Reversed polarity cap killed PSU board

[PCBONEZ]

Re: Reversed polarity cap killed PSU board

Not sure about that.

I am.
Take a good PSU with PI filters and measure the ripple on an o'scope.
Then change [only] one of the OP caps to one with much lower ESR and measure the ripple again.
It will usually be more, not less.
.
Oklahomawolf [from jonnyguru] embarrassed himself publicly by making that mistake.
Recapped unit had worse ripple than it did with bad caps.
.

[Th3_uN1Qu3]

Re: Reversed polarity cap killed PSU board

Lacquer only works if the signing winding is in the outer wraps and that isn't always the case.

Not if you take the transformer out and dip it for a couple hours.

[PCBONEZ]

Re: Reversed polarity cap killed PSU board

Ripple filters in most PSUs are lowpass not bandpass, so no they aren't tuned.

Tweaking them for lowpass IS tuning.

I've heard people blaming caps with too low ESR for supply whine, but in my experience it was always a case of loose windings in a transformer or coil and a bit of lacquer always the fix. Also if it didn't work when you got it, you have no way of knowing if the noise wasn't there in the first place.

First version of Antec Smart power was famous for supply whine and it was found to be a poorly chosen capacitor.

Lacquer only works if the signing winding is in the outer wraps and that isn't always the case.

[Th3_uN1Qu3]

Re: Reversed polarity cap killed PSU board

Ripple filters in most PSUs are lowpass not bandpass, so no they aren't tuned. Besides, the variations in ESR that occur naturally, would throw off tuning anyway.

I've heard people blaming caps with too low ESR for supply whine, but in my experience it was always a case of loose windings in a transformer or coil and a bit of lacquer always the fix. Also if it didn't work when you got it, you have no way of knowing if the noise wasn't there in the first place.

[Pyr0Beast]

Re: Reversed polarity cap killed PSU board

Output filters are 'tuned' to optimize filtering at a certain 'expected' noise frequency and the value of the inductors, the caps uF, and the caps ESR are all take into account when the designer specs the caps.

If you go haphazardly changing values around you will most likely de-tune the filters and the result is more ripple noise in the output.

.

Not sure about that. I do think it has to deal with loop compensation instability.
Lower ESR helps greatly with reducing ripple.

[PCBONEZ]

Re: Reversed polarity cap killed PSU board

Ask Mr. 25 Watt soldering iron (aka: mockingbird) to explain...

And thus the bird of repetition learns two lessons.

1: Size does matter.

2: It is better to walk softly and carry a big stick than to walk loudly and carry a little one.

.

[PCBONEZ]

Re: Reversed polarity cap killed PSU board

I don't know why, but some PSU's don't like very low esr and/or high capacitance on output.

Output filters are 'tuned' to optimize filtering at a certain 'expected' noise frequency and the value of the inductors, the caps uF, and the caps ESR are all take into account when the designer specs the caps.

If you go haphazardly changing values around you will most likely de-tune the filters and the result is more ripple noise in the output.

.

[Pyr0Beast]

Re: Reversed polarity cap killed PSU board

Ya lost me there...

Extending the leads -increases- the ESR. The Rubycon was Higher than the Panasonic, and Lower than the OEM Capxon.

I don't know why, but some PSU's don't like very low esr and/or high capacitance on output.

But again, some don't start if ESR is too high ...

[romansroad]

Re: Reversed polarity cap killed PSU board

I thought I had read he was under-wattaged.....the plot thickens! Perhaps we should have a "help mockingbird get a better soldering iron fund" set up?

[Toasty]

Re: Reversed polarity cap killed PSU board

Ahhh, but you don't know the whole story.

Ask Mr. 25 Watt soldering iron (aka: mockingbird) to explain...

[romansroad]

Re: Reversed polarity cap killed PSU board

One trick I've used is to cut off a shorter piece of the braid, lay it on the PCB where you want to suck up the solder, and then apply the iron. The braid can be quite the heatsink itself, so shorter pieces sometimes are better. You can also just slightly tin the braid with rosin flux lead based solder (I have a 1lb spool of Multicore), then apply lead based solder to the part you want to de-solder....yeah, I know, seems counterintuitive, but the heat conduction is better. Then, apply the braid and suck up the non-lead with the lead solder. BTW, lead based melts at a lower temp, and makes it all easier.

[mockingbird]

Re: Reversed polarity cap killed PSU board

Did you get Solder-Wick yet?

Yes, I bought some "MG Chemicals" braid, and eventhough it shows in the catalog as fluxed, it performs rather poorly even after I dip it in some flux. It's not what I thought it would be. I thought it was gonna suck up solder as soon as I applied heat to it but you really have to put it an effort to get it to absorb something.

[Toasty]

Re: Reversed polarity cap killed PSU board

>>You told me to remove it!!!<<

Yeah, but not expecting a tinker-toy soldering iron either...

Did you get Solder-Wick yet?

>>Check capacitance of Rubycon cap, not ESR.<<
As stated, Rubycon is fine per spec meaning ESR & capacitance.

>>how do you test the dual anode diode for failure<<
Cathode is common center leg. Use diode check with black on center and red to both outside. A failure would either be open or shorted, one or both sides. Can be difficult to check because of connected components. Unsoldering the center leg on a single-sided board is usually the way to go. Through hole board with via's is not easy and will require removal. Experience is the teacher here. Knowing that a reading is good/bad comes with time.

>>can *DEAD* fuses still show continuity if they're tested in circuit?<<
Not typically.

Toast

[mockingbird]

Re: Reversed polarity cap killed PSU board

Very nice. The diode pack all I did was run a tiny piece of wire between it and the nearest compatible trace. You told me to remove it!!!

Check capacitance of Rubycon cap, not ESR. Cap can die with same ESR but has lost its capacitance. I suspect the 10V 2200uF cap I put in the wrong way killed it.

This is good to know for the future. One more question if I may, how do you test the dual anode diode for failure? Second question is can *DEAD* fuses still show continuity if they're tested in circuit?

[Toasty]

Re: Reversed polarity cap killed PSU board

Ya lost me there...

Extending the leads -increases- the ESR. The Rubycon was Higher than the Panasonic, and Lower than the OEM Capxon.

[Pyr0Beast]

Re: Reversed polarity cap killed PSU board

>>I do not know (nor understand) why the Rubycon would not allow it to run >>as the original Capxon is higher ESR by the datasheet.

Chip detects 'short' on the output and shuts down.
I extend the leads if that happens.

[Toasty]

Re: Reversed polarity cap killed Acer AL2017 PSU board

Acer AL2017 Mfg. 12/2006

PSU Board: Delta DAC-12M033 AF, Rev: 02A

Panel: Chi Mei - A201P1


>>Can you show me the fuse that was blown?<<
https://www.badcaps.net/forum/showpo...20&postcount=6

The other pics as promised attached.

Fuse replacement with Pico fuse. Cleaned mask from trace and drilled holes to accept said fuse type, shown on right.

Repair of damaged diode pack traces. Removed and remounted diode pack with fresh thermal grease. Cleaned off mask from traces. Drilled holes oversize and inserted wires I pre-bent at 90° to fit alongside diode leads. Fluxed and soldered in place and over-flowed wire back to good trace areas.

Was not getting any output from main transformer as measured at diode packs. Laid scope probe across transformer and found only a low frequency sine wave. Not correct. With nothing else apparently damaged, and the missing signal at the transformer, I suspected the small cap.

Even though the cap was a Rubycon YXG and seemed in good condition and in spec on capacitance and ESR. The specs for all caps used there follow. I do not know (nor understand) why the Rubycon would not allow it to run as the original Capxon is higher ESR by the datasheet.

[B]22uF @ 50v[/B]
   Original: Capxon GL - 135mA - 1.00Ω
Replacement 1: Rubycon YXG - 180mA - 0.70Ω => Actual 0.46Ω
Replacement 2: Panasonic FM - 250mA - 0.34Ω => Actual 0.23Ω

I wonder if there is another problem that I should investigate, but after it ran for 10 days with no issues on either test patterns, browsing, or videos, I am not going looking for trouble.

Toast

[mockingbird]

Re: Reversed polarity cap killed PSU board

Congrats!

I tested the startup cap for continuity, so I didn't suspect it. Can you show me the fuse that was blown?

[Toasty]

Re: Reversed polarity cap killed PSU board - FIXED!

Done. Working! Running 4 days continuously.

Short version:
1 - 2200uF/10v FM - replaced missing one (One that was reversed)
1 - 4A pico fuse - replaced blown SMD one on 5v line (W crosses to 4A in 2 different manuf sheets)
1 - 22uF/50v FM - replaced Rubycon YXG in start up section

More details and some other photos will be up later.

Toast

[Toasty]

Re: Reversed polarity cap killed PSU board

Okay. PM sent.