EEVblog #848 - Home Energy Savers BUSTED!

[Circlotron]

^^^ Yep, typical smps that has just a bridge rectifier feeding a big electrolytic cap at the input has a power factor of about 0.5 to 0.6 and just putting a cap across the AC input will not help at all.

[Circlotron]

Ah, classic PG&E verbiage. Why provide an equation when a few hundred words will do.

Because lawyers can't argue against equations.

[gildasd]

Where I work we have capacitor banks around in the manufacturing building. They stand a good 1/3 taller than an average man.  Some store 14.5KV, some are 13.2KV and some are 12.5KV. Complete with arc flash hazard signs and all the fun stuff like that.

Sometimes in cramped areas you may have to stand close or lean on them while working on other things.

Here's an example of a similar looking bank. That's right, you can buy industrial capacitor banks on Ali Baba. Ours are a little more "Schmick" made by Chris Gammell's company, ABB instead.

Most boats have something like that.
It's one of the few parts that the mechs have no right to stick a wrench in

[SeanB]

Just a note that airconditioner compressors, aside from the inverter types, which will likely have active PFC on the input just to keep the main bridge current down from the charging of 8000uF of ultra low ESR capacitors, do not have PFC capacitors at all. The 40-75uF capacitor in there is to provide phase shift on the permanent split capacitor motor, which will typically have a PF of around 0.89 lagging in most instances.  You can add an external PFC capacitor, but this will both add cost and mean you need a much more robust contactor to handle the capacitor inrush, and the start up current pulse will then be high enough to trip breakers. As it is inrush current is in the order of 60A for a few cycles as the compressor starts, pretty big for most loads but typical in air conditioning.

Fridges as well you can add external PFC, and there putting it before the control will work, as then it will mostly be PFC for the rest of the house, and the inrush there is generally lower at under 20A in most cases.

[nwvlab]

We had problems in one lab because there were too many PCs with cheap PSUs inside (these 400W PSU are almost empty, compared to "regular" one), and they were triggering the local breaker 

That occurred because the cos(phi) was so low, that the current was incredibly high, even if the real power was around 1kW, maybe 2kW or so. Unfortunately, the technician solved the problem by changing the local breaker with one of higher ratings (instead of purchasing decent PSUs, with good PFC).

That said, the "home energy saver" would have even worsened the effect! (As it appears in the video)

Still

Cheers!

[NiHaoMike]

Some Dell rack servers I have worked on have a selection in the BIOS to disable PFC. Apparently, even in some commercial environments, PFC is not worth the efficiency loss and/or the UPS does the PFC function.

And then there are tricks that can be played with digital PFC controllers in order to boost efficiency and still have a pretty good power factor.

[mjkuwp]

I made a meter based on ADE7953 that can show the power factor and plot the waveforms

http://themzlab.tumblr.com/wattmeterwaveforms

My screenshots are a little bit out of date in that they are using 'inductive' or 'capacative' labels for the loads and even listing power factor as negative on one screenshot.  That comes from me being inexperienced at the time and having found some old terminology.  My meter no longer displays those terms.

I don't like the 'leading' and 'lagging' descriptions of power factor and especially when they describe the waveforms as sine waves.  having taken lots of snapshots with these meters, I hardly ever see a true sine wave.  It is more correct to just describe power factor as a ratio between the active power (integration of V*A) divided by apparent power (RMS V * RMS A)

[coppice]

It is more correct to just describe power factor as a ratio between the active power (integration of V*A) divided by apparent power (RMS V * RMS A)

That is not how most specs. define power factor for non-sinusoidal signals, although it is one of the definitions.

[Pentium100]

That occurred because the cos(phi) was so low, that the current was incredibly high, even if the real power was around 1kW, maybe 2kW or so. Unfortunately, the technician solved the problem by changing the local breaker with one of higher ratings (instead of purchasing decent PSUs, with good PFC).

PFC on computer PSUs can be trouble. If the PSU is cheaper and its hot side capacitor fails, the PFC transistor (or something else) fails short, tripping the circuit breaker and blowing up the PSU (seen three PSUs where this happened, repaired two of them, the third one is waiting for its turn).

Another problem that even the more expensive PSUs with PFS have is non-sinusoidal input waveform (for example, the output from a cheap UPS) or an input with unstable frequency and voltage (for example, the output from a generator without an inverter).