Capacitors before bridge rectifier

[bnetz]

Hey Everyone,

I'm repairing a battery charger and there is 2 capacitors before the bridge rectifier. i'm not sure of the purpose? picture below

[amyk]

Are you sure you've traced the schematic correctly? Are these electrolytics or some other type? If they were 220n instead of 220u, they might be for interference suppression.

[bnetz]

Yes, they are 220uf electrolytic capacitors. picture below (the black and red wire are just ac sensing)

[edpalmer42]

They're for noise suppression.

When you connect two electrolytic capacitors like that, they look like a single non-polarized capacitor with half of the capacity of a single capacitor.  So, in this case, the two together look like a single 110 uf capacitor.  However, that value is rather large.  Combining those two little capacitors and the size of the wire visible in the photo, it's puzzling.

This arrangement is somewhat controversial.  Many people think it's a very bad idea and will eventually fail.  Other people think it's fine.     YMMV.

[bnetz]

Thanks for that info!
one of the capacitors did fail. The 2 aluminum plates are the positive and negative. There's 35a button diodes sandwiched between them.

[MrAl]

Hey Everyone,

I'm repairing a battery charger and there is 2 capacitors before the bridge rectifier. i'm not sure of the purpose? picture below

Hello,

Another possibility is for power factor correction.

The power factor of the bridge is notoriously bad and so some form of power factor correction is sometimes employed.  As others have said though, connecting two capacitors like that is not that great of an idea.  It's a very old idea.  What came after that was connecting two diodes, one across each capacitor, so that when the polarity reverses one of the diodes conducts, and when it reverses again the other diode conducts.  That helps to protect the capacitors from a reverse voltage.  They only see a reverse polarity of about 0.7v then.

To be sure it is for power factor correction you could go through the calculations.
This is of course if you have the circuit drawn correctly.

[floobydust]

I say it's a mistake - these capacitors are the wrong value and type.
Typically an RC snubber network is there. Say 10 ohms + 22-100nF film cap there to lessen spikes from the transformer which would damage the button diodes.
Battery chargers can be made pretty ugly, made in a village literally including using whatever is lying around for parts.

I notice your PC board is badly cooked on the (+) connection, same for the transformer wire it's dark brown too. It must be a high resistance poor connection there that is making lots of heat. This would also damage the caps there.

[bnetz]

Picture below shows before the repair. Had to cut out the old board trace and solder in copper. appears to have been a high resistance connection

[MrAl]

Picture below shows before the repair. Had to cut out the old board trace and solder in copper. appears to have been a high resistance connection

Hello again,

Were you able to verify that the original schematic you drew here was correct with regard to the two electrolytic capacitors?

Would you be able to estimate the year this charger was sold?  That would help determine if it was an attempt to provide power factor correction or not.  A lot of stuff did not have that many many years ago.

[Zero999]

Hey Everyone,

I'm repairing a battery charger and there is 2 capacitors before the bridge rectifier. i'm not sure of the purpose? picture below

Hello,

Another possibility is for power factor correction.

The power factor of the bridge is notoriously bad and so some form of power factor correction is sometimes employed.  As others have said though, connecting two capacitors like that is not that great of an idea.  It's a very old idea.  What came after that was connecting two diodes, one across each capacitor, so that when the polarity reverses one of the diodes conducts, and when it reverses again the other diode conducts.  That helps to protect the capacitors from a reverse voltage.  They only see a reverse polarity of about 0.7v then.

To be sure it is for power factor correction you could go through the calculations.
This is of course if you have the circuit drawn correctly.

But a bridge rectifier an smoothing capacitor is a non-linear load, which is capacitive so adding a large capacitor to the input would  make the power factor worse, not better.

[CaptDon]

Zero, A bridge rectifier with filter bank looks hugely inductive!! Think about it, the voltage will rise until it begins to add current to the filter bank at its existing voltage level. So there may be no current flow until the voltage rises to 80% or more of the existing state of charge. So capacitive power factor correction makes sense. The way they implemented it is a piss poor design and destined to fail as it surely did. A back to back pair of polarized capacitors makes a damn poor non-polarized capacitor, especially for non-symetrical waveforms. Non-polarized capacitors from reputable manufacturers ARE NOT internally a pair of polarized capacitors although there have been examples of some garbage made that way,

[Zero999]

Zero, A bridge rectifier with filter bank looks hugely inductive!! Think about it, the voltage will rise until it begins to add current to the filter bank at its existing voltage level. So there may be no current flow until the voltage rises to 80% or more of the existing state of charge. So capacitive power factor correction makes sense. The way they implemented it is a piss poor design and destined to fail as it surely did. A back to back pair of polarized capacitors makes a damn poor non-polarized capacitor, especially for non-symetrical waveforms. Non-polarized capacitors from reputable manufacturers ARE NOT internally a pair of polarized capacitors although there have been examples of some garbage made that way,

I've thought about this more and no, it's not inductive either.

In reality, a bridge rectifier and smoothing capacitor is neither capacitive, nor inductive. It's just non-linear.

I simulated this to prove it. Notice how adding an inductor or capacitor of different values doesn't reduce the current draw? All you get is the current spikes plus the additional current drawn by the inductor or capacitor.


 Rectifier Bridge C.asc (1.6 kB - downloaded 10 times.)


 Rectifier Bridge L.asc (1.64 kB - downloaded 8 times.)

Conclusion: the capacitor in parallel doesn't do anything to improve the power factor of the bridge rectifier. If it did, then it would be very common, but it isn't.

[CaptDon]

Plug that battery charger into a Voltec power analyzer and look at the displayed readings, power factor, VA, VAR, True power etc. The power factor is horrible when the charger is unloaded but improves under load. The equivalent 110Uf probably really doesn't correct the power factor by much so they may be there to swamp out spikes and protect the diodes. Your sim seems to say 10 milliohms as a source impedance. I would guess the in-situ is closer to .5 ohms.

[johansen]

Folks the diodes have not needed protecting since the days of copper oxide.

If the diodes are making noise in a sensitive radio circuit it was common to put 0.01uf ceramic caps across them, with resistors in series.

Op has a battery charger....!

Personally i have never seen anyone ever use back to back electrolytics to make an ac capacitor except on some experiments with wind turbines. And you dont want resistors on them, because its the non linear leakage resistance of the capacitor that keeps them having a dc bias such that they both develop a positive voltage.

 Tuns out you can get a cubic power curve by putting capacitors in series with the generator. But, they dont last all that long.


The only way to cancel out the harmonics pulled from a power source by a rectifier... Is to create those harmonics.

You can get close, you put a big inductor in series with the power source. And then you build a series of harmonic traps. 3rd, 5th, 7th, 9th... And then you have the bride rectifier.. followed by an inductor to smooth out the current. Then the battery load


The problem is...when you do that you get more of a constant current source rather than a voltage source...which means that now you have to control the current into the battery with an auto transformer, or transformer taps. And harmonic traps are lossy especially at low powers so.. its like at best a 50% efficient way of doing it

[johansen]

Zero, A bridge rectifier with filter bank looks hugely inductive!! Think about it, the voltage will rise until it begins to add current to the filter bank at its existing voltage level. So there may be no current flow until the voltage rises to 80% or more of the existing state of charge. So capacitive power factor correction makes sense. The way they implemented it is a piss poor design and destined to fail as it surely did. A back to back pair of polarized capacitors makes a damn poor non-polarized capacitor, especially for non-symetrical waveforms. Non-polarized capacitors from reputable manufacturers ARE NOT internally a pair of polarized capacitors although there have been examples of some garbage made that way,

I've thought about this more and no, it's not inductive either.

Both simulations are wrong. Replace the load with a zener diode and resistor to hold the voltage down to say, 12 to 14 volts with 1 volt of ripple.

Add a resistance to the ac source such that you lose about 15% power in the resistor. Then sweep the capacitor or inductor. (After the inline resistor)

[MrAl]

Zero, A bridge rectifier with filter bank looks hugely inductive!! Think about it, the voltage will rise until it begins to add current to the filter bank at its existing voltage level. So there may be no current flow until the voltage rises to 80% or more of the existing state of charge. So capacitive power factor correction makes sense. The way they implemented it is a piss poor design and destined to fail as it surely did. A back to back pair of polarized capacitors makes a damn poor non-polarized capacitor, especially for non-symetrical waveforms. Non-polarized capacitors from reputable manufacturers ARE NOT internally a pair of polarized capacitors although there have been examples of some garbage made that way,

I've thought about this more and no, it's not inductive either.

In reality, a bridge rectifier and smoothing capacitor is neither capacitive, nor inductive. It's just non-linear.

I simulated this to prove it. Notice how adding an inductor or capacitor of different values doesn't reduce the current draw? All you get is the current spikes plus the additional current drawn by the inductor or capacitor.

(Attachment Link)
(Attachment Link)

(Attachment Link)
(Attachment Link)

Conclusion: the capacitor in parallel doesn't do anything to improve the power factor of the bridge rectifier. If it did, then it would be very common, but it isn't.

Hi,

Sorry, but I do not see any place in that post where you are showing the power factor.

You've shown waveforms, but not the power factor.
It does not matter if the load is linear, nonlinear, or some combination, or spikes or no spikes, it only matters what the power factor is before and after the power factor adjustment component(s) are added.  If the power factor improves, then it works, if it does not improve, then it does not work, and some other method must be used (if there actually is going to be any power factor correction for the circuit that is).
Note also that I am not sure if we can assume what the load is just yet.

This does not mean that your conclusion will not be right, it just means that you have not yet proved that because we don't see anything about power factor we just see waves.

There also remains some question about why they used such a poor method for connecting the capacitors together regardless of what they are supposed to be doing to improve the circuit.

[johansen]

This does not mean that your conclusion will not be right, it just means that you have not yet proved that because we don't see anything about power factor we just see waves.

The term power factor is used interchangeably with both phase angle and harmonic content.

You cannot generate the 3rd 5th 7th 9th 11th 13th harmonics drawn by a diode rectifier feeding a non zero voltage load....by a capacitor or an inductor connected across the supply.

All that capacitor does is shift the zero crossing of the first harmonic, which is often called power factor. The bridge rectifier does usually draw a little leading power factor, because the current starts on the rising edge of the sine wave minus the diode drop minus the capacitor filter voltage, then the capacitor voltage rises and the current falls on the trailing edge of the sine wave faster than it rose.

An infinite inductor after the rectifier draws a perfectly square wave of current..which is neither leading nor lagging and it is a pproximately 50% power factor because it draws only the 3rd 5th 7th 9th etc out to infinity, all of which add up to half of the area in a sine wave.

[coppercone2]

maybe replace it with this and forget about it
https://www.digikey.com/en/products/detail/nichicon/UES1V101MPM1TD/4317422

I think trying to plug in some old battery charger into power factor correction equipment to determine the need for a part that now costs 1 dollar and its more reliable then the ones they put is not worth your time. I think its a wild goose chase.

It should do whtever the old ones were doing but also should not have any problems from misunderstandings related to making dual polarity electrolytic capacitors from single polarity capacitors.

[Haenk]

maybe replace it with this and forget about it
https://www.digikey.com/en/products/detail/nichicon/UES1V101MPM1TD/4317422

Wasting a MUSE cap on this? Uh...

[Zero999]

Zero, A bridge rectifier with filter bank looks hugely inductive!! Think about it, the voltage will rise until it begins to add current to the filter bank at its existing voltage level. So there may be no current flow until the voltage rises to 80% or more of the existing state of charge. So capacitive power factor correction makes sense. The way they implemented it is a piss poor design and destined to fail as it surely did. A back to back pair of polarized capacitors makes a damn poor non-polarized capacitor, especially for non-symetrical waveforms. Non-polarized capacitors from reputable manufacturers ARE NOT internally a pair of polarized capacitors although there have been examples of some garbage made that way,

I've thought about this more and no, it's not inductive either.

Both simulations are wrong. Replace the load with a zener diode and resistor to hold the voltage down to say, 12 to 14 volts with 1 volt of ripple.

Why?

We don't know what the load on the bridge rectifier is.

My suggestion my be wrong, but it's likely yours is too.

A constant load is correct, if it's a linear regulator, which is a constant current load.

A zener + resistor is correct, if it's a shunt regulator.

Add a resistance to the ac source such that you lose about 15% power in the resistor. Then sweep the capacitor or inductor. (After the inline resistor)

Fair point. 10m is probably much lower than the internal resistance of a small transformer. Increasing the series resistance, will reduce the current spikes and improve the power factor, at the expense of increased power loss.

Zero, A bridge rectifier with filter bank looks hugely inductive!! Think about it, the voltage will rise until it begins to add current to the filter bank at its existing voltage level. So there may be no current flow until the voltage rises to 80% or more of the existing state of charge. So capacitive power factor correction makes sense. The way they implemented it is a piss poor design and destined to fail as it surely did. A back to back pair of polarized capacitors makes a damn poor non-polarized capacitor, especially for non-symetrical waveforms. Non-polarized capacitors from reputable manufacturers ARE NOT internally a pair of polarized capacitors although there have been examples of some garbage made that way,

I've thought about this more and no, it's not inductive either.

In reality, a bridge rectifier and smoothing capacitor is neither capacitive, nor inductive. It's just non-linear.

I simulated this to prove it. Notice how adding an inductor or capacitor of different values doesn't reduce the current draw? All you get is the current spikes plus the additional current drawn by the inductor or capacitor.

(Attachment Link)
(Attachment Link)

(Attachment Link)
(Attachment Link)

Conclusion: the capacitor in parallel doesn't do anything to improve the power factor of the bridge rectifier. If it did, then it would be very common, but it isn't.

Hi,

Sorry, but I do not see any place in that post where you are showing the power factor.

You've shown waveforms, but not the power factor.
It does not matter if the load is linear, nonlinear, or some combination, or spikes or no spikes, it only matters what the power factor is before and after the power factor adjustment component(s) are added.  If the power factor improves, then it works, if it does not improve, then it does not work, and some other method must be used (if there actually is going to be any power factor correction for the circuit that is).
Note also that I am not sure if we can assume what the load is just yet.

This does not mean that your conclusion will not be right, it just means that you have not yet proved that because we don't see anything about power factor we just see waves.

There also remains some question about why they used such a poor method for connecting the capacitors together regardless of what they are supposed to be doing to improve the circuit.

It should be fairly obvious that the power factor is not improved by addition of L or C, by looking at the waves. If it was, then the current consumption would reduce, but that's clearly not what's happening.

For conclusive proof, the power can be calculated from the waveforms, which LTSpice can do, but I've not got time at the moment.

[Kleinstein]

The distortion from the rectifier plus filter are mainly harmonic and the added capacitor should not help much, if at all. To really reduce this one would want inductance or added resistance (e.g. from a fuse).  The details can still vary as the magnetizing current also contains harmonics, especially with transformers driven relatively close to saturation.

Some capacitance at the scondary side of the transformer can still help, as part of the magnetizing current can than be provided from the secondary side. This is the inductive load part of the transformer. The added capacitor can improve the power factor and even reduces the losses at the transformer a little (likely not that relevant). Another point is reducing sharp slopes in the voltage waveform (especially when the mains waveform is no longer a sine) that can cause current peaks from reverse recovery (that is the part that may occasionally cause EMI). Normal rectfiers are relatively slow, similar to 1N4007.

It is still not good to used the 2 electrolytic capacitors in series. With no really good, cheap bipolar capacitor the question is if it is really worth it. With a relatively high voltage (like 50 V) one could get away with a not so large film capacitors as the needed capacitance scales like 1/U². For the EMI point aready a relatively small capacitor like 100 nF should be sufficient.

[MrAl]

This does not mean that your conclusion will not be right, it just means that you have not yet proved that because we don't see anything about power factor we just see waves.

The term power factor is used interchangeably with both phase angle and harmonic content.

You cannot generate the 3rd 5th 7th 9th 11th 13th harmonics drawn by a diode rectifier feeding a non zero voltage load....by a capacitor or an inductor connected across the supply.

All that capacitor does is shift the zero crossing of the first harmonic, which is often called power factor. The bridge rectifier does usually draw a little leading power factor, because the current starts on the rising edge of the sine wave minus the diode drop minus the capacitor filter voltage, then the capacitor voltage rises and the current falls on the trailing edge of the sine wave faster than it rose.

An infinite inductor after the rectifier draws a perfectly square wave of current..which is neither leading nor lagging and it is a pproximately 50% power factor because it draws only the 3rd 5th 7th 9th etc out to infinity, all of which add up to half of the area in a sine wave.

Hi,

You do realize that power factor correction is often used with rectifier inputs for power supplies?

I do not think there is anything here to talk about yet.  Power factor is power factor, it's not a bunch of waves that you interpret by sight.
You either calculate power factor or you do not calculate power factor.  If the calculation comes out showing the power factor was improved, then it was improved.  If it comes out that it was not improved, then it was not improved.
So far I have not seen the power factor calculation therefore none of this is certain yet.
To recap, previously I noted that any spikes or various peaks does not rule out a power factor improvement.  You can't look at a wave (unless it is very simple) and assume that the power factor is either good or bad.  You can only do that with sine waves or possibly other well known waves that have known power factors already.

Do a power factor calculation then we will have something to talk about.  Without that, it's just random opinion that may or may not be correct.

[MrAl]

The distortion from the rectifier plus filter are mainly harmonic and the added capacitor should not help much, if at all. To really reduce this one would want inductance or added resistance (e.g. from a fuse).  The details can still vary as the magnetizing current also contains harmonics, especially with transformers driven relatively close to saturation.

Some capacitance at the scondary side of the transformer can still help, as part of the magnetizing current can than be provided from the secondary side. This is the inductive load part of the transformer. The added capacitor can improve the power factor and even reduces the losses at the transformer a little (likely not that relevant). Another point is reducing sharp slopes in the voltage waveform (especially when the mains waveform is no longer a sine) that can cause current peaks from reverse recovery (that is the part that may occasionally cause EMI). Normal rectfiers are relatively slow, similar to 1N4007.

It is still not good to used the 2 electrolytic capacitors in series. With no really good, cheap bipolar capacitor the question is if it is really worth it. With a relatively high voltage (like 50 V) one could get away with a not so large film capacitors as the needed capacitance scales like 1/U². For the EMI point aready a relatively small capacitor like 100 nF should be sufficient.

I think it's the inductor loads that can benefit from a capacitor to improve the power factor, but I'd have to look up some notes I don't do this every day anymore.  I am not going to guess this one though because the waveforms can be very deceiving.  Also, we have not been given the load yet I don't think.  The load would be part of the calculation so we'd need that.

The ultimate answer would come from a calculation of the power factor why bother with anything else.  If the load is not given then this may require calculating with various types of simple load types.  If we fond that one load type has an improved power factor, then we know it works with that kind of load and possibly not with other types of loads.  This might be good even if we were given the load as it would help understand power factor correction a little bit better.

[Fryguy]

The way they implemented it is a piss poor design and destined to fail as it surely did.

Piss poor design - maybe - destined to fail as it surely did ? This one surely failed because it was cooked by the red glowing screw next to it . . . !   

Otherwise these caps would still be doing their job right now - whatever their job may be . . .  The designer of this charger placed these two caps there for a reason and i'm curious why he did it .

It seems the reason behind it has yet to be found - good hunting ! 

BTW - maybe someone can simply pick up some old transformer , a few diodes and caps out of the trash bin , hook it up to a scope and find out what it does in the real world . . .  just an idea .

[floobydust]

I've seen bottom-dollar (linear) chargers and power supplies with nonsensical electronics. Especially from Korea or the bowels of china.
You'll see there's no knowledge, only outright cheap-as-possible manufacturing that keeps making things silly- but it works.

As if power factor matters at all to them. Ripple current spec for those caps is too tiny to help out. And it's after the transformer, so pushing energy back to the primary is extra inefficient I would say hopeless. You would need a reasonable transformer Spice model. I measure on a similar, smaller 10A 200mH, 16mH. Leakage inductance I did not check but it's going to be big enough for trouble.

Folks the diodes have not needed protecting since the days of copper oxide. [...]

Copper oxide and selenium rectifiers behave differently than silicon diodes.
Here, button diodes start at 50PIV. You think they went higher? Spent more pennies? Didn't use low rung parts? (I find button diodes usually fail due to bad connections and heat transfer unless they are spot-welded). But in avalanche?

This transformer also feeds the controller PC board and you can see they shrunk the capacitors to the smallest that will still work. I think the mystery caps are covering for some problems they had, likely transients.

[coppercone2]

maybe replace it with this and forget about it
https://www.digikey.com/en/products/detail/nichicon/UES1V101MPM1TD/4317422

Wasting a MUSE cap on this? Uh...

its a $1.21 part , very reasonable for a capacitor in something that you don't want to break possibly a $200 battery