"Ken Smith" a écrit dans le message de news:dfs65p$7q5$ snipped-for-privacy@blue.rahul.net...
About a month ago, a good friend of mine who's doing expertise for insurance companies called me for analysing some PC PSU disasters. I, for the first time, saw a PC supply that used a heavy iron cored inductance head filter on primary side to extand conduction angle. IIRC it was Compaq, ca. 1999-2000
I think most of them naturally draw something like this from each line:
**************** ........... 0A ****
When a 3 phase bridge is running into an inductor, each leg spends 1/3rd the time passing no current. The other two are providing all the power as it goes through the zero crossing. The result is that it already has a fairly low harmonic content compared to the capacitive filtered case.
I read in sci.electronics.design that Terry Given wrote (in ) about 'Toridal Mains Transformer', on Sun, 11 Sep 2005:
It's actually quite close to sinusoidal, but of course the apparent frequency is much higher than the supply frequency.
Even more that for a square wave if the conduction angle is anywhere in the practical range. Even if the power supply just meets IEC/EN
61000-3-2 (conduction angle 65 degrees), the 3rd harmonic is 0.78 of the fundamental, not the 0.33 of a full-half period square wave, the 5th is
0.43, not 0.2, etc.
I used to think that you couldn't get conduction angles below 30 degrees, but now low-power products have very big filter caps, so as to meet requirements immunity to supply voltage dips and interruptions, and this results in conduction angles down to 20 degrees.
--
Regards, John Woodgate, OOO - Own Opinions Only.
If everything has been designed, a god designed evolution by natural selection.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk
I think you are trying to say that (most) PFC units are ac-input boost converters, whose DC output is used to directly run the smps, which no longer has an input rectifier etc. I suppose you can tack PFC units (vicor et al make such beasts) directly in front of a conventional smps, but that wastes an entire set of EMI filters, rectifiers etc.
Note also that there are many different types of PFC, and an input rectifier is not mandatory.
You are miles off the mark with linear supplies. Whats the 2nd biggest component in a linear supply? the DC bus capacitor. Why? because the conduction angle is only around 1-2ms (I'll leave the maths as an exercise), so for 8-9ms (at 50Hz) the cap is supplying the load, and for
1-2ms the supply re-charges the cap (this is the steep rising edge on the 100Hz ripple voltage) with a narrow, high current "spike".
This spike must have at least as much Integral(Idt) = charge as the average output current * 10ms (its DC so the integral is simple), which (assuming a rectangle) is 5* the average output current for 2ms conduction time. If you work out the frequency spectrum, you will see there are *shitloads* of (decreasing) odd harmonics - 3rd, 5th, 7th are dreadful. The bigger the DC bus cap, the smaller the conduction angle so the peak input current gets even higher!
Get an AC CT, and clip it onto *ANY* linear psu line cord, and look for yourself. Hell, solder 10 1R resistors in parallel, in series with the DC bus cap -ve terminal, and look directly with your scope.
Those harmonics are one of the reasons the AC voltage waveform often has flat tops. Harmonic propagation through PFCs/active filters is an interesting problem.
there has always been a lot of confusion around power factor and displacement factor. Lots of dodgy companies publish DF rather than PF figures for their rectifier-capacitor circuits, because DF=~1 (first harmonic is almost perfectly lined up with the voltage waveform, by the rectifier), and PF = crappy due to harmonics.
A former boss of mine spent a decade or two specifying DF in our manuals, for precisely that reason. Then him and a couple of other guys went and started up a (now very successful) business fixing the power factor messes he'd helped create. Guess what his first white paper was about - DF vs PF....
I read in sci.electronics.design that Terry Given wrote (in ) about 'Toridal Mains Transformer', on Mon, 12 Sep 2005:
The figure I gave show that it's a 'good approximation' if you will accept an underestimate of less than half the true harmonic amplitude.
--
Regards, John Woodgate, OOO - Own Opinions Only.
If everything has been designed, a god designed evolution by natural selection.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk
yep. The rectangle approach is a good first-order approximation though.
I'd believe that. I've noticed with lifetime calcs that increasing the capacitor size eventually decreases the cap lifetime, because the peak current is that much higher.
Having a (post-) regulator that can tolerate a wide input range is a much better solution. And eventually segues into full-blown PFC.
yeah, but its all IIRC, its been quite a few years since i read up on it.
the guts of the issue is that many PFCs *assume* the line voltage waveform is sinusoidal, and its often not. A common technique is to divide the rectified line voltage down and multiply it with the voltage loop error amplifier to get a current reference. If the voltage is distorted, so is the current reference.
Filter that through the system dynamics, and interesting things happen. For a start, its not drawing a nice sinusoidal current anymore. IIRC odd harmonics on the line voltage end up creating even harmonics in the line current, but that was for three-phase systems.
The sensed voltage is almost always distorted by the bridge rectifier, which is one of the reasons many PFC circuits have THDs of around 3-5%. Likewise not all multipliers are created equal.
I don't know what you mean. For a square wave, the 3rd harmonic is 1/3 of the fundamental. For a rectifier with 65 degree conduction angle, the third harmonic is 0.78 of the fundamental.
--
Regards, John Woodgate, OOO - Own Opinions Only.
If everything has been designed, a god designed evolution by natural selection.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk
we are talking about the same thing, ultimately - a thumping great narrow current spike, that kinda looks sinusoidal, and has truckloads of harmonics.
I have somewhere a mathcad worksheet I wrote to calculate THD, harmonics and PF for a rectifier/capacitor load, given only RMS current and crest factor.
Last century we were doing compliance tests for a range of small single- and three-phase drives (0.5kW - 10kW). When one of the techs gave me the RMS input current measurements, I was initially surprised - a 4.5A drive had about 20% more RMS current than a 12A drive. So we double-checked the measurements, which were correct. The reason? 5% line chokes on the larger drives, increasing the conduction angle.
You may be the expert, but I think the general definition of power factor is watts delivered divided by the product of RMS volts applied times RMS current passed. RMS current includes square root of the sum of the squares of the (possibly phase shifted) fundamental and each of the harmonics.
formatting link
In 3 phase systems, 3rd harmonic currents in the 3 lines add together in the neutral. They have been implicated in fires where the unfused neutral got so hot they set buildings on fire.
"Engineers are concerned with harmonics on building power distribution systems for a number of reasons. The main reason is that harmonic currents are additive in three-phase neutrals, causing them to overheat. Since the neutral conductor has no circuit breaker protection, as do the main conductors."
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.