I edit the title thread to say that I got it fixed: the noise problem returns but the replaced parts are not overheating anymore.
Now every time when the monitor is switched on, the woofer makes a loud pop, the cone moves forward and stays in that position until I switch it off.

See attached video. The hissing noise is very loud, overpowered the sound from my TV.

 

[ccaudle]

the cone moves forward and stays in that position until I switch it off.

That makes it sound like you haven't even put a meter on the amp yet. Cone moving forward would be DC present on the amp output. How much? Did you ever put a scope on the power supply rails of the amp output stage?

 

[chilidawg]

That makes it sound like you haven't even put a meter on the amp yet. Cone moving forward would be DC present on the amp output. How much? Did you ever put a scope on the power supply rails of the amp output stage?

2.8V DC at the woofer amp speaker output.
-5.0V DC at the source pins for IRF530.
2.9V DC at the source pins for IRF9530.

Haven't checked with the other working monitor if the voltages are the same. Will do that later tonight.

Don't have a scope.

 

[chilidawg]

I should also tell you that at this point the replaced Darlington transistor (TIP112) started to overheat very fast again, and I didn't even do anything to the board at all. What the heck is going on?!

 

[Khron]

Treating the symptom rather than the root cause rarely leads to good results, unfortunately... Don't ask me how i know

Not sure you can get away without reverse-engineering at least part of the circuit.

 

[chilidawg]

Treating the symptom rather than the root cause rarely leads to good results, unfortunately... Don't ask me how i know

Not sure you can get away without reverse-engineering at least part of the circuit.

I totally agree with you, hence I'm still chasing it down.

 

[chilidawg]

Aside from the output MOSFET source voltages, I found another discrepancy.
There's a 47uF/35V electrolytic in the woofer amp section that the positive pin measured less than 1V (0.960mV) when compared to the reference board, that is 13V in difference! Negative pin is -14.9V, likewise with the reference board.

-55.6V at TIP112 emitter pin for all boards, but only this one is overheating.

All values measured with loads attached, and no input signal.

 

[chilidawg]

Ok, basically after taking a lot of measurements from both boards, here's what I know so far:

Only the positive signal amplifier is not working correctly.
Some switching transistors and diodes have different voltages compared with the negative signal amplifier and also compared with the reference board, like for an example what should be a negative voltage instead became a positive voltage. Got my head scratching a bit as to why that could happen.
DC offset at the output occurs because of the imbalance.
Power supplies are ok (both primary and secondary)

I feel I'm getting close to pinpoint the true culprit. Just have to take my time.

 

[ccaudle]

-5.0V DC at the source pins for IRF530.
2.9V DC at the source pins for IRF9530.

That doesn't tell a lot without knowing the circuit configuration.

Don't have a scope.

As I mentioned before is it somewhere between "very difficult" and "can't be done" to debug a switching power supply without a scope. Since the entire principle of operation relies on dynamically switching states, you can't tell much about how it is operating if you can't view the states and the transitions between.

Only the positive signal amplifier is not working correctly.

That is a strange phrase, what exactly do you mean by that? Are you referring to a complementary pair style amplifier, and only the negative voltage side is operating?

Some switching transistors and diodes have different voltages compared with the negative signal amplifier a

If they are truly switching transistors then all you can see is the DC average, so again you can't really tell much about what is wrong. If they are actually operating as linear devices then you can measure the base, collector, and emitter to see if they are biased correctly. You can almost do the same thing with MOSFETs by measuring the voltage at gate, drain, and source, but the on/off threshold has so much more variation than a bipolar junction that you have to do more work to determine if it is biased correctly or not.

Power supplies are ok (both primary and secondary)

I would put an asterisk there if you don't have a scope. Since the power supply for the bass amp is generated by a variable switching supply, I would consider noise and stability of that supply to still be an open question, or at least partially open.

 

[chilidawg]

That is a strange phrase, what exactly do you mean by that? Are you referring to a complementary pair style amplifier, and only the negative voltage side is operating?

If you look at where the woofer's positive lead (red cable) and the negative lead (black cable) are connected to, what do you see?

because I see a layout and components that are copies of each other. Does this indicate a complementary pair? Idk. All I know is that when I checked the monitor that has no issue, voltages that appear on both sides are identical, within a very tight 0.1V tolerance.



Blue lines surrounded the sections where the voltages are very different.

For the sake of simplicity, I named the two Solo6 monitors as GOOD and BAD.
And some measurements:

U501 near the 47uF/35V capacitor, which is a TL082C op-amp.
Pin 1 is connected to a resistor (R513) in series with a BAS16 switching diode (D514)
GOOD: pin 1 is -13.6V, anode -13.6V, cathode -13.9V
BAD: pin 1 is +13.2V, anode +1.6V, cathode 0.9V

Q309 next to the trim pot P300, a MMBT4401 PNP transistor.
GOOD: base +1.65V, collector +13.2V, emitter +1.1V
BAD: base +4.0V, collector +14.9V, emitter +12.6V

D307, a BAS16 diode.
GOOD: anode -13.2V, cathode -13.8V
BAD: anode -7.2V, cathode -7.98V

Not identical.

On the negative lead side,

Q31? something next to trim pot P301, also a MMBT4401 PNP transistor.
Both monitors: base +1.65V, collector +13.2V, emitter +1.1V

D316, a BAS16 diode.
Both monitors: anode -13.2V, cathode -13.8V

Identical.

I took these measurements many times, in different sessions. Always the same result, which is why I believe that the power supply is ok, and the negative lead side is also ok.

If they are truly switching transistors then all you can see is the DC average, so again you can't really tell much about what is wrong. If they are actually operating as linear devices then you can measure the base, collector, and emitter to see if they are biased correctly. You can almost do the same thing with MOSFETs by measuring the voltage at gate, drain, and source, but the on/off threshold has so much more variation than a bipolar junction that you have to do more work to determine if it is biased correctly or not.

Sorry, I meant to say switching diode. No idea if the transistors are also being used for switching.
Aside from the MOSFETs, they are MMBT4401, MMBT4403, MMBT5551S, MMBT5401S.

 

[ccaudle]

layout and components that are copies of each other. Does this indicate a complementary pair?

No, looks like a bridged output. Basically the speaker level version of a symmetrical balanced output. So yes, agree, you should be able to check between the + and - speaker leads and both sides should be identical (at DC, they move in opposite directions at AC).

The measurements you have show that the NPN transistors on the bad side are obviously not biased on, but not enough information to tell why. Sketching out the circuit would probably make it more obvious where the fault lies.

The picture problem_board_caps looks like possibly the negative supply is not running (around -14V on the good amp, only -1V on the bad amp). Can't tell just from the pictures whether those are power supply bypass or in the signal path, though, so could just show that the bad amp is way off on DC levels. Do you know the function of those caps in the circuit?

Q309 next to the trim pot P300, a MMBT4401 PNP transistor.
GOOD: base +1.65V, collector +13.2V, emitter +1.1V

That indicates a properly biased NPN (not PNP) transistor, emitter is 0.55V more negative than base.

BAD: base +4.0V, collector +14.9V, emitter +12.6V

That looks like the upstream circuit is trying to bias the transistor on, and is able to get the base up to 4V (compared to other side, so apparently around 1.6V or 1.7V is the normal level), but the emitter has 12V on it for some reason, so can't be biased on. Without knowing what is connected to the emitter no way to tell why it is so far off. Need to sketch out the circuit so you can get a rough estimate of what the voltages should be, then it will probably be easier to see where things go wrong.

 

[chilidawg]

Wanna know the real culprit is? A resistor that is supposed to be connected to a capacitor with vias and a bottom layer track for an odd reason isn't connected anymore, even though the vias and the track are not damaged at all.






In continuity mode, GOOD board gives the long infinite beep, and R311 is measured in circuit approximately 15K.
BAD board no beep and R311 is measured in circuit approximately 100K.

Here's TL082C pin 1 again, measured +13V, when it's supposed to be -13.6V



Added a solid jumper wire to remake the broken connection. If the MOSFETs aren't in the way, I could have done a better job, but this is pretty much "nice and clean".



Here's TL082C pin 1 again for the last time, measured at -13.6V


Re-checked all the voltages, they're now identical to the negative lead side and identical to GOOD board.
No DC voltage at output (measured 0.02mV)
Connected the woofer, tweeter, and no loud pop when switching it on/off.
No hissing, no crackling, and woofer doesn't get stuck in forward position anymore.

 

[Khron]

If anything, you could get away with maybe half that length of jumper wire, and lay it flat across the board.

Or, alternately, strip enough insulation off the two ends, that there's enough (bare, tinned) wire to poke through those two vias from the bottom of the board, and reach the components it's connecting (effectively "being" what the pcb trace used to be).

But great job tracking down the actual issue!