PC P&C Silencer 420W

[Colt45ws]

Re: PC P&C Silencer 420W

Yes, I thought I had posted something about that incident, but I must have forgotten to do that. I have a 60W temp-controlled unit I bought to be able to do caps, and I was trying to remove the heatsink and related components with it. I could probably do caps but something that heavy was a no go.
As far as the coil overheating, I dont think so, it was dead cold when I fired it up. Ran 20 minutes and *POP*.


I just went and found the bag of components I was able to pull. Neither of the primary switchers are shorted. Just for further info.

[Toasty]

Re: PC P&C Silencer 420W

Heard from Colt45ws via PM...

No. I the process of removing the primary heatsink and hardware, I lifted a couple traces as I dont have a desoldering setup. I still have it but it just spare parts now...

Unfortunate.

[Toasty]

Re: PC P&C Silencer 420W

Hey Colt45ws...

Did this get fixed or figured out?

[seanc]

Re: PC P&C Silencer 420W

https://www.badcaps.net/forum/showpo...4&postcount=24

That does to me... unless of course my Seasonic S12 is not a Seasonic.

[TheLaw]

Re: PC P&C Silencer 420W

This is a re-badged Seasonic isn't it.
Looks identical to the Antec EA-380 and Seasonic S12II 500W I have. If it's 2007 dated, you're about right for the OST caps to fail.

Do you want me to check them out for comparative parts/readings?

Doesn't really look like a Seasonic topology.

[kaboom]

Re: PC P&C Silencer 420W

If you turn down 'boosted B+' too much, you may not be able to sustain the output voltages under load.

If the inductor is overheating because too much current is used to boost 120V AC/170pk to ~400 (copper loss), or if the core is getting too hot (too much flux from too much current), I'd do something about that. After all, the PWM max duty cycle and transformer turns ratio was calculated based on the switch(es) being supplied with ~400V.

If you want to cheat, put a load on it that represents what it'll ultimately power. 10A on 5V, 10-15A on 12V, 8-10A on 3.3V, for example. Now find the voltage adjust point for the PFC booster. While watching the voltage, turn it down until your output voltages start to collapse. Now bring it back up by 20 percent or so. Ten percent may be possible, but it leaves little room for line voltage dips and I2R losses in the PFC choke and TX as they heat up.

If the transformer was wound with a low ratio, you may be able to run the PFC voltage as low as 240V dc. I think this would be unusual though, since at nominal voltage, the duty cycle would have to be quite low. Possibly to where it affects stability.

The best thing would be to recalculate/redesign the PFC- heavier wire on the inductor, different/bigger core, different frequency to suit core and/or range of boosted voltage, etc.

-Paul

[everell]

Re: PC P&C Silencer 420W

What I don't like about these APFC units is that they seem to be designed for 200-240V input. At 120V in, the PFC booster has to supply much more power to get the main bus up to the 360-400V it's regulated at.

-Paul

Paul, I have a AGI HP-W4307F3 400 watt power supply that uses the CM6800 chip for APFC and PWM. Its output voltage is set at 400 volts (403 volts measured). Looking at the CM6800 data sheet, they use 385 volts as the suggested set voltage. But.....here in USA we have 120 volts input. It is not too difficult to locate the feedback resistor which sets the output voltage. I just reset it using a trimpot, now measuring 380 volts. In order to balance the voltage needed for the pwm switching transistors and lowering the voltage to reduce power generated by the booster coil, what would you suggest as an optimum output voltage set point?

[4m0k]

Re: PC P&C Silencer 420W

Yah, that's Seasonic S12-II 500, same internals as in my Corsair VX450.

[seanc]

Re: PC P&C Silencer 420W

Sorry Everell, I meant the OPs unit.
I did read your thread back when you were originally fixing that PSU, I'm amazed that you got it working again!

[hoodlikegaza]

Re: PC P&C Silencer 420W

My 750 watt silencer has recently started to make some electrical sounding noise when under load by my video card, or cpu (reproducible by any 3d app, or prime95). After googling my symptoms, I found this informative site Looks like I'll take a look at the internals tomorrow, but do any of you do cap replacements for a fee? The descriptions of how to diy are a bit intimidating. I'm located in NJ.

Pm me!

[everell]

Re: PC P&C Silencer 420W

I believe it is a Delta. Here is the thread I started to describe the repair work done on it

https://www.badcaps.net/forum/showth...hlight=everell

[seanc]

Re: PC P&C Silencer 420W

This is a re-badged Seasonic isn't it.
Looks identical to the Antec EA-380 and Seasonic S12II 500W I have. If it's 2007 dated, you're about right for the OST caps to fail.

Do you want me to check them out for comparative parts/readings?

[everell]

Re: PC P&C Silencer 420W

...till the wire insulation melts and turns start shorting to each other. Boom.

Like this!

[Th3_uN1Qu3]

Re: PC P&C Silencer 420W

I would also like to thank our Romanian engineer.....that he says the wrong core is being used on the booster coil if it is overheating. That is correct, but if the Chinese manufacturer used a cheaper (cost out) core, we are stuck with a time bomb, which in a few years goes "kaboom".

That would be me. Thanks... although i have about 3 more years till i can be officially called an engineer. Now that you mention it, if the manufacturer used a cheaper core with significant thermal aging, in the long term the core losses increase so the coil gets hotter and hotter... till the wire insulation melts and turns start shorting to each other. Boom.

I fixed one.....after feeding several sets of FET transistors to failure. The coil is hard to troubleshoot as it may measure OK, but when in operation is bad.

It is hard to troubleshoot if you don't have the adequate equipment. I don't either... but i've read about a simple circuit that hooks up to your oscilloscope and measures saturation current of an inductor. Stay tuned.

[everell]

Re: PC P&C Silencer 420W

Looks like this thread is going in the same direction as our previous interesting (and heated) discussion on APFC. I would like to thank kaboom for his take on the matter....the booster coil gets too hot when psu is operated at 110 VAC. It is having to do twice as much work as those psu operated on 230 volts in Europe. I would also like to thank our Romanian engineer.....that he says the wrong core is being used on the booster coil if it is overheating. That is correct, but if the Chinese manufacturer used a cheaper (cost out) core, we are stuck with a time bomb, which in a few years goes "kaboom". My take on it is, the two FET transistors are shorted, the coil is either shorted or intermittently shorting due to overheating, and some surrounding components may also be bad. Changing only the FET transistors will not solve the problem. I fixed one.....after feeding several sets of FET transistors to failure. The coil is hard to troubleshoot as it may measure OK, but when in operation is bad.

And special thanks to pcbones......APFC does not help or save money for the computer owner. AND it does not necessarily help the power company either. For sure, only the politicians and lobbyists are pocketing extra money.

[Th3_uN1Qu3]

Re: PC P&C Silencer 420W

Unfortunately as PCBONEZ said, this doesn't help as much as one would think, because there's still all those crap swirly bulbs with not even an attempt at PFC. Computer power supplies are far from being the worst offenders when it comes to mains waveform distortion.

[kaboom]

Re: PC P&C Silencer 420W

PFC is not about power supply efficiency it's about power grid efficiency.

I realize that. Instead of 5-10 amps just before every positive and negative voltage peak, when instantaneous line voltage exceeds that of the input caps and they recharge, the PFC 'booster' spreads that current out over the entire 360 degree cycle (hopefully). Those old secondary dist ckts, with transformers too far apart, get 'flat topping' of the sine wave from the line conductor resistance.

Power supply _unit_ efficiency goes down since we have a 'power supply feeding a power supply.' The POCOs love it- people think they're 'being efficient' but every one of those PFC boosters has losses. 10 watts here, 15 cents there...

-Paul

[PCBONEZ]

Re: PC P&C Silencer 420W

Even on the grid it only increases efficiency on paper.
Specifically the paper in the accountant's office.
.

[Th3_uN1Qu3]

Re: PC P&C Silencer 420W

The very fact that those coils overheat and the semis need a heatsink proves that they _reduce_ the overall efficiency of the supply...

PFC is not about power supply efficiency it's about power grid efficiency. As for the booster circuit, it's the designer's fault if they aren't spec'd for 120v operation too. Universal input supplies have been around since way before PFC.

[kaboom]

Re: PC P&C Silencer 420W

What I don't like about these APFC units is that they seem to be designed for 200-240V input. At 120V in, the PFC booster has to supply much more power to get the main bus up to the 360-400V it's regulated at.

In my opinion, the booster coils are way too small- not big enough or enough copper. They get way too hot and when they finally get shorted turns, the semis go bang. Or if the DC bus cap opens up, there's no place for the inductor to dump its current. You get insane voltages that destroy everything.

This whole idea was forced upon us by a few 'influential' countries in Europe. Instead of upgrading their backwards secondary-level (416Y/240, 380Y/220V) power distribution systems, the whole world has these little afterthoughts instead. The very fact that those coils overheat and the semis need a heatsink proves that they _reduce_ the overall efficiency of the supply...


One thing to verify is whether the IC on that little board has been damaged. If the PFC FETs turned into a piece of wire, the PFC IC will be backfed line voltage. In that case, the replacement FETs will be destroyed by an IC that puts line voltage to the gates. If you can find the datasheet for your particular PFC IC, see if you can power it up by providing a sort of aux supply for it. Now you can see if the gate drivers are OK.

-Paul