capacitor droppers and power factor

During a discussion with a friend, the subject of using capacitors as mains droppers came up. It seems to be common practice these days to use this technique in cheap wall warts (plug top PSU's) and LED lamps. In view of the probably millions of these in use we wondered how this was affecting the power factor of the mains supply and whether the energy companies are concerned about this. I have attempted to search the internet and have established that EU regs state that LED lamps between 5 and 25 watts can have a power factor as low as 0.5.

Comments from our electrical guru's would be welcome.

Years ago I remember seeing an article about how the proliferation of PCs in offices was causing issues with power factors due to all the plugged in power supplies.

Martin C

The issue that I have come accross in the Lab is Earth leakage. Virtually every device now has mains filters and with UK assymetric supplies all contribute to earth leakage. With the old Lab much of the earth current was finding it's way into the buildig metalwork, pipes, reinforcing etc and for field sensitive devices such as electron microscopes causing beam wobble and loss of resolution. Important when you are attempting to work to atomic resolution. I once measured 6 and a bit amps on a water pipe running close to a microscope suite. We reduced it by changing the grounding with a couple of heavy earth straps but the ultimate fix was building an active magnetic shield. The new Lab has much less problems with this as the power distribution was especially engineered in view of these issues.

regards Martin

Just one guru's thoughts...other gurus may also chip in...

If by a dropper you mean a circuit where most of the mains voltage appears across a series capacitor rather than a resistor, this clearly won't meet any sensible power factor spec. Since the load current has to flow through the capacitor as well, a moment's thought shows that for smallish output voltages (~5V say) the power factor is approximately the ratio of output voltage to mains votage, or ~0.02, nowhere near 0.5 minimum. In addition, say for example you wanted 5V at 1 amp - 5 watts output. The capacitor would need to be ~14 uF, capable of withstanding mains voltage continuously. That's physically quite large - think motor run capacitor, certainly not something that would fit in an LED bulb.

I somehow doubt that any cheap mains devices will use capacitor droppers. All the ones I have seen have very small and simple switch mode PSUs.

Capacitors as voltage droppers work great at 50Hz, but are a terrible idea if there are harmonics or spikes.

I had a fancy Philips coffee machine on my canal boat, well that was till I switched from shore power to inverter. A bang and cloud of acrid smoke reminded me that I had left it plugged in.

The inverter was a quasi sine wave, the higher harmonics of the square steps in voltage went through the voltage dropping capacitor and fried the electronics.

It was still under guarantee so I took it back, a few weeks later I got a call from Philips asking how it happened. I was so impressed they were following it up I told them how it happened.

Adrian

I suspect that in that application the dropper was used only for a low power auxiliary supply, for example for a clock. As the required current would be very low the PF would be similar to a suppressor capacitor. But of course, if the mains waveform had a lot of HF then the dropping effect of the capacitor would be much reduced resulting in magic smoke.

A few years ago the forum discussed a very cheap LED lamp's unsuitability for workshop use, and many felt it was dangerous in the home. A small lamp with a bendy neck and magnetic base, potentially useful on a headstock for lighting the cutting area. Inside a capacitor dropper power supply like this example from Wikipedia.

The example confirms John's point about capacitor size because the power output from a cheap capacitive dropper is tiny - this example, 10mA at 5V.

Apart from the cheap lamp, I've not come across another capacitive dropper. All the transformer-less wall-warts I've disembowelled were switch-mode types. I think capacitor droppers are only attractive for a narrow range of cheap low current devices. Above a certain current they're expensive and there's little reason not to use a better circuit.

I'm not convinced the power factor of devices like this have any significant effect on the network. The major cause of power factor bother is motors. Megawatts of inductive load on the grid and a factory might concentrate hundreds of them in a small area. Compared with motors, LED lamps are insignificant. Capacitive loads might even be a good thing. Although their power factor is poor, capacitive phase shift is in the opposite direction to inductive phase shift, and tends to restore balance. Hence racks full of big capacitors are the usual way to apply power factor correction.

Motors are the evil one - I can't think of a common example of a major capacitive load requiring inductive PFC?

Digging into the design of larger switch-mode supplies, say more than 50W, it seems these have become ever more likely to include PFC over the last 25 years or so.

Power Factor and harmonics make my head spin faster than the average electron!

Dave

Is a squirrel cage motor running off a VFD capacitative or inductive?

Must be inductive I think. Copper coil stator, and - although the rotor 'coils' are shorted out bars, they're still one turn coils.

Bit of web searching came up with motor start capacitors, buried cables and PFC capacitor banks. None of them are major loads compared with resistive and inductive loads of which there are many common examples.

Possibly a wind-turbine charging a battery? Dunno.

Dave

Edited By SillyOldDuffer on 17/08/2020 13:36:33

Capacitor dropper power supplies can in fact be found in small appliances. For example, we have a hot air fryer, sold in the UK as a Tefal Actifry. It decided after the warranty ran out to shutdown regularly with an error code. Being me, I took it apart to investigate and found it has a small controller board powered by a capacitor dropper 1/2 wave power supply (on the same board), It powers the mcu (an 8bit ST model) display etc direct from the mains with no isolation. A bit of research indicates these types of supplies are more common than one would think, because they are very cheap and in the case of the Actifry saves putting in optoisolators to sense the AC zero crossing point detection and for the triac motor control.

Be very careful with these supplies on the bench. I do my test work with the board connected via an isolation transformer as otherwise they are dangerous (even then working carefully).

If anyone is interested in knowing more about the capacitor dropper design the following Applications Note from Microchip may be of interest, which is quite often referenced by other sources on the internet..

https://www.microchip.com/wwwAppNotes/AppNotes.aspx?appnote=en021083

As for power factor I have not investigated as I don't yet have a meter for that.

On the point about wall warts raised earlier, the limited number I have taken apart have either had a transformer, the older type, or have a genuine switch mode power supply, both types giving genuine mains isolation.

Paul

An inductive load isn't a load, strictly speaking, as it doesn't dissipate any power. Many years ago when I was designing VFDs for use as electric vehicle drives the company I worked for had the idea of using the VFD as a power factor corrector for industry. The mains connected to the VFD motor terminals and the battery was replaced by a capacitor bank. The software then controlled the capacitor bank to look like a varying capacitative load, cancelling out the normal inductive load of the factory and ensuring the power factor was unity. The unit didn't consume much power although the circulating currents in the capacitors could be large - hundreds of amps. Personally I though this was a better bet for making money than electric vehicles, although in the event neither worked out and the company went phut soon after I moved on.

An induction motor will be inductive on the output of a VFD, but it's isolated from the mains via the DC link. So what the VFD looks like from the mains is determined by the input of the VFD, not the load.

Andrew

No,
Anything on a cheap VFD will be capcitive and with poor conduction angle to boot. VFDs turn AC into DC then back into AC. the inductive effects of the motor don't cross the DC link. Better ones have power factor correction either active or a AC reactor (choke).

Robert G8RPI.

Very true, it's so necessary to be careful with words. How about a "load with a reactive component due to inductance"?

Isn't the failure of a reactive load to dissipate power why such a fuss is made about Power Factor? Domestic customers don't give a hoot about PF because we're only charged for real power, a feature of the meter. Distributors have to care about PF because their networks suffers I²R loses due to current no one is paying for.

Dave

Not sure losses are the main problem. At the generator the output voltage and current will be in phase. So the generator creates an amount of real energy X. The old style electricity meters (with the little rotating disc) only measure real energy consumed, ie, they take into account the power factor. So if the customer has a power factor less than unity he only pays for the real energy consumed, Y. The generator is paying to create X amount of energy but only charging for Y amount, where Y<X. Equals a hole in the bottom line. Modern electronic meters measure power factor as well as voltage and current. What is done with it is up to the electricity company. As SoD says I expect that for the ordinary consumer it makes little difference. Unless one has a garage full of machine tools with large motors.

Andrew

All this seems to confirm that the only sensible application is for very low power auxiliary circuits, which may be controlling much higher power things like heaters. At the 50 mW level mentioned above, the load and power factor is virtually equivalent to an ordinary suppressor capacitor billions of which appear across the mains anyway. I do have a tentative plan to link the Estop circuit on my lathe CNC controller to the mains direct to the motor from the contactor, using a suppressor cap driving an opto isolator through a bridge rectifier. There are some nice little optos that include a bridge rectifier driving the LED. So as long as the supply is on the CNC will run but there's a single Estop button which will stop everything.

Low current yes. I recently looked inside a defunct central heating controller. I was basically one large chip, LCD display and a small relay providing a set of change over contacts to the outside world. The whole thing was powered from the mains via a capacitor/resistor potential divider, rectifier and possibly a zener diode. I didn't investigate further as it was dead, so it went for electrical goods recycling.

The main problem has usually been, as stated above, inductive loads such as washing machines, vacuum cleaners and men with industrial machine tools tucked away in their sheds. There is also (according to what I have read) a problem with distortion of the sine wave caused by millions of switched mode power supplies, where the rectifiers all star to conduct towards the peak of the cycle as they top up the reservoir capacitors. I don't know how true this is, as there have always been rectifier circuits used on the mains supply. Maybe not so many.

Two small points.

1. At a generator's terminals the output is in MVA. It can consist of MW and MVAr's. The MW output is determined by the input to the generator of energy, from a fuel source. More energy in, equals more energy out, as the energy input to the generator attempts to increase the generator's rotational speed above the system frequency. The MVAr component is a function of the excitation of the generator, and the transformer ratio between the generator and the system. The result cn be a generator only making MW, or only MVAr's - a synchronous generator.

2. Switched mode power supplies (includes wind turbines) tend to generate odd order harmonics (3rd. and 5th. are the biggest. 7th., 9th. etc are less. Further reading: Fourier Series) back into the power system, unless they are filtered out at their source - wind turbines are an example of this. This causes, as others have said, heating losses in the network, but more importantly in the transformers in the network. High neutral currents also, as others have said. All this shortens the life of expensive assets - like transformers.

If all this doesn't make sense, don't worry! The outcome is that you need to pay more for your electricity!

Was a problem, rather than is a problem!

Here's my go at understanding what's going on.

First, a rectifier circuit that doesn't cause PFC or harmonic problems:

This is OK for Power Factor because the diodes rectify throughout the AC cycle, and the un-smoothed DC output is fine for electric railways, heating, lead acid battery chargers, electroplating and other unfussy applications. Good Power factor, rotten DC because the circuit outputs 100Hz half cycles obnoxious to amplifiers, radios and many other requirements.

The 100Hz half cycle DC can be improved considerably simply by adding a big capacitor. It stores electricity during peaks and releases it during valleys, putting cleaner DC on the load by filling in the gaps.

Unfortunately adding Capacitor C has a profound effect on the rectifier diodes. They no longer conduct throughout the AC cycle, instead they're called on to charge the capacitor mostly at the top of each AC cycle. This causes the transformer to draw current from the mains in far from sinusoidal pulses causing harmonic distortion.

Power Factor, in the sense of a phase shift caused by an inductive motor doesn't cover this situation so the governing standards express acceptable distortion in terms of harmonic content at various frequencies. Seems that basic capacitor smoothed power supplies meet these standards up to about 60W, beyond that power supplies need some form of Power Factor Correction.

The traditional approach is to add a choke: Choke.

The choke resists the rush of current into the capacitor, and reduces it's wham-bam thank you ma'am effect on the rectifier and transformer. Simple solution, but chokes with enough inductance to work at 50Hz are big, heavy, contain miles of copper wire and are thus expensive.

The modern approach (over the last 30 years) has been to replace massive chokes with some form of active power factor control.

PFC can be improved with a suitable arrangement of rectifiers, capacitors, and resistors. The circuit manages the rate at which C is charged, buffering the rectifier from the brutal current requirements of the simple circuit. However, more likely than this simplicity, a switch mode boost converter circuit is used. It draws near sinusoidal current from the rectifier, generates high-frequency AC from crude DC, and much smaller inductors and capacitors can exploited to filter fast switching pulses into clean DC. Rather than being a simple filter, the convertor will be a fully regulated circuit. The main disadvantage is the interference caused by high speed switching.

I don't think it's necessary to worry about PFC in domestic products.

Dave

Edited By SillyOldDuffer on 18/08/2020 12:18:36

Would this be one of these beasties? It's from my bandsaw, appears to have gone short circuit and to be delivering 300v to the LED which made that understandably unhappy.

4 legged chip has 810 on it and I think is a bridge recitifier, there is an obvious electrolytic but I haven't attempted to decipher the rest of the circuit yet.

Edited By Adrian R2 on 12/01/2021 09:31:54

The brown tubular thing looks like a power resistor to me.