PSU board, fault hunting

[dark_hawk]

Hi,

I'm trying to repair this PSU board for a Creative S750 surround sound system.
Original fault was a blown diode (D9 1N4004), I replaced it with a 1N4007 diode. I also replaced U6 7920 negative voltage regulator which used the configuration in the schematic to output -24V , I replaced it with a 7924 while removing R20, C68 and R19.
I also replaced all the crap CAPXON capacitors on the board with quality caps.
The board now outputs all the voltages it's supposed to output correctly with one problem.
When you plug the power cord the green light turns on. When you switch the unit on from the control pod for the first time after you plug the cord I hear something like a spark coming from the relay (I think), If you switch the unit off from the control pod (without unpluging the power cord) and then switch it back on, the relay click is normal without that spark sound.
One time I switched the unit off, unplugged the power cord, and then quickly plugged it in the switched it on, and it blew the D9 again.
My question, what part of this circuit can cause such a sparking sound when being turned on for the first time, or is responsible for a soft starting of the relay (if that is a thing)?

Thanks.

Stand by section of the board:


The section that gets turned on by the relay:

[xavier60]

Does D9 go open or short? Check TH1.

[dark_hawk]

Does D9 go open or short? Check TH1.

Thanks for the reply.

It goes short.

The Thermistor seems fine, reads about 10.5 ohm when at room temperature, and quickly drops to around 7 ohm with a light heating of one of it's legs.

[xavier60]

The only thing Im able to suspect is the relay. It's doing a tough job which has possibly caused the contacts to erode causing a lot of contact bounce.

BTW, you have not included your country in your profile.

[xavier60]

Also carefully check the bridge rectifier, being certain that the forward drop of each of the 4 diodes is noted. In other words, none are open.

[Audiorepair]

I have a small PSU I've just looked at using a TOP247 I think, it is not doing anything and I suspect the TOP device.

Interestingly, the R49 snubber resistor in mine is open circuit, so you might want to check yours.
Don't know if this might have killed the TOP device or vice versa.

[dark_hawk]

The only thing Im able to suspect is the relay. It's doing a tough job which has possibly caused the contacts to erode causing a lot of contact bounce.

BTW, you have not included your country in your profile.

If it was the relay, why does it do that spark sound the first time it clicks, and not every time its clicks, as I mentioned it only does that sound the first time after connecting the power cord, every other relay click after that is normal.

The country is Egypt, I don't know why that's relevant .

The bridge is ok, all diodes are measuring fine. Please remember the unit is working fine

[dark_hawk]

I have a small PSU I've just looked at using a TOP247 I think, it is not doing anything and I suspect the TOP device.

Interestingly, the R49 snubber resistor in mine is open circuit, so you might want to check yours.
Don't know if this might have killed the TOP device or vice versa.

I checked all resistors around the TOP243, they all seem fine. Please remember the unit is working, I'm just trying to find why the relay makes that sparky sound first time it turns on.

[xavier60]

The only thing Im able to suspect is the relay. It's doing a tough job which has possibly caused the contacts to erode causing a lot of contact bounce.

BTW, you have not included your country in your profile.

If it was the relay, why does it do that spark sound the first time it clicks, and not every time its clicks, as I mentioned it only does that sound the first time after connecting the power cord, every other relay click after that is normal.

The country is Egypt, I don't know why that's relevant .

The bridge is ok, all diodes are measuring fine. Please remember the unit is working fine

Im thinking that it's to do with whether the capacitors are charged or not, maybe. Knowing your country allows us the figure out the main voltage there. Can you easily remove the relay's cover?

[dark_hawk]

The only thing Im able to suspect is the relay. It's doing a tough job which has possibly caused the contacts to erode causing a lot of contact bounce.

BTW, you have not included your country in your profile.

If it was the relay, why does it do that spark sound the first time it clicks, and not every time its clicks, as I mentioned it only does that sound the first time after connecting the power cord, every other relay click after that is normal.

The country is Egypt, I don't know why that's relevant .

The bridge is ok, all diodes are measuring fine. Please remember the unit is working fine

Im thinking that it's to do with whether the capacitors are charged or not, maybe. Knowing your country allows us the figure out the main voltage there. Can you easily remove the relay's cover?

No, it's potted.
Looks like this one, expect it's 9v.


I'm thinking the same thing regarding charged capacitors. Maybe while replacing the capacitors I made a mistake. But where should I look?
What part of the circuit will cause such a spark or is responsible for suppressing such a spark when it's first turned on.
I think (not 100% sure) that before replacing the caps on the board, when I first got the unit and just replaced the D9 it didn't make that sparky sound. Or maybe it did and I didn't notice.

[xavier60]

Im not blaming the capacitors for being faulty or not fitted right, just the state of charge they happen to be in when the relay turns on.
You could just replace the relay if it's not to difficult to find a replacement.
Im still not certain if it's responsible for D9 shorting. While D9 is shorted, does it cause a problem? I wouldn't expect so.

[xavier60]

The purpose of D9 is to prevent C61 discharging into the 470uF capacitors when the relay first turns on. This would be audible and bad for the relay also.
Is D9 shorted at the moment?

[dark_hawk]

Replacing the relay won't be easy since I can't find it locally, plus I still don't think it's an issue with the relay since all subsequent clicking sounds of the relay after the initial one is normal.

You are correct about when the D9 shorted, the unit continued to function the 0.5 minute I gave it before turning it off, didn't blow the fuse.

The purpose of D9 is to prevent C61 discharging into the 470uF capacitors when the relay first turns on. This would be audible and bad for the relay also.
Is D9 shorted at the moment?

No, I replaced it once it went bang.
How do I prevent C61 from discharging into the 470uF capacitors? What other part of that circuit should I check?

[xavier60]

Im out of clues for now. Only D9 is responsible for preventing C61 discharging into the 470uF capacitors at relay turn on.

[dark_hawk]

Here is a photo of the back of my board, maybe there is something I'm missing.
The board had that conductive glue which was very hard to remove, the result was some broken pads, hence the the jumper wires.



And here is the front of the board (not mine):


Overlayed:

[xavier60]

Take care with that board, the clearances between live tracks and the pour look way too close. Don't do anything that will cause reduced clearances.
 Is that diode on the bottom doing the same as D9?

[dark_hawk]

Take care with that board, the clearances between live tracks and the pour look way too close. Don't do anything that will cause reduced clearances.
 Is that diode on the bottom doing the same as D9?
[/quote]

What do you mean by "the clearances between live tracks and the pour look way too close"? What is the "pour"?

Yes, that D9, It's pads are gone. I will glue it and add some insulation when I'm done.

[xavier60]

https://en.wikipedia.org/wiki/Copper_pour

[dark_hawk]

https://en.wikipedia.org/wiki/Copper_pour

I see, thank you.

Where exactly do you think there might be problems? I can add an extra layer of uv curable solder mask at these areas.

[xavier60]

https://en.wikipedia.org/wiki/Copper_pour

I see, thank you.

Where exactly do you think there might be problems? I can add an extra layer of uv curable solder mask at these areas.

I can't see with enough detail to be specific. The main thing is that you are aware.

[eblc1388]

A diode is not a fuse. Repeated blowing of a diode can only be due to overcurrent as 1N4007 should not fail in your case due to exceeding reversed voltage, which the diode is rated for 1000V.

The answer is simple, use a diode with higher current rating like 1N5408,3A instead of 1N400x diodes.

As a general rule, increasing the current rating of a diode should not cause problem to any properly designed circuit.

[dark_hawk]

The purpose of D9 is to prevent C61 discharging into the 470uF capacitors when the relay first turns on. This would be audible and bad for the relay also.
Is D9 shorted at the moment?

The positive terminal of C61 was somehow connected to the linear regulators/diode heatsink, which was in turn connected to the output of the 70V diodes through L11 and two 10N capacitors (3rd schematic top right). Removing that connection removed the issue, now it clicks normally the first time.

Thank you for pointing me to the function of D9 and the possibility of C61 being discharged into the 470uF capacitors.

[xavier60]

The purpose of D9 is to prevent C61 discharging into the 470uF capacitors when the relay first turns on. This would be audible and bad for the relay also.
Is D9 shorted at the moment?

The positive terminal of C61 was somehow connected to the linear regulators/diode heatsink, which was in turn connected to the output of the 70V diodes through L11 and two 10N capacitors (3rd schematic top right). Removing that connection removed the issue, now it clicks normally the first time.

Thank you for pointing me to the function of D9 and the possibility of C61 being discharged into the 470uF capacitors.

I would like to make better sense of that, actually sounds potentially dangerous.
Can you describe the connection? Does it seem intentional?

[dark_hawk]

The purpose of D9 is to prevent C61 discharging into the 470uF capacitors when the relay first turns on. This would be audible and bad for the relay also.
Is D9 shorted at the moment?

The positive terminal of C61 was somehow connected to the linear regulators/diode heatsink, which was in turn connected to the output of the 70V diodes through L11 and two 10N capacitors (3rd schematic top right). Removing that connection removed the issue, now it clicks normally the first time.

Thank you for pointing me to the function of D9 and the possibility of C61 being discharged into the 470uF capacitors.

I would like to make better sense of that, actually sounds potentially dangerous.
Can you describe the connection? Does it seem intentional?

No, not intentional.
This board has a serious heat dissipation issue. If you search around you will find 100 pages worth of people trying to fix their units. For example the primary side transistors gets around 70-80 C while in standby doing nothing. Also the whole thing is packed inside the sub woofer box with no ventilation what so ever.

Heat dissipation from the primary side transistors is through a heat sink hanging around, there is also a heat sink for the 4 diodes and the standby linear regulators.
What I did was getting the support heat sink and bolted it to the two heat sinks which in turn bolted to the back of the sub with thermal paste in between all connections. I didn't notice that the positive terminal of C61 was connected somehow to one of the heat sink holes. Might be the conductive/corrosive glue used originally degraded some solder mask and made the connection.

Original configuration:


My changes:




The point where it was connected:

[xavier60]

The top pour in the AUX supply appears to be connected to the cathode of D9, explaining it being live with HVDC.
Off-line power supplies are designed in a way that minimizes the chances of live side circuitry contacting circuitry on the isolated output side, making it dangerously live. 
Having everything sharing the same heatsink structure is not safe design practice.
Im not sure what else to say at this stage except to ask, is there a mains earth connection to the isolated output side?