Reversed polarity cap killed PSU board

[Pyr0Beast]

Re: Reversed polarity cap killed PSU board


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.
I recap PSU's before failing. If I leave the 5V cap before the choke as it was and recap rest everything is ok.

Replace that cap with some ULESR and it whines. Replace with some poorer cap, whine is gone.

And no, it doesn't respond much to load.

Tweaking them for lowpass IS tuning.
But PI filters ain't even lowpass. They are DC pass-trough.

It will usually be more, not less.
That depends on the suitability of the cap. High ESR can mean higher losses in cap, not necessarily better filtering.

Also. Unloaded stuff isn't worth anything.

Same with temperature.

And same with feedback.

Low ESL cap can act differently as well.

Even PWM regulation is different then.

Oklahomawolf [from jonnyguru] embarrassed himself publicly by making that mistake.
Recapped unit had worse ripple than it did with bad caps.
Can you link me to that thread ?

[PCBONEZ]

Re: Reversed polarity cap killed PSU board

It will usually be more, not less.
That depends on the suitability of the cap. High ESR can mean higher losses in cap.

NONSENSE
- Assuming a Ripple Voltage of 100 mV show me how a higher ESR [more resistance] will result in more heat being created [aka:"losses" aka:watts] in the cap.
.
I'll look at the rest later.
.

[Pyr0Beast]

Re: Reversed polarity cap killed PSU board

NONSENSE
- Assuming a Ripple Voltage of 100 mV show me how a higher ESR [more resistance] will result in more heat being created [aka:"losses" aka:watts] in the cap.
.
I'll look at the rest later.
.

Assumption of Ripple voltage for power losses in capacitor is wrong. Ripple voltage is result of ripple current and resistance of a cap. If cap cannot handle current - ripple voltage will go up and so will resistive losses.
Current always stays the same.
Now tell me. Where will be bigger losses:
Perfect cap with 1Ohm series
Perfect cap with 2Ohm series

[PCBONEZ]

Re: Reversed polarity cap killed PSU board

Assumption of Ripple voltage for power losses in capacitor is wrong. Ripple voltage is result of ripple current and resistance of a cap. If cap cannot handle current - ripple voltage will go up and so will resistive losses.
Current always stays the same.
Now tell me. Where will be bigger losses:
Perfect cap with 1Ohm series
Perfect cap with 2Ohm series

WRONG
Capacitors don't -make- Ripple.
The switchers do.
.
And Voltage comes before Current.
Without Voltage there is no Current...
.
That's why Voltage is called Potential.
It's the 'Potential' to have 'Flow'.
.

[PCBONEZ]

Re: Reversed polarity cap killed PSU board

Oklahomawolf [from jonnyguru] embarrassed himself publicly by making that mistake.
Recapped unit had worse ripple than it did with bad caps.
Can you link me to that thread ?

They revamped their site and my link doesn't work.
I think 370forlife has the new link because as I recall he's posted it since the new site version.

[370forlife]

Re: Reversed polarity cap killed PSU board

http://www.jonnyguru.com/modules.php...tory&ndar_id=8

[Pyr0Beast]

Re: Reversed polarity cap killed PSU board

WRONG
Capacitors don't -make- Ripple.
The switchers do.
.
And Voltage comes before Current.
Without Voltage there is no Current...
.
That's why Voltage is called Potential.
It's the 'Potential' to have 'Flow'.
.

There is no voltage difference in supper-conductor yet the current flows.

[PCBONEZ]

Re: Reversed polarity cap killed PSU board

There is no voltage difference in supper-conductor yet the current flows.

That's not true either but that's OT.
.

[PCBONEZ]

Re: Reversed polarity cap killed PSU board

Piece by piece [Now that I have time.]

Assumption of Ripple voltage for power losses in capacitor is wrong.

No it isn't.
- You obviously don't understand what the Ripple Current rating is.
- The Ripple Current rating is based on not overheating the cap internally due to heat generation [aka losses] inside the cap.
Ripple Voltage and ESR determine the Ripple Current.
Since the ESR is fixed:
-- Ripple Voltage determines Ripple Current
-- Ripple Current and ESR determine losses. [The losses are in Watts]

Ripple voltage is result of ripple current and resistance of a cap.

No.
Caps don't create Voltage.
The Ripple Voltage is voltage spikes created by the switching transistors switching their lil butts off.
The Ripple Voltage would be there even if you had no caps.

If cap cannot handle current - ripple voltage will go up and so will resistive losses.

No.
Capacitors do not -actively- -regulate- either current or voltage.
If the Ripple Current exceeds the rating of the cap then the cap will simply overheat.
[Just as a resistor does when too many amps pass through it.]

Current always stays the same.

Yes it does.
So does ESR.
That is why the voltage doesn't change.
But if that current that isn't changing exceeds the rating of the cap then the cap will overheat.

Now tell me. Where will be bigger losses:
Perfect cap with 1Ohm series
Perfect cap with 2Ohm series

That's what I asked you and you couldn't tell me.

So on with what you couldn't figure out.

Starting with some switchers generating 100 mV and no cap.
[Of course the standard 100kHz is assumed.]

Add the 10Ω ESR cap across the 100mV
E/R = I
0.1v/10Ω = 0.01a
-
I²R = W
0.01a x 0.01a x 10Ω = 0.001w

.

Now replace the 10Ω with the 20Ω.
E/R = I
0.1v/20Ω = 0.005a
-
I²R = W
0.005a x 0.005a x 20Ω = 0.0005w

Lookie there...
Higher ESR results in ~LESS~ losses in the cap.
- Just like I told you.

.

[Pyr0Beast]

Re: Reversed polarity cap killed PSU board

That's not true either but that's OT.
.

The electrical resistivity of a metallic conductor decreases gradually as the temperature is lowered. However, in ordinary conductors such as copper and silver, this decrease is limited by impurities and other defects. Even near absolute zero, a real sample of copper shows some resistance. Despite these imperfections, in a superconductor the resistance drops abruptly to zero when the material is cooled below its critical temperature. An electric current flowing in a loop of superconducting wire can persist indefinitely with no power source.

http://en.wikipedia.org/wiki/Superconductivity

No resistance = No voltage potential

[PCBONEZ]

Re: Reversed polarity cap killed PSU board

The electrical resistivity of a metallic conductor decreases gradually as the temperature is lowered. However, in ordinary conductors such as copper and silver, this decrease is limited by impurities and other defects. Even near absolute zero, a real sample of copper shows some resistance. Despite these imperfections, in a superconductor the resistance drops abruptly to zero when the material is cooled below its critical temperature. An electric current flowing in a loop of superconducting wire can persist indefinitely with no power source.

http://en.wikipedia.org/wiki/Superconductivity

No resistance = No voltage potential

No voltage potential = No electrical flow.

"can persist indefinitely with no power source"
"can persist"
~ But it can not -start- without a voltage potential.
Once you remove the voltage the movement isn't electricity anymore, it's just particle motion.
Electrons move around the nucleus of an atom too, that is not an electric current it's just particle movement.
.

Anyone mention to you that I used to split atoms for a living?
.

[Pyr0Beast]

Re: Reversed polarity cap killed PSU board

No voltage potential = No electron flow.

1. So can you say: Voltage potential means electron flow ?
220V in electric socket. Electrons stay put (if we ignore air conductivity)

No it doesnt. Just because there is a potential it doesn't mean anything will flow.


2. Take a shunt resistor and measure voltage across it. Lets go for 1 Ohm and 1Amp = 1 Volt
Go with 0.1Ohm
0.1V across resistor
Go with 0.01Ohm
0.01V across it
Go (or calculate a limit) with 0 Ohm = perfect conductivity
Still 1A across it
Yet no voltage.

Same with electron flow in vacuum. Nothing to cause resistance and there is nothing that causes voltage potential, therefore vacuum is superconductive.

EDIT: Seems post has changed.


~ But it can not -start- without a voltage potential.

Flux Pumping
http://en.wikipedia.org/wiki/Flux_pumping

Law of induction helps


Anyone mention to you that I used to split atoms for a living?
No. Glad you did. Must be fun stuff.

Once you remove the voltage the movement isn't electricity anymore, it's just particle motion.
Electron movement is electricity. Electrons still hold their kinetic energy. If you would be brave enough to brutally open/cut a superconductive loop, you should expect electrons and perhaps even ripped off atoms to spurt out in a form of powerful electric arc.


Electrons move around the nucleus of an atom too, that is not an electric current it's just particle movement.
That is unorganised movement which doesn't do anything (electrons however do have their kinetic energy). Organised one is in which such energy can be released.

[Pyr0Beast]

Re: Reversed polarity cap killed PSU board

No voltage potential = No electrical flow.


Seems you have changed that as well

Electrical flow = flow of electrons (well, in opposite direction due to negative charge)

Speed of moving electrons over resistance = electric potential over resistance
Number of moving electrons over resistance = current passing trough resistance which in turn causes electric potential

You can also say potential causes electron flow
Two sides of the same coin in non superconductive conditions

Also, found this interesting graph:

[Pyr0Beast]

Re: Reversed polarity cap killed PSU board

Piece by piece [Now that I have time.]


No it isn't.
- You obviously don't understand what the Ripple Current rating is.
- The Ripple Current rating is based on not overheating the cap internally due to heat generation [aka losses] inside the cap.
Ripple Voltage and ESR determine the Ripple Current.
Since the ESR is fixed:
-- Ripple Voltage determines Ripple Current
-- Ripple Current and ESR determine losses. [The losses are in Watts]

- You made a false assumption I do not understand specifications of capacitors.
- Ripple current is calculated from max. allowed dissipation of the package and its measured internal resistance.
- ESR is not fixed and it varies greatly with temperature, same as capacitance, but for the sake of simplicity I will leave that alone

- Ripple voltage does not determine Ripple Current, With no capacitors in switching regulator (non-continuous operation mode) ripple voltage = output voltage of transformer from one tap to the other/ground
In buck from input voltage to ground
In boost from induced voltage+input voltage to ground
(Ignore diode drops)

Constant current load would eat all the 'ripple current' which would be equal to the output current of transformer/source at the expense of rising voltage on the load.


To keep it simple.
Capacitors are constant voltage 'load' -> Pushing it with higher voltage will demand amps
Switchers constant power load -> Output needs to be constant power regardless of input voltage of current. If lacking voltage it will demand more current and vice versa.
Inductors constant current 'load' -> Voltage across them will rise when you want to increase the current trough them

To keep voltage from rising you add capacitance which will store all the excess 'current' and 'fix' the voltage to certain point.

Once we have voltage fixed we can experiment with resistive and inductive elements.

Adding series resistance to the capacitor will shift the current handling to the load, in turn, raising the rail ripple voltage.

Adding series inductance before capacitor will increase magnetic flux in coil but keep the current virtually the same. Voltage will rise on switching pulse and when pulse is removed it would be still provided by disintegrating flux trying to maintain current.

Caps don't create Voltage.
Never said they do.

The Ripple Voltage would be there even if you had no caps.

Yup.


Capacitors do not -actively- -regulate- either current or voltage.
Caps and inductors are passive elements. All regulation they do is by storing and releasing energy.

If the Ripple Current exceeds the rating of the cap then the cap will simply overheat.
Yes. But what I meant is, when series resistance goes up ripple voltage will go up as well to compensate for the lack of current flow in and out the cap.

Starting with some switchers generating 100 mV and no cap.
Which is wrong by default.

You can design a switcher with 100mV ripple limit. But that is a design choice based on experience and not a calculated value from esr esl and similar parameters, just take 1% of output voltage of 10V, that is 100mV
But it is good to stay in the 100mV-200mV range anyway.

From there you calculate other parametrs, like desired capacitance.

Higher ESR results in ~LESS~ losses in the cap.
- Just like I told you.

You represented problem completely wrong.

Also explain why poor quality (high ESR) caps have lower specified ripple current if losses ain't greater ?

Now replace the 10Ω with the 20Ω.
Now replace 20Ohm with open circuit and there suddenly is no need for capacitor.

Or replace 10Ohm with 1Ohm ...
P=U²/R
0.01V²/1
0.01W losses

0.1Ohm
0.1W losses

0.01W
1W losses


... then use a perfect capacitor (no losses) and watch it go supercritical.

Those calculation are completely wrong because they assume constant voltage and completely ignore constant current demands.
You can see they are absurd given in the real limits.

What will run as 100mA over 1Ohm esr will run as 100mA over 10Ohm esr.
And what would once be 0.1V ripple voltage will suddenly become 1V
10mW vs 100mW

[PCBONEZ]

Re: Reversed polarity cap killed PSU board

First to debunk you OT gibberish. [Will look at the rest later on.]

As I said before the 'electron flow' in super conductors is not electricity.
The electrons moving in super conductors are comprised of Cooper Particles.
-
Cooper Particles are two electrons bound together and are a simply a charged particle.
A (negative) Beta particle is a single electron and simply a charged particle.
-
Both Beta and Cooper particles can be put into motion by applying a voltage however Cooper Particles in motion are no more 'electricity' than are Beta Particles in motion.
-
The fact that they keep moving when the voltage is removed proves that.

.

Wikipedia [overall] is wrong as often as not.
That is due to the bulk of it's authors being laymen within the topic they write. All you have to do to know this is look at the 'Capacitor Plague' article to know this true. It's been AFU (both Wikipedia and the Cap Plague article) since day one and no amount of correcting it changes that for more than a day.
Wikipedia is a good place to find references but it is not itself a good reference.
Wikipedia is to an Encyclopedia as the National Enquirer is to a News Paper.

[Yeah, I use it too. But only when it makes sense and what it says can be verified from a reliable source.]
.

[Pyr0Beast]

Re: Reversed polarity cap killed PSU board


As I said before the 'electron flow' in super conductors is not electricity.
The electrons moving in super conductors are comprised of Cooper Particles.
And Cooper pairs can consist of electrons.

Open charged superconductor up and tape what happens. Energy has to go somewhere.

In a battery, there can be no electron flow either, just ion flow.

That is due to the bulk of it's authors being laymen within the topic they write.
Most school books aren't any better either.

[Th3_uN1Qu3]

Re: Reversed polarity cap killed PSU board

Mmm, PCBONEZ, so you proved that bad capacitors don't get hot anymore. And this helps us how? In the vast majority of devices with bad caps, the caps didn't cook themselves, it's the components around them that heat them up.

Scope voltage before capacitor. Multiply by 0.3535. Square. Divide by ESR. Compare with datasheet. But if you have a temperature probe there's no need for all this.

[PCBONEZ]

Re: Reversed polarity cap killed PSU board

Mmm, PCBONEZ, so you proved that bad capacitors don't get hot anymore.

Go study. - You need it.
.
And yes, depending on the failure mode that can be a true statement.
.
Perhaps you should study failure modes of caps.
.

[Th3_uN1Qu3]

Re: Reversed polarity cap killed PSU board

Go study. - You need it.

True in some aspects but not on topic.