TOP258 SMPS IC keeps blowing. (Bose L1 compact speaker system)

[Yamin]

Hey Folks,
I've recently got a Bose L1 compact sound system in for repair. The unit failed to power up one fine day.
After opening the unit up the first thing I noticed was the TOP258 IC was cracked open. Before replacing the IC I had a look for any other components which might have failed. I checked for the obvious culprits and I couldn't find anything which has failed. I checked the diodes and resistors around the IC, checked the main filter caps and the electrolytics near the IC. Checked the secondary side for any shorts or failed caps and nothing  . I checked the capacitors with the ESR meter, I removed some capacitors and checked them with my multimeter for the capacitance value as well.
So after much contemplation, I decided to change the IC only but the thing was I couldn't get hold of any TOP258 ICs. So I placed an order for those. However I had TOP256 IC in my parts bin. It can handle low power than the 258 and to see if anything else failed I replaced it with it (until I got the 258). Powered it up and it was working. Obviously I didn't run it at full volume or had it on for long.
After week of intermittently testing again the unit stopped powering up. Opened it up the SMPS IC was cracked open. Same story as before nothing else seems to have failed. I now have got the TOP258 IC but am afraid the same thing would happen unless I could find the culprit.
I have checked the speakers  thinking they might have partly shorted and might have been drawing too much current. They also test out ok. The amplifier IC is TDA8920B. On the datasheet the amp chip and the TOP258 seems to have all the protection in the world and seems to be bullet proof.
I was hoping someone could shed some light on how to proceed and what else to look for.
I have attached a pdf of the power supply schematic.
For a while I was pretty sure that the TVS diode (DZ1) might have failed as that would put high voltage on the drain pin, but it shows forward voltage of 0.6 and doesn't show anything in the reverse direction. I'm not sure how I can check it  for proper 'TVS' operation.
The full schematic can be found https://elektrotanya.com/bose_l1_compact_system_service_manual.pdf/download.html

Thanks in advance.
P.S: I have got another unit which came in for repairs and the capacitor and the resistor of the LPF was burned up and bulged on that one. The SMPS IC was cracked open on that one as well. I also checked those components on the unit I'm currently working on and they test out ok.

 

[wraper]

For a while I was pretty sure that the TVS diode (DZ1) might have failed as that would put high voltage on the drain pin, but it shows forward voltage of 0.6 and doesn't show anything in the reverse direction. I'm not sure how I can check it  for proper 'TVS' operation.

The same way as Zener diode. You just need a high voltage because it has a high clamping voltage. Primary suspect should be here:

[Yamin]

For a while I was pretty sure that the TVS diode (DZ1) might have failed as that would put high voltage on the drain pin, but it shows forward voltage of 0.6 and doesn't show anything in the reverse direction. I'm not sure how I can check it  for proper 'TVS' operation.

The same way as Zener diode. You just need a high voltage because it has a high clamping voltage. Primary suspect should be here:

Thanks @wraper. So I would need a 150V supply right?
I have checked almost all the components that you have circled. The resistors were ok. C7 measured out of circuit was 2nF which I assume is ok. D2 tests out ok and I measured the resistance of it reversed bias and shows no value.
I was wondering to check the TVS diode would it be a good idea to use mix/max function of my fluke meter during power up to see whether its clamping at that voltage?
Appreciate the help.

[wraper]

You can use rectified mains voltage + resistor in series with high enough resistance so TVS does not overheat.

[Yamin]

Thanks, what do you reckon the best value for a resistor would be for current limiting. Should I try limit the current to 1mA? and should I have a load resistor connected or could I leave it?
I was wondering would my initial idea of min/max function work?
Thanks

[wraper]

1mA is fine. What load resistor? TVS is the load.

[Yamin]

I was thinking like this one.

[wraper]

It's not a test circuit. And in that case TVS is not event intended to conduct any current during normal operation. Only clamping transients.

[Yamin]

It's not a test circuit. And in that case TVS is not event intended to conduct any current during normal operation. Only clamping transients.

Understood thank you

[Yamin]

Update:
I tried measuring the value of C42 but my meter does not read in the pico range. I'm not sure if there's any other way of checking that. Visually it looks ok no cracks or burned marks.

Update on the test for the TVS diode:
My test setup was isolation transformer -- variac -- rectifier --test circuit.
Important note - there were no smoothing capacitors connected across the rectifier output so the DC was not smoothed.
I had two 20k resistors in parallel connected to the TVS diode.
TVS diode specs according to the datasheet
Reverse Stand off Voltage (VR): 128.00V
Breakdown Voltage VBR @1mA: Min 143.00V , Max 158.00V.
Here are the results:
When I set the input voltage to 154V = I measure around 120V across the diode.
When I set the input voltage to 250V = I measure around 130V across the diode.

What would your thoughts be on the results?

[wraper]

Is the rectified voltage smoothed with capacitor? Otherwise it will be pulsating and measured DC value will be wrong.

[Yamin]

Is the rectified voltage smoothed with capacitor? Otherwise it will be pulsating and measured DC value will be wrong.

Nope it was not, I also thought about this. I was using the rectifier circuit of a handheld mixer it didn't have a smoothing cap. What value of capacitor should I use for this test purpose.
I'm now realising that I'm not 100% sure of the TVS parameters either. What is the significance of Reverse Stand off Voltage?
Should I measure the voltage across and take note of the reading when the input voltage = breakdown voltage?
Appreciate the help.

[wraper]

Is the rectified voltage smoothed with capacitor? Otherwise it will be pulsating and measured DC value will be wrong.

Nope it was not, I also thought about this. I was using the rectifier circuit of a handheld mixer it didn't have a smoothing cap. What value of capacitor should I use for this test purpose.

A few uF is enough.

I'm now realising that I'm not 100% sure of the TVS parameters either. What is the significance of Reverse Stand off Voltage?

That it meets the spec and thus is not faulty

Should I measure the voltage across and take note of the reading when the input voltage = breakdown voltage?

Input voltage should be higher, so TVS acts as a Zener diode and limits the voltage.

[Yamin]

Thanks @wraper, so lets say if the input voltage is around 160V the voltage across the diode should be 150V?
I guess my trouble is with a zener diode specific zener voltage is stated for example 5.6V so I would expect to see around 5.6V across the diode if the input voltage is above the zener voltage.
So with this TVS (P6KE150A) diode should I see 128V or the Breakdown voltage?
What should I observe if the voltage is below the breakdown voltage or should it be below the reverse stand off voltage? It should be the open circuit voltage or exactly the input voltage right?
Thanks so much

[wraper]

If current is 1mA you should get between

Min 143.00V , Max 158.00V.

If current is lower, you may see a slightly lower voltage. Or a bit higher voltage if current is higher.

[Yamin]

Thanks
So I'm thinking of changing all the resistors and the caps on the HV side. Since the resistors also looks a bit tarnished even though it measures ok. Also I'm ordering in some diodes.
Anyways I'm having some doubts about the capacitor. C42 is a 47pF 1kV Ceramic disc cap and the dielectric material stated on the service manual was X7R. I couldn't find the same from mouser.
I found same rating capacitors but different dielectric material. So my question is matching the dielectric material that important? and can I use MLCC? any merits to it?I'm asking for future reference too.

[wraper]

Original Part you have probably is MLCC to begin with. But it does not matter is it's single layer or multilayer. Just checked and Mouser has 47pF 1kV X7R. But you can also use C0G and U2J, the difference is they have more stable capacitance.

[Yamin]

But you can also use C0G and U2J, the difference is they have more stable capacitance.

X7R is not available. C0G and U2J is more stable than X7R right?

[wraper]

But you can also use C0G and U2J, the difference is they have more stable capacitance.

X7R is not available.

https://www.mouser.com/ProductDetail/KEMET/C317C470KDR5TA?qs=nfiUXiPFtERNdjkk6ztTQg%3D%3D

C0G and U2J is more stable than X7R right?

Yes, they are better.

[Yamin]

Got stuck on a problem. The service manual says "*ALLRESISTORSINOHMS,5%1/10W(0603)."

I searched for those on mouser and the one which I found which matches the specs had a voltage rating of 75V. For instance R1 (110k) which is the clamping circuit for the mosfet - wouldn't it exceed the rated voltage?
Its my understanding the typical voltage rating for 0603 size resistors are in the 75V range.
I'm slightly confused at the moment.

[wraper]

Is it really 0603? There are also resistors which are rated for higher voltage than usual for the size. Not that you need to replace it. https://www.yageo.com/upload/media/product/products/datasheet/rchip/PYu-RV_51_RoHS_L_10.pdf
350V continuous, 500V peak.

[Yamin]

Further development:
Turns out the resistors I needed were actually 1206 size it was mentioned in the parts list. I'm still waiting for the parts including the TOP258.

Anyway I have a got a theory of why the units failed from a recent experience. I had a similar kind of system (JBL EON one) come in for repair with similar kind of problem. The power supply had failed. However this time the customer did notice some subtle distortion coming from the high frequency driver when it failed. Repaired the power supply powered it back on and there was no sound from the HF tower. Measured the resistance of the HF tower no obvious fault, there are total of 6 drivers - 2 in series and the pair is paralleled with each other so total of 3 rings.
So I opened up the tower disconnected all the drivers and checked each driver with a tone generator and one of the driver was quieter than the rest and it was distorting. So I bypassed that driver which also bypasses the one connected to it in series (all the drivers shows same dc resistance). The unit started to work. (the customer is going to replace the complete tower).
Thinking that the same thing might have happened to the bose L1 units I opened one of the tower. The drivers were connected in the same manner total of 6 drivers but the wires were wrapped to the terminals. I am a bit unsure on unwrapping the wires since I don't have a wrapping tool. So I used the tone generator with all the 6 drivers connected. I thought I heard some distortion.
Could this cause the failure of the TOP IC? As I read the TDA8920B (amplifier chip) sort of shut down for 100ms if it detects impedance is below the threshold as a protection. So
Could I unwrap the wire from the terminals to check the drivers individually and re-wrap them and solder them?   
I also have a some question about how I am testing the drivers - I'm using a small headphone amplifier to drive it, could the distortion I hear be due to the amplifier not being able to handle the load?
Any suggestions and thoughts are greatly appreciated