Help troubleshooting Seasonic X-650 (ATX power supply)

[Otto W]

Hello fellow EEVBloggers! Long time fan of the channel -- first time here.

I'm repairing a failed 650W Seasonic X-650 active PFC power supply. I have not yet been able to trace down the problem, so I'll ask here for some general advice.

Symptoms:
- When I try to power on the PSU, the fan spins up, relay clicks and power is immediately cut off. To me this looks like something trips the overcurrent/protection relay. The clicking relay is the one right next to the mains input in the top right corner of the attached image.
- 5V standby power works normally.
- One of the diodes next to the big transformer in the middle had its other leg "flapping in the breeze". It didn't look burnt or blown, and it tested fine. I simply soldered it back, and you can see traces of fresh soldering around it.

Troubleshooting I've done thus far:
- Electrolytic caps tested and all seem fine. It's supposed to be a decent quality PSU, and hence all caps are Nippon Chemicon. I tested them with my Fluke 189 capacitance test mode.
- All TO-220 transistors and Schottky diode bolted to the heatsink measures OK.
- Bridge rectifier measures OK.
- Attached to this post are single-shot captures of rail voltages when the PSU tries to power on. To me this looks like 3.3V and 5V rails are OK, but 12V rail fails, trips the protection relay and tapers away. The PSU design looks like 3.3V and 5V are tapped from the 12V rail, which I think is quite common in modern PSU's.

This is how far I've gotten, and I'm running out of ideas on what to look for next. Any advice is greatly appreciated. Thanks!

3.3V:


5V:


12V:



What the PSU looks like:

[Whales]

I don't think the relay is a form of protection.  IIRC they are only added on expensive units to bypass parts of the input stages (NTC? or possibly shuts off some input filtering during standby?) and improve unit efficiency by a small %.  It will click on an off as a result of a fault, not necessarily as a cause.

Indeed sounds like self-protection shutoff.  Could be over-voltage or under-voltage  too, there are varying amounts of trip conditions depending on the supervision design.

Are you familiar with the approximate topology of ATX PSUs?  400VDC rail followed by a chopper,  one or more transformers, rectifiers on the other side?  Group regulated vs secondary regulators?  If not then ask, there are also various resources explaining the approximate topology.

Approximate ideas I would try to hunt for clues (I'm not an expert, what I suggest here might not be the best):

• Start with the first power rails.  Probe the 400V rail/heatsink (if you can do so safely).  See if anything weird happens there during turn on.  If you do this wrong then you'll damage yourself.
• With the unit off: gently try backfeeding some voltage into the output rails (use a current limited power supply and keep it very low, eg 100mA max!)  see if they all behave similarly or if one behaves very differently to the others (eg one shorted others fine).
• Start it up with 12V connected to a heavy load.  See if that keeps it online longer than a split second.

I presume that because this unit is active PFC it won't have a 220/110V switch, but it's worth asking anyway

[TheMG]

The relay is only there to squeeze out that tiny little bit of efficiency that would be lost if the inrush limiting NTC were to remain in-circuit at all times. It simply energizes and bypasses the NTC once the power supply is started up.

When the power supply shuts down, so does the relay, as keeping the relay energized with the power supply in standby would waste more energy.

Even if this relay were to fail, nothing would happen other than a barely noticeable loss of efficiency, as the full current of the power supply input is capable of passing through the NTC without the relay.

So the relay clicking on and off is just a symptom and likely nothing else, the fault is elsewhere.

[PKTKS]

(.)
This is how far I've gotten, and I'm running out of ideas on what to look for next. Any advice is greatly appreciated. Thanks!

One more PFC thing just stopped... 
Not surprised.

The lack of ideas really calls for a proper schematic.
hard to figure that crammed stuff just by using our eyes

Too many "new supervisors" and designs made that almost
an impossible "by guess" thing.

They have on the factory the required stuff like schematic
and daughter boards and proper replacement parts.

PFC designs are suited to satisfy legislation not to be user
friendly - RMA not by chance has been made by replacement.

Expensive solution - need to consider if your time worth it

I have been there.

Just to reach the conclusion that: they do not worth the hassle.

Previous designs are easier to fix and maintain.
More reliable predictable stuff

2 cents IMHO before you go to deep without the proper
schema and parts.. yes parts are also a nightmare to find..

Paul

[asis]

Hi
You just need to download the description on the CM6901C.
https://www.alldatasheet.com/view_datasheet.jsp?Searchword=CM6901
There are circuit diagrams on 12.13 pages. STANDBAY circuit.
ECAP requires a thorough check of capacitors C34, C38, C28, C29, C31, C32, C35 and ZD3.
Capacitor indexing may differ from your PSU. If you have a laboratory source of + 15V DC, you can supply voltage from 11V to 12.6 V via a terminating resistor to the + 12V STB terminal when AC PSU is completely disconnected from the mains.
Check the operation of the optocoupler PC3, PC2 (817), conclusions 3.4.
Note ZD3 has a current limit, do not exceed the voltage applied strongly from a laboratory source.
I think the problem is in the STANDBY power source U101 (VIPER201).
Vladimir

[f14]

check voltage of 3.3v 5v rails output by external 12v source . I got some case have dead APM7159 buck pwm controller