Since my job is related to led drivers and ballasts i get confused with some questions. I see that the capacitor that is placed after diode bridges decreases the power factor and when we remove it we can achieve a higher pf. So i think that our power factor is a leading pf.

How do you compensate leading power factor?

Also, why we know most of the loads as inductance loads? and why do we just have capacitor banks, not inductors bank?


3 Answers 3


The capacitor in a power supply following the bridge rectifier is not causing a phase shift in the current. Instead, it is contributing to harmonic distortion of the current waveform, which is a different method of creating a poor power factor.

When you remove the capacitor, the conduction through the rectifier is more continuous. When you add the capacitor, the conduction now occurs only in short pulses at the peak of the voltage waveform.

Putting an inductor in parallel with such a load would solve nothing — you would now have a lagging current component PLUS a harmonic distortion component.

However, putting an inductor in series with such a load can have benefits. It will filter out the higher-frequency components of the harmonic distortion by forcing the rectifier diodes to switch on and off more slowly. Note that it will still cause a lag in the current relative to the voltage, but you can probably find a "sweet spot" value for the inductance that balances the two effects to create the best overall power factor.


Traditionally, large electrical power consumers (factories) 'spend' most of their power consumption driving large inductive loads (like motors).
Inductive loads cause a lagging power factor (since the current in the circuit 'lags' behind the voltage).
In order to compensate for these inductive loads, these factories can install power-factor correction equipment which, in its simplest form, is a bank of capacitors connected up to the power lines in parallel with their existing inductive loads with the size of the bank of capacitors calculated to approximately offset 'lag' caused by the inductive loads.

As you've noticed, LED drivers (and most other electronic devices running from the mains) have large-ish smoothing capacitors in them and yes you are correct - this does cause a leading power factor (opposite to inductive loads).

So - why don't we need inductor banks to compensate for these?
A couple of reasons:

  • Lighting is a much smaller consumer of power than most other industrial processes, and so the leading power factor that it could introduce is easily swamped by the much larger lagging power factor cause by all of the inductive loads
  • Most of the even half-way decent LED drivers on the market include power-factor-correction circuitry into their design (I know all of the drivers made by the company I work for do). You'll see that the mains input supply in these drivers doesn't simply run from a bridge-rectifier to a large capacitor - there is a much more complicated system there to rectify & smooth the supply. This automatically 'corrects' the effective power-factor of the device to 0.9 or better and so any external correction is usually unnecessary.
  • \$\begingroup\$ Thanks alot.as you said most of the led drivers have pfc correction;but some which are less power dont.at least in my country.about the first reason that you mentioned,i realized that as an example in my house all of electrical devices powers will be added together at first and since the collected load would be inductive load so we need a capacitor for compensating.am i correct? \$\endgroup\$
    – s.rahmati
    Feb 11, 2016 at 14:50
  • \$\begingroup\$ The less power they use, the less of an effect they'll have on the overall power factor. A 10kVA load with a power factor of 0.8 is much, much worse than a 10VA load with a power factor of 0.3 ... Since the PFC circuitry adds complexity & cost, we only add it where it will be useful. \$\endgroup\$
    – brhans
    Feb 11, 2016 at 14:51
  • \$\begingroup\$ Yes, for an installation you would typically look at the overall power-factor of everything added together and compensate that. Houses though are usually exempt from having to do PFC since (from the electricity provider's point of view) residential loads are typically not 'bad' enough to worry about. \$\endgroup\$
    – brhans
    Feb 11, 2016 at 14:53
  • \$\begingroup\$ Your second paragraph is incorrect. The capacitor after the rectifier causes harmonic distortion, not a leading current waveform. See my answer for a fuller explanation. \$\endgroup\$
    – Dave Tweed
    Feb 11, 2016 at 15:09
  • \$\begingroup\$ Dave is right - I was over-simplifying. That's why I work in networked controls and not power electronics ... :P \$\endgroup\$
    – brhans
    Feb 11, 2016 at 15:17

I guess that it is cheaper to buy high power factor led drivers than try to fix lots of low power factor drivers, but if you are in a huge trouble (already bought, installed and got complaints), let us think. I'm not thinking about dimmable drivers.

Solution #1: about 23 years ago I had problems with low power factor and harmonics when designing a 2x32W 220V electronic ballast for fluorescent lamps. Since the load could be considered almost constant, the solution was to add a passive LC elliptical filter to the input tunned in the 4th harmonics (240Hz in a 60Hz grid) in order to attenuate strongly the 3rd and 5th order harmonics; 7th, 9th and above not so much.

Cons: needs you to find the calculation, not useful for small quantities (many of them with their inputs tied together would be better to lower the amount of filters), the DC voltage over the electrolytic capacitor inside the driver will drop about 20%, need someone who makes the inductor. Inductor if not built properly will generate "hummmm noise".

Pros: the PF will rise close to 0.96 and harmonics will be lowered.

Solution #2 - Will work only if the components of the driver circuit (electrolytic capacitor, mosfet, integrated circuit and other) can handle 400VDC (if your mains is about 220VAC). If your mains is about 120VAC, then you can work with output voltages close to 300VDC, but always watch out for the other internal components of the driver.

1)Tie the inputs of the AC drivers of a room together (in two or more sets); 2)Know the input power of them together; 3)Build a circuit (or ask somebody to do it for you) using a high power factor integrated circuit (plus other components: mosfet, inductor, etc.). L6562, FAN7527B are some of the good examples of ICs and have some suggested circuits in their datasheets. 4)The output of the circuit (power factor with DC voltage output) will be connected to the inputs of the led drivers. Don't worry where to connect the positive or negative: there is a rectifier bridge inside your drivers that can handle this problem.

CONS: find somebody to do it for you if you won't, distribute DC to the drivers, the wall switch must turn on/off the PFC circuit (not the DC outit worksput of it!), when making maintenance DC is more dangerous than AC at the same voltage value.

PROS: power factor close to 0.98 and extremely low harmonics. This circuit makes no "hummmm" noise.

Solution #3: Valley-fill circuit. Search the web. It is not miraculous but sometimes it works.

CONS (that I remember from 15 years ago): Power factor close to 0.92, current harmonics about 25 to 30%. Weird DC output voltage. Necessary to know if your driver will handle the DC waveshape without stroboscopic effect or overheat, must change the input circuit inside the driver (oh my...).

PROS: low cost.

Best regards

  • \$\begingroup\$ Hi again i realized that when the load is linear we have lag or leading power factor and because most of the loads are lagging we use capacitor banks in distrbution systems.but my question is that when our load is nonlinear(such as led drivers)and the driver doesnt have pfc so in this case how the distrbution system compensate this low pf and how should the system produse the source current needed for the driver? \$\endgroup\$
    – s.rahmati
    Feb 13, 2016 at 21:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.