Is PFC worth it??

[aboubkr90]

PFC has become the standard now, a requirement actually. But no matter how much I tried I just can't wrap my head around that mosfet switching to ground.
I mean, What if we don't. Mains bulge? noise?
We still have to deal with noise and nobody lost 10% of the world's energy dealing with noise!!
Maybe because I come from software where efficiency outweighs functionality sometimes, But I find it really hard to accept that much wasted energy.
I believe that PFC should be restricted to medical and voltage sensitive devices. I personally don't mind the lighting flickering.

[coromonadalix]

https://www.us.lambda.tdk.com/resources/blogs/20080208.html

so   yes

i would say  for critical / sensitive / or precise stuff / long term stability

and getting rid of harmonics is very important too,  telecomm / communications ..

[Wolfram]

From a legal perspective, PFC is required for loads above 100 W or so, due to limits on harmonic current drawn from the mains grid.

From a technical perspective, power factor correction introduces typically 1 - 10 % loss depending on power level and the technology used, while reducing input current between 25 and 50 %. For higher power loads, this puts much less of a burden on the power network, while introducing manageable losses within the product, so I suspect it would be common even in the absence of legal requirements. PFC stages also usually provide the following stage with a fixed voltage, greatly simplifying their design, and enabling efficient LLC converters to be widely used for fixed output voltage applications, while also allowing universal mains voltage input.

There's also the subject of harmonics. Non-PFC loads with a capacitor input filter draw significant low frequency harmonics, leading to voltage distortion in the mains network. The overall level of harmonic currents in the supply network have to be limited in order to not induce gross mains voltage distortion, which can pose challenges for other loads on the grid.

[Phil1977]

My practical experience shows that PFC-power-supplies are 50% more expensive than the cheapest non-PFC-modules but practically as efficient and mostly of much better engineering quality. I´ve seen quite a few non-PFC-modules that really had terrible EMI-emissions (especially LED drivers...) but so far most PFC-designs have been well filtered.

[schmitt trigger]

PFC not worth it? ....this question means that you don't understand the implications of a low power factor in the electrical generation and distribution industry.

For a given amount of generated, transformed and conducted volt-amps, the amount of watts which are the real power that actually performs useful work, decreases with a decreasing power factor.

This fact was immediately recognized when AC power generation and distribution became mainstream. This is the reason that all utilities severely penalize the user with a low power factor load.(*)

For decades, the #1 reason for a low power factor was caused by electric motors. But in the last decades, electronic loads have become ubiquitous, and the traditional rectification and filtering methods create a low power factor situation. Additionally it generates harmonic distortion which itself produces a series of very significant headaches.

Think of the PFC circuit as similar to your car's catalytic converter. Yes, it may rob the engine of some power, but would you rather live with the air pollution?

(*) In many countries, small users i.e. homes, are exempt.

[MrAl]

PFC has become the standard now, a requirement actually. But no matter how much I tried I just can't wrap my head around that mosfet switching to ground.
I mean, What if we don't. Mains bulge? noise?
We still have to deal with noise and nobody lost 10% of the world's energy dealing with noise!!
Maybe because I come from software where efficiency outweighs functionality sometimes, But I find it really hard to accept that much wasted energy.
I believe that PFC should be restricted to medical and voltage sensitive devices. I personally don't mind the lighting flickering.

Hi,

Power factor is a smooth operation if the voltage and current are both sinusoidal.  A low power factor in that case is still not too good  because it draws more current than would be needed to produce the required power.  Correcting the PF would lower that current.
The problem here is that all the lines that bring that power in have to handle more current, and wire with more current dissipates more energy which loads the grid more than necessary, and on top of that, the power company does not get paid for that extra current.  That's with pure AC loads.

With non sinusoidal loads, the situation is even worse.  That's because then we have currents that are not sinusoidal anymore.  I good example I think is with a full wave bridge rectifier and capacitor filter.  The current is almost like a pulse rather than a sinusoid, and that means high distortion as well as high peak current.  Fixing that reduces the distortion as well as allows the capacitor to last longer.

The formula for power factor for pure AC loads is:
PF=cos(theta), where theta is the angle of lead or lag.

The formula for any load is:
PF=Pavg/(Vrms*Irms), where Pavg is the average power and Vrms is the RMS voltage and Irms is the RMS current.

[PGPG]

The main task of PFC is to make load looking like being pure resistive.
If in 3 phase supply system you have the same resistive load at each phase their currents add up in zero wire, but as they have phase shift the sum is 0 (they are shifted by 1/3 of the period). If load in phases are not the same some current flows in zero wire but a maximum (0 load at one phase, or 0 load at two phases) is the same as in phase wires.
If you have bridge+capacitor load than you get high current pulses far different than sinus. Such pulses have a big 3-th harmonic component. These harmonics adds in zero wire as previously but this time they are 'in phase' as shift by 1/3 of the main frequency period means shift for the 1 period of 3-th harmonic.
To avoid fire the zero wire in all distribution should have 2..3 times the cross-section of phase wires or loads having big 3-th harmonic should be avoided. And PFC does this job.

[madires]

From a legal perspective, PFC is required for loads above 100 W or so, due to limits on harmonic current drawn from the mains grid.

IEC 61000-3-2 Ed. 5.1: PSUs >75W and LED drivers >25W

[coppice]

From a legal perspective, PFC is required for loads above 100 W or so, due to limits on harmonic current drawn from the mains grid.

IEC 61000-3-2 Ed. 5.1: PSUs >75W and LED drivers >25W

This is probably the reason 23W and 24W LED luminaires are so very popular.

[Doctorandus_P]

I say this PFC stuff is a whole lot of bollocks that governments do just because they can.

The apparent and imaginary powers are real, but that does not mean you should shove it into the faces of consumers. My vote is that the electricity company has much better opportunities to do this in an efficient way. For example by putting FPC correctors on sub stations. But also, I do not know anything about mains power instabilities and oscillations. I once talked to a guy working in the high voltage distribution stuff, and apparently it was quite difficult to keep the harmonics on the distribution net under control. (Yeah, I know, harmonics are not PFC, but bad power factor (for example triac dimmers) are a big generator of harmonics).

Another idea is to let (industrial) customers pay for apparent power, instead of real power. This gives them a choice to whether they want to do PFC themselves, or let the power company do it for them. This would not work for consumers, mass produced gadgets are cheap ("price sensitive") and no house wife (or guy) is going to pay attention to the power factor of their latest gadget.

Doing PFC on a house-by-house basis instead of for each individual gadget is probably also a much better solution, and cheaper in the long run, but who is going to pay for and install all those PFC boxes in each house.

Edit: Coppice (below) is right, and I formulated my remark about PFC / harmonics differently.

[coppice]

(Yeah, I know, harmonics are not PFC, but if if PFC is 1 there are no harmonics either).

Er, no. If all the harmonics in the current and voltage line up, you have a PF of 1, with a large amount of real harmonic energy. The chances of all of them lining up might be low, but the statement above is wrong.

[ejeffrey]

I say this PFC stuff is a whole lot of bollocks that governments do just because they can.
<snip>
 But also, I do not know anything about mains power instabilities

Ok, see how that sounds dumb?

Another idea is to let (industrial) customers pay for apparent power, instead of real power. This gives them a choice to whether they want to do PFC themselves, or let the power company do it for them.

It's been this way for decades.  All industrial consumers pay for poor power factor. 

but that does not mean you should shove it into the faces of consumers

Nobody is making individual consumers care about power factor correction.  That would indeed be crazy. That's exactly why there are regulations for device manufacturers.

Doing PFC on a house-by-house basis instead of for each individual gadget is probably also a much better solution, and cheaper in the long run, but who is going to pay for and install all those PFC boxes in each house.

It seems unlikely that adding a 50 kW power supply to every house would be cheaper than devices building it in as needed.

[Phil1977]

Is there anyone who is annoyed by the PFC-obligation beside manufacturers of cheap battery chargers?


I can't see that PC-power supplies have got much more expensive. Yes, maybe they would be €5 cheaper without PFC, but does any private PC buyer care about this?

[Wolfram]

From a legal perspective, PFC is required for loads above 100 W or so, due to limits on harmonic current drawn from the mains grid.

IEC 61000-3-2 Ed. 5.1: PSUs >75W and LED drivers >25W

Is that what the standard says? I'm not sure I've read the latest version, but from my memory the limits are on absolute harmonic currents and not on a given power level. Practically, capacitor input filters will fail above a certain power level in that range, but the exact value depends on smoothing capacitor sizing and ESR, so the exact transition point is not clearly defined. But if they changed the way in which this was specified, that's definitely interesting. As an example, induction stoves don't use PFC but also use very little smoothing capacitance across the bus, staying within the IEC-61000-3-2 (or -3-12 for hard-wired units above 16 A) at power levels above 2 kW.

[aboubkr90]

Sorry for the late reply. There are a million ways to get rid of harmonics that doesn't cost 10% efficiency.

[Phil1977]

Why do you assume PFC costs 10% efficiency?

Modern PC power supplies keep around 90% total efficiency over a wide range of loads. Do you think all 10% of losses only account on the PFC-circuit?

[aboubkr90]

Fixed voltage is important for voltage sensitive device. I like it as well. But not on every device. I mean, a 4kw split-ac doesn't need it enough to lose 40-400w on it. And that's my point.

[aboubkr90]

That's implication dependant. The extra e-waste I'm ok with. It falls into our hands and we play with it.

[aboubkr90]

I don't agree with the correlation between power and power factor.
Pollution kills people. CMI doesn't kill circuits. It reduces their stability. what I'm asking, Is stability as important to every device? Do we need to lose energy for the stability of a battery charger for example? It's dead stable, and it doesn't mind the noise.

[Phil1977]

Ah AC inverters...

Let´s try to see it on a wider scale: The losses of all house cabling and also of the grid transformers are generated by the apparent and not by the active power that is drawn by the devices.

That means, if you improve the power factor of an AC inverter by the PFC circuit from 0.45 top 0.9 then you reduce the supply losses to their half. It depends on your cabling and grid connection how much losses there are, but it´s easily possible that the overall consumption of a house goes down by using PFC.

That´s no big oil conspiracy that´s just physics.

[aboubkr90]

Ok everyone!
Your are explaining why we PFC. I know why we PFC.
I'm saying there are other ways that doesn't have much impact on the efficiency of our devices. Like ferrite beads. The rest of the slow CMI spectrum, if ignored, will become white noise that will average to zero (not exactly, but close).
For critical devices, where a human could die if it goes wrong, there we can PFC. and it will regulate the stable voltage the device needs to operate safely.

[aboubkr90]

If I may add. I also know that mains frequency will double if we do that.
Still okay.

[Andy Chee]

Like ferrite beads.

No, ferrite beads do not affect power factor.

I believe you are confusing power factor with EMI.  They are two very different things.

[Siwastaja]

Simple explanation to beginners:

Poor power factor means pulling large amounts of power during short moments. It can (but not necessarily) also mean of pushing some power back into grid (alternatively "borrowing" and "giving back" for no reason).

Someone has to generate (in a loose way, e.g. time-shift by bidirectional storage) and transfer this power at those exact right moments your crappy contraption decides to want it. It is not that this could not be done, but more that it makes more sense to just design products in a way that such special snowflake power delivery for them is not needed.

Analogy: try asking your employer to pay whole year of salary upfront and see how they like the idea. Of try asking them to pay for two year's salary in advance and promise to pay one year worth back after a while.

[PGPG]

I'm saying there are other ways that doesn't have much impact on the efficiency of our devices. Like ferrite beads.

What you propose is like have hope that car shock absorber (absorbs ms range pulses) will ensure your car be traveling strictly horizontally when you drive through hills (hour range pulses).
Ferrite beads work at 100MHz..GHz while PFC eliminates 150Hz, 250Hz ,.. from 50Hz (the ratio between these frequencies is 1000000 times).